rdelaage
/
postgresql
mirror of https://git.postgresql.org/git/postgresql.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
postgresql/src/backend/utils/misc/guc.c

8542 lines
213 KiB
C
Raw Normal View History

The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
/*--------------------------------------------------------------------
* guc.c
*
* Support for grand unified configuration scheme, including SET
* command, configuration file, and command line options.
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
* See src/backend/utils/misc/README for more information.
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
*
*
Stamp copyrights for year 2011.
2011-01-01 19:18:15 +01:00
* Copyright (c) 2000-2011, PostgreSQL Global Development Group
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
* Written by Peter Eisentraut <peter_e@gmx.net>.
In the continuing saga of FE/BE protocol revisions, add reporting of initial values and runtime changes in selected parameters. This gets rid of the need for an initial 'select pg_client_encoding()' query in libpq, bringing us back to one message transmitted in each direction for a standard connection startup. To allow server version to be sent using the same GUC mechanism that handles other parameters, invent the concept of a never-settable GUC parameter: you can 'show server_version' but it's not settable by any GUC input source. Create 'lc_collate' and 'lc_ctype' never-settable parameters so that people can find out these settings without need for pg_controldata. (These side ideas were all discussed some time ago in pgsql-hackers, but not yet implemented.)
2003-04-25 21:45:10 +02:00
*
* IDENTIFICATION
Remove cvs keywords from all files.
2010-09-20 22:08:53 +02:00
* src/backend/utils/misc/guc.c
In the continuing saga of FE/BE protocol revisions, add reporting of initial values and runtime changes in selected parameters. This gets rid of the need for an initial 'select pg_client_encoding()' query in libpq, bringing us back to one message transmitted in each direction for a standard connection startup. To allow server version to be sent using the same GUC mechanism that handles other parameters, invent the concept of a never-settable GUC parameter: you can 'show server_version' but it's not settable by any GUC input source. Create 'lc_collate' and 'lc_ctype' never-settable parameters so that people can find out these settings without need for pg_controldata. (These side ideas were all discussed some time ago in pgsql-hackers, but not yet implemented.)
2003-04-25 21:45:10 +02:00
*
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
*--------------------------------------------------------------------
*/
#include "postgres.h"
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
#include <ctype.h>
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
#include <float.h>
Implement prefetching via posix_fadvise() for bitmap index scans. A new GUC variable effective_io_concurrency controls how many concurrent block prefetch requests will be issued. (The best way to handle this for plain index scans is still under debate, so that part is not applied yet --- tgl) Greg Stark
2009-01-12 06:10:45 +01:00
#include <math.h>
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
#include <limits.h>
#include <unistd.h>
Whack some sense into the configuration-file-location patch. Refactor code into something reasonably understandable, cause use of the feature to not fail in standalone backends or in EXEC_BACKEND case, fix sloppy guc.c table entries, make the documentation minimally usable.
2004-10-08 03:36:36 +02:00
#include <sys/stat.h>
Fix syslog bug: if any messages are emitted to write_syslog before the facility has been set, the facility gets set to LOCAL0 and cannot be changed later. This seems reasonably plausible to happen, particularly at higher debug log levels, though I am not certain it explains Han Holl's recent report. Easiest fix is to teach the code how to change the value on-the-fly, which is nicer anyway. I made the settings PGC_SIGHUP to conform with log_destination.
2005-10-14 22:53:56 +02:00
#ifdef HAVE_SYSLOG
#include <syslog.h>
#endif
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
Remove hard-wired lists of timezone abbreviations in favor of providing configuration files that can be altered by a DBA. The australian_timezones GUC setting disappears, replaced by a timezone_abbreviations setting (set this to 'Australia' to get the effect of australian_timezones). The list of zone names defined by default has undergone a bit of cleanup, too. Documentation still needs some work --- in particular, should we fix Table B-4, or just get rid of it? Joachim Wieland, with some editorializing by moi.
2006-07-25 05:51:23 +02:00
#include "access/gin.h"
Put back yet another improperly-removed #include, per Mark Kirkwood.
2006-12-13 06:54:48 +01:00
#include "access/transam.h"
Two-phase commit. Original patch by Heikki Linnakangas, with additional hacking by Alvaro Herrera and Tom Lane.
2005-06-18 00:32:51 +02:00
#include "access/twophase.h"
Allow include files to compile own their own. Strip unused include files out unused include files, and add needed includes to C files. The next step is to remove unused include files in C files.
2006-07-13 18:49:20 +02:00
#include "access/xact.h"
Create a new GUC variable search_path to control the namespace search path. The default behavior if no per-user schemas are created is that all users share a 'public' namespace, thus providing behavior backwards compatible with 7.2 and earlier releases. Probably the semantics and default setting will need to be fine-tuned, but this is a start.
2002-04-01 05:34:27 +02:00
#include "catalog/namespace.h"
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
#include "commands/async.h"
RESET SESSION, plus related new DDL commands. Patch from Marko Kreen, reviewed by Neil Conway. This patch adds the following DDL command variants: RESET SESSION, RESET TEMP, RESET PLANS, CLOSE ALL, and DEALLOCATE ALL. RESET SESSION is intended for use by connection pool software and the like, in order to reset a client session to something close to its initial state. Note that while most of these command variants can be executed inside a transaction block (but are not transaction-aware!), RESET SESSION cannot. While this is inconsistent, it is intended to catch programmer mistakes: RESET SESSION in an open transaction block is probably unintended.
2007-04-12 08:53:49 +02:00
#include "commands/prepare.h"
Instead of having a configure-time DEFAULT_ATTSTATTARGET, store -1 in attstattarget to indicate 'use the default'. The default is now a GUC variable default_statistics_target, and so may be changed on the fly. Along the way we gain the ability to have pg_dump dump the per-column statistics target when it's not the default. Patch by Neil Conway, with some kibitzing from Tom Lane.
2002-07-31 19:19:54 +02:00
#include "commands/vacuum.h"
Alphabetically order reference to include files, "G" - "M".
2006-07-11 19:04:13 +02:00
#include "commands/variable.h"
Changes pg_trigger and extend pg_rewrite in order to allow triggers and rules to be defined with different, per session controllable, behaviors for replication purposes. This will allow replication systems like Slony-I and, as has been stated on pgsql-hackers, other products to control the firing mechanism of triggers and rewrite rules without modifying the system catalog directly. The firing mechanisms are controlled by a new superuser-only GUC variable, session_replication_role, together with a change to pg_trigger.tgenabled and a new column pg_rewrite.ev_enabled. Both columns are a single char data type now (tgenabled was a bool before). The possible values in these attributes are: 'O' - Trigger/Rule fires when session_replication_role is "origin" (default) or "local". This is the default behavior. 'D' - Trigger/Rule is disabled and fires never 'A' - Trigger/Rule fires always regardless of the setting of session_replication_role 'R' - Trigger/Rule fires when session_replication_role is "replica" The GUC variable can only be changed as long as the system does not have any cached query plans. This will prevent changing the session role and accidentally executing stored procedures or functions that have plans cached that expand to the wrong query set due to differences in the rule firing semantics. The SQL syntax for changing a triggers/rules firing semantics is ALTER TABLE <tabname> <when> TRIGGER|RULE <name>; <when> ::= ENABLE | ENABLE ALWAYS | ENABLE REPLICA | DISABLE psql's \d command as well as pg_dump are extended in a backward compatible fashion. Jan
2007-03-20 00:38:32 +01:00
#include "commands/trigger.h"
As discussed on several occasions previously, the new anonymous composite type capability makes it possible to create a system view based on a table function in a way that is hopefully palatable to everyone. The attached patch takes advantage of this, moving show_all_settings() from contrib/tablefunc into the backend (renamed all_settings(). It is defined as a builtin returning type RECORD. During initdb a system view is created to expose the same information presently available through SHOW ALL. For example: test=# select * from pg_settings where name like '%debug%'; name | setting -----------------------+--------- debug_assertions | on debug_pretty_print | off debug_print_parse | off debug_print_plan | off debug_print_query | off debug_print_rewritten | off wal_debug | 0 (7 rows) Additionally during initdb two rules are created which make it possible to change settings by updating the system view -- a "virtual table" as Tom put it. Here's an example: Joe Conway
2002-08-15 04:51:27 +02:00
#include "funcapi.h"
Make the location of the Kerberos server key file run time configurable (rather than compile time). For libpq, even when Kerberos support is compiled in, the default user name should still fall back to geteuid() if it can't be determined via the Kerberos system. A couple of fixes for string type configuration parameters, now that there is one.
2000-08-25 12:00:35 +02:00
#include "libpq/auth.h"
Add large object access control. A new system catalog pg_largeobject_metadata manages ownership and access privileges of large objects. KaiGai Kohei, reviewed by Jaime Casanova.
2009-12-11 04:34:57 +01:00
#include "libpq/be-fsstubs.h"
In the continuing saga of FE/BE protocol revisions, add reporting of initial values and runtime changes in selected parameters. This gets rid of the need for an initial 'select pg_client_encoding()' query in libpq, bringing us back to one message transmitted in each direction for a standard connection startup. To allow server version to be sent using the same GUC mechanism that handles other parameters, invent the concept of a never-settable GUC parameter: you can 'show server_version' but it's not settable by any GUC input source. Create 'lc_collate' and 'lc_ctype' never-settable parameters so that people can find out these settings without need for pg_controldata. (These side ideas were all discussed some time ago in pgsql-hackers, but not yet implemented.)
2003-04-25 21:45:10 +02:00
#include "libpq/pqformat.h"
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
#include "miscadmin.h"
#include "optimizer/cost.h"
#include "optimizer/geqo.h"
#include "optimizer/paths.h"
Teach planner how to rearrange join order for some classes of OUTER JOIN. Per my recent proposal. I ended up basing the implementation on the existing mechanism for enforcing valid join orders of IN joins --- the rules for valid outer-join orders are somewhat similar.
2005-12-20 03:30:36 +01:00
#include "optimizer/planmain.h"
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
#include "parser/parse_expr.h"
Support type modifiers for user-defined types, and pull most knowledge about typmod representation for standard types out into type-specific typmod I/O functions. Teodor Sigaev, with some editorialization by Tom Lane.
2006-12-30 22:21:56 +01:00
#include "parser/parse_type.h"
Move some declarations in the raw-parser header files to create a clearer distinction between the external API (parser.h) and declarations that only need to be visible within the raw parser code (gramparse.h, which now is only included by parser.c, gram.y, scan.l, and keywords.c). This is in preparation for the upcoming change to a reentrant lexer, which will require referencing YYSTYPE in the declarations of base_yylex and filtered_base_yylex, hence gram.h will have to be included by gramparse.h. We don't want any more files than absolutely necessary to depend on gram.h, so some cleanup is called for.
2009-07-12 19:12:34 +02:00
#include "parser/parser.h"
Fix bug that allowed any logged-in user to SET ROLE to any other database user id (CVE-2006-0553). Also fix related bug in SET SESSION AUTHORIZATION that allows unprivileged users to crash the server, if it has been compiled with Asserts enabled. The escalation-of-privilege risk exists only in 8.1.0-8.1.2. However, the Assert-crash risk exists in all releases back to 7.3. Thanks to Akio Ishida for reporting this problem.
2006-02-12 23:32:43 +01:00
#include "parser/scansup.h"
Alphabetically order reference to include files, "G" - "M".
2006-07-11 19:04:13 +02:00
#include "pgstat.h"
Integrate autovacuum functionality into the backend. There's still a few loose ends to be dealt with, but it seems to work. Alvaro Herrera, based on the contrib code by Matthew O'Connor.
2005-07-14 07:13:45 +02:00
#include "postmaster/autovacuum.h"
Separate out bgwriter code into a logically separate module, rather than being random pieces of other files. Give bgwriter responsibility for all checkpoint activity (other than a post-recovery checkpoint); so this child process absorbs the functionality of the former transient checkpoint and shutdown subprocesses. While at it, create an actual include file for postmaster.c, which for some reason never had its own file before.
2004-05-30 00:48:23 +02:00
#include "postmaster/bgwriter.h"
#include "postmaster/postmaster.h"
Alphabetically order reference to include files, "G" - "M".
2006-07-11 19:04:13 +02:00
#include "postmaster/syslogger.h"
Create a new dedicated Postgres process, "wal writer", which exists to write and fsync WAL at convenient intervals. For the moment it just tries to offload this work from backends, but soon it will be responsible for guaranteeing a maximum delay before asynchronously-committed transactions will be flushed to disk. This is a portion of Simon Riggs' async-commit patch, committed to CVS separately because a background WAL writer seems like it might be a good idea independently of the async-commit feature. I rebased walwriter.c on bgwriter.c because it seemed like a more appropriate way of handling signals; while the startup/shutdown logic in postmaster.c is more like autovac because we want walwriter to quit before we start the shutdown checkpoint.
2007-07-24 06:54:09 +02:00
#include "postmaster/walwriter.h"
Efficient transaction-controlled synchronous replication. If a standby is broadcasting reply messages and we have named one or more standbys in synchronous_standby_names then allow users who set synchronous_replication to wait for commit, which then provides strict data integrity guarantees. Design avoids sending and receiving transaction state information so minimises bookkeeping overheads. We synchronize with the highest priority standby that is connected and ready to synchronize. Other standbys can be defined to takeover in case of standby failure. This version has very strict behaviour; more relaxed options may be added at a later date. Simon Riggs and Fujii Masao, with reviews by Yeb Havinga, Jaime Casanova, Heikki Linnakangas and Robert Haas, plus the assistance of many other design reviewers.
2011-03-06 23:49:16 +01:00
#include "replication/syncrep.h"
Send status updates back from standby server to master, indicating how far the standby has written, flushed, and applied the WAL. At the moment, this is for informational purposes only, the values are only shown in pg_stat_replication system view, but in the future they will also be needed for synchronous replication. Extracted from Simon riggs' synchronous replication patch by Robert Haas, with some tweaking by me.
2011-02-10 20:00:29 +01:00
#include "replication/walreceiver.h"
Introduce Streaming Replication. This includes two new kinds of postmaster processes, walsenders and walreceiver. Walreceiver is responsible for connecting to the primary server and streaming WAL to disk, while walsender runs in the primary server and streams WAL from disk to the client. Documentation still needs work, but the basics are there. We will probably pull the replication section to a new chapter later on, as well as the sections describing file-based replication. But let's do that as a separate patch, so that it's easier to see what has been added/changed. This patch also adds a new section to the chapter about FE/BE protocol, documenting the protocol used by walsender/walreceivxer. Bump catalog version because of two new functions, pg_last_xlog_receive_location() and pg_last_xlog_replay_location(), for monitoring the progress of replication. Fujii Masao, with additional hacking by me
2010-01-15 10:19:10 +01:00
#include "replication/walsender.h"
Restructure some header files a bit, in particular heapam.h, by removing some unnecessary #include lines in it. Also, move some tuple routine prototypes and macros to htup.h, which allows removal of heapam.h inclusion from some .c files. For this to work, a new header file access/sysattr.h needed to be created, initially containing attribute numbers of system columns, for pg_dump usage. While at it, make contrib ltree, intarray and hstore header files more consistent with our header style.
2008-05-12 02:00:54 +02:00
#include "storage/bufmgr.h"
Replace max_standby_delay with two parameters, max_standby_archive_delay and max_standby_streaming_delay, and revise the implementation to avoid assuming that timestamps found in WAL records can meaningfully be compared to clock time on the standby server. Instead, the delay limits are compared to the elapsed time since we last obtained a new WAL segment from archive or since we were last "caught up" to WAL data arriving via streaming replication. This avoids problems with clock skew between primary and standby, as well as other corner cases that the original coding would misbehave in, such as the primary server having significant idle time between transactions. Per my complaint some time ago and considerable ensuing discussion. Do some desultory editing on the hot standby documentation, too.
2010-07-03 22:43:58 +02:00
#include "storage/standby.h"
Create a GUC parameter max_files_per_process that is a configurable upper limit on what we will believe from sysconf(_SC_OPEN_MAX). The default value is 1000, so that under ordinary conditions it won't affect the behavior. But on platforms where the kernel promises far more than it can deliver, this can be used to prevent running out of file descriptors. See numerous past discussions, eg, pgsql-hackers around 23-Dec-2000.
2001-09-30 20:57:45 +02:00
#include "storage/fd.h"
Implement genuine serializable isolation level. Until now, our Serializable mode has in fact been what's called Snapshot Isolation, which allows some anomalies that could not occur in any serialized ordering of the transactions. This patch fixes that using a method called Serializable Snapshot Isolation, based on research papers by Michael J. Cahill (see README-SSI for full references). In Serializable Snapshot Isolation, transactions run like they do in Snapshot Isolation, but a predicate lock manager observes the reads and writes performed and aborts transactions if it detects that an anomaly might occur. This method produces some false positives, ie. it sometimes aborts transactions even though there is no anomaly. To track reads we implement predicate locking, see storage/lmgr/predicate.c. Whenever a tuple is read, a predicate lock is acquired on the tuple. Shared memory is finite, so when a transaction takes many tuple-level locks on a page, the locks are promoted to a single page-level lock, and further to a single relation level lock if necessary. To lock key values with no matching tuple, a sequential scan always takes a relation-level lock, and an index scan acquires a page-level lock that covers the search key, whether or not there are any matching keys at the moment. A predicate lock doesn't conflict with any regular locks or with another predicate locks in the normal sense. They're only used by the predicate lock manager to detect the danger of anomalies. Only serializable transactions participate in predicate locking, so there should be no extra overhead for for other transactions. Predicate locks can't be released at commit, but must be remembered until all the transactions that overlapped with it have completed. That means that we need to remember an unbounded amount of predicate locks, so we apply a lossy but conservative method of tracking locks for committed transactions. If we run short of shared memory, we overflow to a new "pg_serial" SLRU pool. We don't currently allow Serializable transactions in Hot Standby mode. That would be hard, because even read-only transactions can cause anomalies that wouldn't otherwise occur. Serializable isolation mode now means the new fully serializable level. Repeatable Read gives you the old Snapshot Isolation level that we have always had. Kevin Grittner and Dan Ports, reviewed by Jeff Davis, Heikki Linnakangas and Anssi Kääriäinen
2011-02-07 22:46:51 +01:00
#include "storage/predicate.h"
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
#include "tcop/tcopprot.h"
Tsearch2 functionality migrates to core. The bulk of this work is by Oleg Bartunov and Teodor Sigaev, but I did a lot of editorializing, so anything that's broken is probably my fault. Documentation is nonexistent as yet, but let's land the patch so we can get some portability testing done.
2007-08-21 03:11:32 +02:00
#include "tsearch/ts_cache.h"
User and database-specific session defaults for run-time configuration variables. New commands ALTER DATABASE ... SET and ALTER USER ... SET.
2002-03-01 23:45:19 +01:00
#include "utils/builtins.h"
Support hex-string input and output for type BYTEA. Both hex format and the traditional "escape" format are automatically handled on input. The output format is selected by the new GUC variable bytea_output. As committed, bytea_output defaults to HEX, which is an *incompatible change*. We will keep it this way for awhile for testing purposes, but should consider whether to switch to the more backwards-compatible default of ESCAPE before 8.5 is released. Peter Eisentraut
2009-08-04 18:08:37 +02:00
#include "utils/bytea.h"
Alphabetically order reference to include files, "G" - "M".
2006-07-11 19:04:13 +02:00
#include "utils/guc_tables.h"
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
#include "utils/memutils.h"
Locale support is on by default. The choice of locale is done in initdb and/or with GUC variables.
2002-04-03 07:39:33 +02:00
#include "utils/pg_locale.h"
Changes pg_trigger and extend pg_rewrite in order to allow triggers and rules to be defined with different, per session controllable, behaviors for replication purposes. This will allow replication systems like Slony-I and, as has been stated on pgsql-hackers, other products to control the firing mechanism of triggers and rewrite rules without modifying the system catalog directly. The firing mechanisms are controlled by a new superuser-only GUC variable, session_replication_role, together with a change to pg_trigger.tgenabled and a new column pg_rewrite.ev_enabled. Both columns are a single char data type now (tgenabled was a bool before). The possible values in these attributes are: 'O' - Trigger/Rule fires when session_replication_role is "origin" (default) or "local". This is the default behavior. 'D' - Trigger/Rule is disabled and fires never 'A' - Trigger/Rule fires always regardless of the setting of session_replication_role 'R' - Trigger/Rule fires when session_replication_role is "replica" The GUC variable can only be changed as long as the system does not have any cached query plans. This will prevent changing the session role and accidentally executing stored procedures or functions that have plans cached that expand to the wrong query set due to differences in the rule firing semantics. The SQL syntax for changing a triggers/rules firing semantics is ALTER TABLE <tabname> <when> TRIGGER|RULE <name>; <when> ::= ENABLE | ENABLE ALWAYS | ENABLE REPLICA | DISABLE psql's \d command as well as pg_dump are extended in a backward compatible fashion. Jan
2007-03-20 00:38:32 +01:00
#include "utils/plancache.h"
RESET SESSION, plus related new DDL commands. Patch from Marko Kreen, reviewed by Neil Conway. This patch adds the following DDL command variants: RESET SESSION, RESET TEMP, RESET PLANS, CLOSE ALL, and DEALLOCATE ALL. RESET SESSION is intended for use by connection pool software and the like, in order to reset a client session to something close to its initial state. Note that while most of these command variants can be executed inside a transaction block (but are not transaction-aware!), RESET SESSION cannot. While this is inconsistent, it is intended to catch programmer mistakes: RESET SESSION in an open transaction block is probably unintended.
2007-04-12 08:53:49 +02:00
#include "utils/portal.h"
Add GUC update_process_title to control whether 'ps' display is updated for every command, default to on.
2006-06-28 00:16:44 +02:00
#include "utils/ps_status.h"
Remove hard-wired lists of timezone abbreviations in favor of providing configuration files that can be altered by a DBA. The australian_timezones GUC setting disappears, replaced by a timezone_abbreviations setting (set this to 'Australia' to get the effect of australian_timezones). The list of zone names defined by default has undergone a bit of cleanup, too. Documentation still needs some work --- in particular, should we fix Table B-4, or just get rid of it? Joachim Wieland, with some editorializing by moi.
2006-07-25 05:51:23 +02:00
#include "utils/tzparser.h"
Add support for converting binary values (i.e. bytea) into xml values, with new GUC parameter "xmlbinary" that controls the output encoding, as per SQL/XML standard.
2007-01-19 17:58:46 +01:00
#include "utils/xml.h"
Back out RESET CONNECTION until there is more discussion.
2006-04-25 16:11:59 +02:00
In the continuing saga of FE/BE protocol revisions, add reporting of initial values and runtime changes in selected parameters. This gets rid of the need for an initial 'select pg_client_encoding()' query in libpq, bringing us back to one message transmitted in each direction for a standard connection startup. To allow server version to be sent using the same GUC mechanism that handles other parameters, invent the concept of a never-settable GUC parameter: you can 'show server_version' but it's not settable by any GUC input source. Create 'lc_collate' and 'lc_ctype' never-settable parameters so that people can find out these settings without need for pg_controldata. (These side ideas were all discussed some time ago in pgsql-hackers, but not yet implemented.)
2003-04-25 21:45:10 +02:00
#ifndef PG_KRB_SRVTAB
#define PG_KRB_SRVTAB ""
#endif
Allow kerberos name and username case sensitivity to be specified from postgresql.conf. --------------------------------------------------------------------------- Here's an updated version of the patch, with the following changes: 1) No longer uses "service name" as "application version". It's instead hardcoded as "postgres". It could be argued that this part should be backpatched to 8.0, but it doesn't make a big difference until you can start changing it with GUC / connection parameters. This change only affects kerberos 5, not 4. 2) Now downcases kerberos usernames when the client is running on win32. 3) Adds guc option for "krb_caseins_users" to make the server ignore case mismatch which is required by some KDCs such as Active Directory. Off by default, per discussion with Tom. This change only affects kerberos 5, not 4. 4) Updated so it doesn't conflict with the rendevouz/bonjour patch already in ;-) Magnus Hagander
2005-06-04 22:42:43 +02:00
#ifndef PG_KRB_SRVNAM
#define PG_KRB_SRVNAM ""
#endif
In the continuing saga of FE/BE protocol revisions, add reporting of initial values and runtime changes in selected parameters. This gets rid of the need for an initial 'select pg_client_encoding()' query in libpq, bringing us back to one message transmitted in each direction for a standard connection startup. To allow server version to be sent using the same GUC mechanism that handles other parameters, invent the concept of a never-settable GUC parameter: you can 'show server_version' but it's not settable by any GUC input source. Create 'lc_collate' and 'lc_ctype' never-settable parameters so that people can find out these settings without need for pg_controldata. (These side ideas were all discussed some time ago in pgsql-hackers, but not yet implemented.)
2003-04-25 21:45:10 +02:00
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
#define CONFIG_FILENAME "postgresql.conf"
Adjust configuration-files GUC behavior as per my recent proposal. The vars are renamed to data_directory, config_file, hba_file, and ident_file, and are guaranteed to be set to accurate absolute paths during postmaster startup. This commit does not yet do anything about hiding path values from non-superusers.
2004-10-10 01:13:22 +02:00
#define HBA_FILENAME "pg_hba.conf"
#define IDENT_FILENAME "pg_ident.conf"
Whack some sense into the configuration-file-location patch. Refactor code into something reasonably understandable, cause use of the feature to not fail in standalone backends or in EXEC_BACKEND case, fix sloppy guc.c table entries, make the documentation minimally usable.
2004-10-08 03:36:36 +02:00
SECOND ATTEMPT Dump/read non-default GUC values for use by exec'ed backend, for Win32.
2003-05-03 00:02:47 +02:00
#ifdef EXEC_BACKEND
#define CONFIG_EXEC_PARAMS "global/config_exec_params"
Arrange for the postmaster (and standalone backends, initdb, etc) to chdir into PGDATA and subsequently use relative paths instead of absolute paths to access all files under PGDATA. This seems to give a small performance improvement, and it should make the system more robust against naive DBAs doing things like moving a database directory that has a live postmaster in it. Per recent discussion.
2005-07-04 06:51:52 +02:00
#define CONFIG_EXEC_PARAMS_NEW "global/config_exec_params.new"
SECOND ATTEMPT Dump/read non-default GUC values for use by exec'ed backend, for Win32.
2003-05-03 00:02:47 +02:00
#endif
In the continuing saga of FE/BE protocol revisions, add reporting of initial values and runtime changes in selected parameters. This gets rid of the need for an initial 'select pg_client_encoding()' query in libpq, bringing us back to one message transmitted in each direction for a standard connection startup. To allow server version to be sent using the same GUC mechanism that handles other parameters, invent the concept of a never-settable GUC parameter: you can 'show server_version' but it's not settable by any GUC input source. Create 'lc_collate' and 'lc_ctype' never-settable parameters so that people can find out these settings without need for pg_controldata. (These side ideas were all discussed some time ago in pgsql-hackers, but not yet implemented.)
2003-04-25 21:45:10 +02:00
Convert the arithmetic for shared memory size calculation from 'int' to 'Size' (that is, size_t), and install overflow detection checks in it. This allows us to remove the former arbitrary restrictions on NBuffers etc. It won't make any difference in a 32-bit machine, but in a 64-bit machine you could theoretically have terabytes of shared buffers. (How efficiently we could manage 'em remains to be seen.) Similarly, num_temp_buffers, work_mem, and maintenance_work_mem can be set above 2Gb on a 64-bit machine. Original patch from Koichi Suzuki, additional work by moi.
2005-08-21 01:26:37 +02:00
/* upper limit for GUC variables measured in kilobytes of memory */
Redefine Datum as uintptr_t, instead of unsigned long. This is more in keeping with modern practice, and is a first step towards porting to Win64 (which has sizeof(pointer) > sizeof(long)). Tsutomu Yamada, Magnus Hagander, Tom Lane
2009-12-31 20:41:37 +01:00
/* note that various places assume the byte size fits in a "long" variable */
#if SIZEOF_SIZE_T > 4 && SIZEOF_LONG > 4
Convert the arithmetic for shared memory size calculation from 'int' to 'Size' (that is, size_t), and install overflow detection checks in it. This allows us to remove the former arbitrary restrictions on NBuffers etc. It won't make any difference in a 32-bit machine, but in a 64-bit machine you could theoretically have terabytes of shared buffers. (How efficiently we could manage 'em remains to be seen.) Similarly, num_temp_buffers, work_mem, and maintenance_work_mem can be set above 2Gb on a 64-bit machine. Original patch from Koichi Suzuki, additional work by moi.
2005-08-21 01:26:37 +02:00
#define MAX_KILOBYTES INT_MAX
#else
#define MAX_KILOBYTES (INT_MAX / 1024)
#endif
Include the backend ID in the relpath of temporary relations. This allows us to reliably remove all leftover temporary relation files on cluster startup without reference to system catalogs or WAL; therefore, we no longer include temporary relations in XLOG_XACT_COMMIT and XLOG_XACT_ABORT WAL records. Since these changes require including a backend ID in each SharedInvalSmgrMsg, the size of the SharedInvalidationMessage.id field has been reduced from two bytes to one, and the maximum number of connections has been reduced from INT_MAX / 4 to 2^23-1. It would be possible to remove these restrictions by increasing the size of SharedInvalidationMessage by 4 bytes, but right now that doesn't seem like a good trade-off. Review by Jaime Casanova and Tom Lane.
2010-08-13 22:10:54 +02:00
/*
* Note: MAX_BACKENDS is limited to 2^23-1 because inval.c stores the
* backend ID as a 3-byte signed integer. Even if that limitation were
* removed, we still could not exceed INT_MAX/4 because some places compute
* 4*MaxBackends without any overflow check. This is rechecked in
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
* check_maxconnections, since MaxBackends is computed as MaxConnections
Include the backend ID in the relpath of temporary relations. This allows us to reliably remove all leftover temporary relation files on cluster startup without reference to system catalogs or WAL; therefore, we no longer include temporary relations in XLOG_XACT_COMMIT and XLOG_XACT_ABORT WAL records. Since these changes require including a backend ID in each SharedInvalSmgrMsg, the size of the SharedInvalidationMessage.id field has been reduced from two bytes to one, and the maximum number of connections has been reduced from INT_MAX / 4 to 2^23-1. It would be possible to remove these restrictions by increasing the size of SharedInvalidationMessage by 4 bytes, but right now that doesn't seem like a good trade-off. Review by Jaime Casanova and Tom Lane.
2010-08-13 22:10:54 +02:00
* plus autovacuum_max_workers plus one (for the autovacuum launcher).
*/
#define MAX_BACKENDS 0x7fffff
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
#define KB_PER_MB (1024)
#define KB_PER_GB (1024*1024)
#define MS_PER_S 1000
Provide a HINT listing the allowed unit names when a GUC variable seems to contain a wrong unit specification, per discussion. In passing, fix the code to avoid unnecessary integer overflows when converting units, and to detect overflows when they do occur.
2007-06-21 20:14:21 +02:00
#define S_PER_MIN 60
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
#define MS_PER_MIN (1000 * 60)
Provide a HINT listing the allowed unit names when a GUC variable seems to contain a wrong unit specification, per discussion. In passing, fix the code to avoid unnecessary integer overflows when converting units, and to detect overflows when they do occur.
2007-06-21 20:14:21 +02:00
#define MIN_PER_H 60
#define S_PER_H (60 * 60)
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
#define MS_PER_H (1000 * 60 * 60)
Provide a HINT listing the allowed unit names when a GUC variable seems to contain a wrong unit specification, per discussion. In passing, fix the code to avoid unnecessary integer overflows when converting units, and to detect overflows when they do occur.
2007-06-21 20:14:21 +02:00
#define MIN_PER_D (60 * 24)
#define S_PER_D (60 * 60 * 24)
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
#define MS_PER_D (1000 * 60 * 60 * 24)
In the continuing saga of FE/BE protocol revisions, add reporting of initial values and runtime changes in selected parameters. This gets rid of the need for an initial 'select pg_client_encoding()' query in libpq, bringing us back to one message transmitted in each direction for a standard connection startup. To allow server version to be sent using the same GUC mechanism that handles other parameters, invent the concept of a never-settable GUC parameter: you can 'show server_version' but it's not settable by any GUC input source. Create 'lc_collate' and 'lc_ctype' never-settable parameters so that people can find out these settings without need for pg_controldata. (These side ideas were all discussed some time ago in pgsql-hackers, but not yet implemented.)
2003-04-25 21:45:10 +02:00
/* XXX these should appear in other modules' header files */
This patch brings up to date what I did last year (now unfortunately bitrotted) to allow the logging of the end of a session, enabled by the config setting "log_disconnections". Andrew Dunstan
2004-02-17 04:54:57 +01:00
extern bool Log_disconnections;
pgindent run. Make it all clean.
2001-03-22 05:01:46 +01:00
extern int CommitDelay;
extern int CommitSiblings;
Create 'default_tablespace' GUC variable that supplies a TABLESPACE clause implicitly whenever one is not given explicitly. Remove concept of a schema having an associated tablespace, and simplify the rules for selecting a default tablespace for a table or index. It's now just (a) explicit TABLESPACE clause; (b) default_tablespace if that's not an empty string; (c) database's default. This will allow pg_dump to use SET commands instead of tablespace clauses to determine object locations (but I didn't actually make it do so). All per recent discussions.
2004-11-05 20:17:13 +01:00
extern char *default_tablespace;
Create a GUC parameter temp_tablespaces that allows selection of the tablespace(s) in which to store temp tables and temporary files. This is a list to allow spreading the load across multiple tablespaces (a random list element is chosen each time a temp object is to be created). Temp files are not stored in per-database pgsql_tmp/ directories anymore, but per-tablespace directories. Jaime Casanova and Albert Cervera, with review by Bernd Helmle and Tom Lane.
2007-06-03 19:08:34 +02:00
extern char *temp_tablespaces;
Add a GUC variable "synchronize_seqscans" to allow clients to disable the new synchronized-scanning behavior, and make pg_dump disable sync scans so that it will reliably preserve row ordering. Per recent discussions.
2008-01-30 19:35:55 +01:00
extern bool synchronize_seqscans;
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
extern bool fullPageWrites;
Add configuration parameter ssl_renegotiation_limit to control how often we do SSL session key renegotiation. Can be set to 0 to disable renegotiation completely, which is required if a broken SSL library is used (broken patches to CVE-2009-3555 a known cause) or when using a client library that can't do renegotiation.
2010-02-25 14:26:16 +01:00
extern int ssl_renegotiation_limit;
Make GUC variables for syslog and SSL always visible Make the variables visible (but not used) even when support is not compiled in.
2010-12-18 16:53:59 +01:00
extern char *SSLCipherSuites;
Allow read only connections during recovery, known as Hot Standby. Enabled by recovery_connections = on (default) and forcing archive recovery using a recovery.conf. Recovery processing now emulates the original transactions as they are replayed, providing full locking and MVCC behaviour for read only queries. Recovery must enter consistent state before connections are allowed, so there is a delay, typically short, before connections succeed. Replay of recovering transactions can conflict and in some cases deadlock with queries during recovery; these result in query cancellation after max_standby_delay seconds have expired. Infrastructure changes have minor effects on normal running, though introduce four new types of WAL record. New test mode "make standbycheck" allows regression tests of static command behaviour on a standby server while in recovery. Typical and extreme dynamic behaviours have been checked via code inspection and manual testing. Few port specific behaviours have been utilised, though primary testing has been on Linux only so far. This commit is the basic patch. Additional changes will follow in this release to enhance some aspects of behaviour, notably improved handling of conflicts, deadlock detection and query cancellation. Changes to VACUUM FULL are also required. Simon Riggs, with significant and lengthy review by Heikki Linnakangas, including streamlined redesign of snapshot creation and two-phase commit. Important contributions from Florian Pflug, Mark Kirkwood, Merlin Moncure, Greg Stark, Gianni Ciolli, Gabriele Bartolini, Hannu Krosing, Robert Haas, Tatsuo Ishii, Hiroyuki Yamada plus support and feedback from many other community members.
2009-12-19 02:32:45 +01:00
Add a trace_sort option to help with measuring resource usage of external sort operations. Per recent discussion. Simon Riggs and Tom Lane.
2005-10-04 00:55:56 +02:00
#ifdef TRACE_SORT
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
extern bool trace_sort;
Add a trace_sort option to help with measuring resource usage of external sort operations. Per recent discussion. Simon Riggs and Tom Lane.
2005-10-04 00:55:56 +02:00
#endif
Arrange for large sequential scans to synchronize with each other, so that when multiple backends are scanning the same relation concurrently, each page is (ideally) read only once. Jeff Davis, with review by Heikki and Tom.
2007-06-08 20:23:53 +02:00
#ifdef TRACE_SYNCSCAN
extern bool trace_syncscan;
#endif
Teach tuplesort.c about "top N" sorting, in which only the first N tuples need be returned. We keep a heap of the current best N tuples and sift-up new tuples into it as we scan the input. For M input tuples this means only about M*log(N) comparisons instead of M*log(M), not to mention a lot less workspace when N is small --- avoiding spill-to-disk for large M is actually the most attractive thing about it. Patch includes planner and executor support for invoking this facility in ORDER BY ... LIMIT queries. Greg Stark, with some editorialization by moi.
2007-05-04 03:13:45 +02:00
#ifdef DEBUG_BOUNDED_SORT
extern bool optimize_bounded_sort;
#endif
FixBTree flag still exists and may be used to turn runtime recovery OFF.
2001-02-08 00:36:22 +01:00
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
static int GUC_check_errcode_value;
/* global variables for check hook support */
pgindent run before PG 9.1 beta 1.
2011-04-10 17:42:00 +02:00
char *GUC_check_errmsg_string;
char *GUC_check_errdetail_string;
char *GUC_check_errhint_string;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
Add "source file" and "source line" information to each GUC variable. initdb forced due to changes in the pg_settings view. Magnus Hagander and Alvaro Herrera.
2008-09-10 20:09:20 +02:00
static void set_config_sourcefile(const char *name, char *sourcefile,
int sourceline);
pgindent run before PG 9.1 beta 1.
2011-04-10 17:42:00 +02:00
static bool call_bool_check_hook(struct config_bool * conf, bool *newval,
void **extra, GucSource source, int elevel);
static bool call_int_check_hook(struct config_int * conf, int *newval,
void **extra, GucSource source, int elevel);
static bool call_real_check_hook(struct config_real * conf, double *newval,
void **extra, GucSource source, int elevel);
static bool call_string_check_hook(struct config_string * conf, char **newval,
void **extra, GucSource source, int elevel);
static bool call_enum_check_hook(struct config_enum * conf, int *newval,
void **extra, GucSource source, int elevel);
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
static bool check_log_destination(char **newval, void **extra, GucSource source);
static void assign_log_destination(const char *newval, void *extra);
Remove 'syslog' GUC variable, and add more logical 'log_destination' variable to control logoutput location on Unix and Win32. Magnus Hagander
2004-04-05 05:02:11 +02:00
Remove --enable-syslog option
2002-04-21 02:22:52 +02:00
#ifdef HAVE_SYSLOG
Fix syslog bug: if any messages are emitted to write_syslog before the facility has been set, the facility gets set to LOCAL0 and cannot be changed later. This seems reasonably plausible to happen, particularly at higher debug log levels, though I am not certain it explains Han Holl's recent report. Easiest fix is to teach the code how to change the value on-the-fly, which is nicer anyway. I made the settings PGC_SIGHUP to conform with log_destination.
2005-10-14 22:53:56 +02:00
static int syslog_facility = LOG_LOCAL0;
Make GUC variables for syslog and SSL always visible Make the variables visible (but not used) even when support is not compiled in.
2010-12-18 16:53:59 +01:00
#else
static int syslog_facility = 0;
#endif
Allow default transaction isolation level (a.k.a. set session characteristics) to be set through GUC.
2001-07-01 00:03:26 +02:00
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
static void assign_syslog_facility(int newval, void *extra);
static void assign_syslog_ident(const char *newval, void *extra);
static void assign_session_replication_role(int newval, void *extra);
static bool check_temp_buffers(int *newval, void **extra, GucSource source);
static bool check_phony_autocommit(bool *newval, void **extra, GucSource source);
static bool check_debug_assertions(bool *newval, void **extra, GucSource source);
static bool check_bonjour(bool *newval, void **extra, GucSource source);
static bool check_ssl(bool *newval, void **extra, GucSource source);
static bool check_stage_log_stats(bool *newval, void **extra, GucSource source);
static bool check_log_stats(bool *newval, void **extra, GucSource source);
static bool check_canonical_path(char **newval, void **extra, GucSource source);
static bool check_timezone_abbreviations(char **newval, void **extra, GucSource source);
static void assign_timezone_abbreviations(const char *newval, void *extra);
Simplify handling of the timezone GUC by making initdb choose the default. We were doing some amazingly complicated things in order to avoid running the very expensive identify_system_timezone() procedure during GUC initialization. But there is an obvious fix for that, which is to do it once during initdb and have initdb install the system-specific default into postgresql.conf, as it already does for most other GUC variables that need system-environment-dependent defaults. This means that the timezone (and log_timezone) settings no longer have any magic behavior in the server. Per discussion.
2011-09-09 23:59:11 +02:00
static void pg_timezone_abbrev_initialize(void);
Minor improvements in backup and recovery: - create a separate archive_mode GUC, on which archive_command is dependent - %r option in recovery.conf sends last restartpoint to recovery command - %r used in pg_standby, updated README - minor other code cleanup in pg_standby - doc on Warm Standby now mentions pg_standby and %r - log_restartpoints recovery option emits LOG message at each restartpoint - end of recovery now displays last transaction end time, as requested by Warren Little; also shown at each restartpoint - restart archiver if needed to carry away WAL files at shutdown Simon Riggs
2007-09-27 00:36:30 +02:00
static const char *show_archive_command(void);
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
static void assign_tcp_keepalives_idle(int newval, void *extra);
static void assign_tcp_keepalives_interval(int newval, void *extra);
static void assign_tcp_keepalives_count(int newval, void *extra);
Add GUC variables to control keep-alive times for idle, interval, and count. Oliver Jowett
2005-07-30 17:17:26 +02:00
static const char *show_tcp_keepalives_idle(void);
static const char *show_tcp_keepalives_interval(void);
static const char *show_tcp_keepalives_count(void);
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
static bool check_maxconnections(int *newval, void **extra, GucSource source);
static void assign_maxconnections(int newval, void *extra);
static bool check_autovacuum_max_workers(int *newval, void **extra, GucSource source);
static void assign_autovacuum_max_workers(int newval, void *extra);
static bool check_effective_io_concurrency(int *newval, void **extra, GucSource source);
static void assign_effective_io_concurrency(int newval, void *extra);
static void assign_pgstat_temp_directory(const char *newval, void *extra);
static bool check_application_name(char **newval, void **extra, GucSource source);
static void assign_application_name(const char *newval, void *extra);
Add a log_file_mode GUC that allows control of the file permissions set on log files created by the syslogger process. In passing, make unix_file_permissions display its value in octal, same as log_file_mode now does. Martin Pihlak
2010-07-17 00:25:51 +02:00
static const char *show_unix_socket_permissions(void);
static const char *show_log_file_mode(void);
In the continuing saga of FE/BE protocol revisions, add reporting of initial values and runtime changes in selected parameters. This gets rid of the need for an initial 'select pg_client_encoding()' query in libpq, bringing us back to one message transmitted in each direction for a standard connection startup. To allow server version to be sent using the same GUC mechanism that handles other parameters, invent the concept of a never-settable GUC parameter: you can 'show server_version' but it's not settable by any GUC input source. Create 'lc_collate' and 'lc_ctype' never-settable parameters so that people can find out these settings without need for pg_controldata. (These side ideas were all discussed some time ago in pgsql-hackers, but not yet implemented.)
2003-04-25 21:45:10 +02:00
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
static char *config_enum_get_options(struct config_enum * record,
const char *prefix, const char *suffix,
const char *separator);
Some cleanups of enum-guc code, per comments from Tom.
2008-03-16 17:42:44 +01:00
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
/*
* Options for enum values defined in this module.
Make the enumvals column of pg_settings be text[] instead of just a comma separated string.
2008-11-21 19:49:24 +01:00
*
* NOTE! Option values may not contain double quotes!
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
*/
Split apart message_level_options into one set for server-side settings and one for client-side, restoring the previous behaviour with different sort order for the 'log' level. Also, remove redundant list of available options, since the enum code will output it automatically.
2008-07-01 08:08:31 +02:00
Support hex-string input and output for type BYTEA. Both hex format and the traditional "escape" format are automatically handled on input. The output format is selected by the new GUC variable bytea_output. As committed, bytea_output defaults to HEX, which is an *incompatible change*. We will keep it this way for awhile for testing purposes, but should consider whether to switch to the more backwards-compatible default of ESCAPE before 8.5 is released. Peter Eisentraut
2009-08-04 18:08:37 +02:00
static const struct config_enum_entry bytea_output_options[] = {
{"escape", BYTEA_OUTPUT_ESCAPE, false},
{"hex", BYTEA_OUTPUT_HEX, false},
{NULL, 0, false}
};
Split apart message_level_options into one set for server-side settings and one for client-side, restoring the previous behaviour with different sort order for the 'log' level. Also, remove redundant list of available options, since the enum code will output it automatically.
2008-07-01 08:08:31 +02:00
/*
* We have different sets for client and server message level options because
* they sort slightly different (see "log" level)
*/
static const struct config_enum_entry client_message_level_options[] = {
"debug" level was supposed to be hidden, since it's just an alias for debug2.
2008-07-01 08:36:11 +02:00
{"debug", DEBUG2, true},
Add a field to guc enums to allow hiding of values from display while still accepting them as input, used to allow alternate syntax for the same setting. Alex Hunsaker
2008-05-28 11:04:06 +02:00
{"debug5", DEBUG5, false},
{"debug4", DEBUG4, false},
{"debug3", DEBUG3, false},
{"debug2", DEBUG2, false},
{"debug1", DEBUG1, false},
{"log", LOG, false},
Split apart message_level_options into one set for server-side settings and one for client-side, restoring the previous behaviour with different sort order for the 'log' level. Also, remove redundant list of available options, since the enum code will output it automatically.
2008-07-01 08:08:31 +02:00
{"info", INFO, true},
{"notice", NOTICE, false},
{"warning", WARNING, false},
{"error", ERROR, false},
{"fatal", FATAL, true},
{"panic", PANIC, true},
{NULL, 0, false}
};
static const struct config_enum_entry server_message_level_options[] = {
"debug" level was supposed to be hidden, since it's just an alias for debug2.
2008-07-01 08:36:11 +02:00
{"debug", DEBUG2, true},
Split apart message_level_options into one set for server-side settings and one for client-side, restoring the previous behaviour with different sort order for the 'log' level. Also, remove redundant list of available options, since the enum code will output it automatically.
2008-07-01 08:08:31 +02:00
{"debug5", DEBUG5, false},
{"debug4", DEBUG4, false},
{"debug3", DEBUG3, false},
{"debug2", DEBUG2, false},
{"debug1", DEBUG1, false},
Add a field to guc enums to allow hiding of values from display while still accepting them as input, used to allow alternate syntax for the same setting. Alex Hunsaker
2008-05-28 11:04:06 +02:00
{"info", INFO, false},
{"notice", NOTICE, false},
{"warning", WARNING, false},
{"error", ERROR, false},
Split apart message_level_options into one set for server-side settings and one for client-side, restoring the previous behaviour with different sort order for the 'log' level. Also, remove redundant list of available options, since the enum code will output it automatically.
2008-07-01 08:08:31 +02:00
{"log", LOG, false},
Add a field to guc enums to allow hiding of values from display while still accepting them as input, used to allow alternate syntax for the same setting. Alex Hunsaker
2008-05-28 11:04:06 +02:00
{"fatal", FATAL, false},
{"panic", PANIC, false},
{NULL, 0, false}
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
};
Add a new GUC variable called "IntervalStyle" that decouples interval output from DateStyle, and create a new interval style that produces output matching the SQL standard (at least for interval values that fall within the standard's restrictions). IntervalStyle is also used to resolve the conflict between the standard and traditional Postgres rules for interpreting negative interval input. Ron Mayer
2008-11-09 01:28:35 +01:00
static const struct config_enum_entry intervalstyle_options[] = {
{"postgres", INTSTYLE_POSTGRES, false},
{"postgres_verbose", INTSTYLE_POSTGRES_VERBOSE, false},
{"sql_standard", INTSTYLE_SQL_STANDARD, false},
Add support for input and output of interval values formatted per ISO 8601; specifically, we can input either the "format with designators" or the "alternative format", and we can output the former when IntervalStyle is set to iso_8601. Ron Mayer
2008-11-11 03:42:33 +01:00
{"iso_8601", INTSTYLE_ISO_8601, false},
Add a new GUC variable called "IntervalStyle" that decouples interval output from DateStyle, and create a new interval style that produces output matching the SQL standard (at least for interval values that fall within the standard's restrictions). IntervalStyle is also used to resolve the conflict between the standard and traditional Postgres rules for interpreting negative interval input. Ron Mayer
2008-11-09 01:28:35 +01:00
{NULL, 0, false}
};
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
static const struct config_enum_entry log_error_verbosity_options[] = {
Add a field to guc enums to allow hiding of values from display while still accepting them as input, used to allow alternate syntax for the same setting. Alex Hunsaker
2008-05-28 11:04:06 +02:00
{"terse", PGERROR_TERSE, false},
Remove GUC extra_desc strings that are redundant with the enum value lists.
2008-07-01 23:07:33 +02:00
{"default", PGERROR_DEFAULT, false},
Add a field to guc enums to allow hiding of values from display while still accepting them as input, used to allow alternate syntax for the same setting. Alex Hunsaker
2008-05-28 11:04:06 +02:00
{"verbose", PGERROR_VERBOSE, false},
{NULL, 0, false}
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
};
static const struct config_enum_entry log_statement_options[] = {
Add a field to guc enums to allow hiding of values from display while still accepting them as input, used to allow alternate syntax for the same setting. Alex Hunsaker
2008-05-28 11:04:06 +02:00
{"none", LOGSTMT_NONE, false},
{"ddl", LOGSTMT_DDL, false},
{"mod", LOGSTMT_MOD, false},
{"all", LOGSTMT_ALL, false},
{NULL, 0, false}
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
};
Convert three more guc settings to enum type: default_transaction_isolation, session_replication_role and regex_flavor.
2008-04-02 16:42:56 +02:00
static const struct config_enum_entry isolation_level_options[] = {
Add a field to guc enums to allow hiding of values from display while still accepting them as input, used to allow alternate syntax for the same setting. Alex Hunsaker
2008-05-28 11:04:06 +02:00
{"serializable", XACT_SERIALIZABLE, false},
{"repeatable read", XACT_REPEATABLE_READ, false},
{"read committed", XACT_READ_COMMITTED, false},
{"read uncommitted", XACT_READ_UNCOMMITTED, false},
Convert three more guc settings to enum type: default_transaction_isolation, session_replication_role and regex_flavor.
2008-04-02 16:42:56 +02:00
{NULL, 0}
};
static const struct config_enum_entry session_replication_role_options[] = {
Add a field to guc enums to allow hiding of values from display while still accepting them as input, used to allow alternate syntax for the same setting. Alex Hunsaker
2008-05-28 11:04:06 +02:00
{"origin", SESSION_REPLICATION_ROLE_ORIGIN, false},
{"replica", SESSION_REPLICATION_ROLE_REPLICA, false},
{"local", SESSION_REPLICATION_ROLE_LOCAL, false},
{NULL, 0, false}
Convert three more guc settings to enum type: default_transaction_isolation, session_replication_role and regex_flavor.
2008-04-02 16:42:56 +02:00
};
Convert syslog_facility guc to enum type.
2008-04-03 11:21:15 +02:00
static const struct config_enum_entry syslog_facility_options[] = {
Make GUC variables for syslog and SSL always visible Make the variables visible (but not used) even when support is not compiled in.
2010-12-18 16:53:59 +01:00
#ifdef HAVE_SYSLOG
Add a field to guc enums to allow hiding of values from display while still accepting them as input, used to allow alternate syntax for the same setting. Alex Hunsaker
2008-05-28 11:04:06 +02:00
{"local0", LOG_LOCAL0, false},
{"local1", LOG_LOCAL1, false},
{"local2", LOG_LOCAL2, false},
{"local3", LOG_LOCAL3, false},
{"local4", LOG_LOCAL4, false},
{"local5", LOG_LOCAL5, false},
{"local6", LOG_LOCAL6, false},
{"local7", LOG_LOCAL7, false},
Make GUC variables for syslog and SSL always visible Make the variables visible (but not used) even when support is not compiled in.
2010-12-18 16:53:59 +01:00
#else
{"none", 0, false},
#endif
Convert syslog_facility guc to enum type.
2008-04-03 11:21:15 +02:00
{NULL, 0}
};
Add support for tracking call counts and elapsed runtime for user-defined functions. Note that because this patch changes FmgrInfo, any external C functions you might be testing with 8.4 will need to be recompiled. Patch by Martin Pihlak, some editorialization by me (principally, removing tracking of getrusage() numbers)
2008-05-15 02:17:41 +02:00
static const struct config_enum_entry track_function_options[] = {
Add a field to guc enums to allow hiding of values from display while still accepting them as input, used to allow alternate syntax for the same setting. Alex Hunsaker
2008-05-28 11:04:06 +02:00
{"none", TRACK_FUNC_OFF, false},
{"pl", TRACK_FUNC_PL, false},
{"all", TRACK_FUNC_ALL, false},
{NULL, 0, false}
Add support for tracking call counts and elapsed runtime for user-defined functions. Note that because this patch changes FmgrInfo, any external C functions you might be testing with 8.4 will need to be recompiled. Patch by Martin Pihlak, some editorialization by me (principally, removing tracking of getrusage() numbers)
2008-05-15 02:17:41 +02:00
};
Turn xmlbinary and xmloption GUC variables into enumsTurn xmlbinary and xmloption GUC variables into enums..
2008-04-04 10:33:15 +02:00
static const struct config_enum_entry xmlbinary_options[] = {
Add a field to guc enums to allow hiding of values from display while still accepting them as input, used to allow alternate syntax for the same setting. Alex Hunsaker
2008-05-28 11:04:06 +02:00
{"base64", XMLBINARY_BASE64, false},
{"hex", XMLBINARY_HEX, false},
{NULL, 0, false}
Turn xmlbinary and xmloption GUC variables into enumsTurn xmlbinary and xmloption GUC variables into enums..
2008-04-04 10:33:15 +02:00
};
static const struct config_enum_entry xmloption_options[] = {
Add a field to guc enums to allow hiding of values from display while still accepting them as input, used to allow alternate syntax for the same setting. Alex Hunsaker
2008-05-28 11:04:06 +02:00
{"content", XMLOPTION_CONTENT, false},
{"document", XMLOPTION_DOCUMENT, false},
{NULL, 0, false}
Turn xmlbinary and xmloption GUC variables into enumsTurn xmlbinary and xmloption GUC variables into enums..
2008-04-04 10:33:15 +02:00
};
Convert three more guc settings to enum type: default_transaction_isolation, session_replication_role and regex_flavor.
2008-04-02 16:42:56 +02:00
Convert backslash_quote guc to use enum.
2008-04-04 13:47:19 +02:00
/*
* Although only "on", "off", and "safe_encoding" are documented, we
* accept all the likely variants of "on" and "off".
*/
static const struct config_enum_entry backslash_quote_options[] = {
Add a field to guc enums to allow hiding of values from display while still accepting them as input, used to allow alternate syntax for the same setting. Alex Hunsaker
2008-05-28 11:04:06 +02:00
{"safe_encoding", BACKSLASH_QUOTE_SAFE_ENCODING, false},
{"on", BACKSLASH_QUOTE_ON, false},
{"off", BACKSLASH_QUOTE_OFF, false},
{"true", BACKSLASH_QUOTE_ON, true},
{"false", BACKSLASH_QUOTE_OFF, true},
{"yes", BACKSLASH_QUOTE_ON, true},
{"no", BACKSLASH_QUOTE_OFF, true},
{"1", BACKSLASH_QUOTE_ON, true},
{"0", BACKSLASH_QUOTE_OFF, true},
{NULL, 0, false}
Convert backslash_quote guc to use enum.
2008-04-04 13:47:19 +02:00
};
Create a third option named "partition" for constraint_exclusion, and make it the default. This setting enables constraint exclusion checks only for appendrel members (ie, inheritance children and UNION ALL arms), which are the cases in which constraint exclusion is most likely to be useful. Avoiding the overhead for simple queries that are unlikely to benefit should bring the cost down to the point where this is a reasonable default setting. Per today's discussion.
2009-01-07 23:40:49 +01:00
/*
* Although only "on", "off", and "partition" are documented, we
* accept all the likely variants of "on" and "off".
*/
static const struct config_enum_entry constraint_exclusion_options[] = {
{"partition", CONSTRAINT_EXCLUSION_PARTITION, false},
{"on", CONSTRAINT_EXCLUSION_ON, false},
{"off", CONSTRAINT_EXCLUSION_OFF, false},
{"true", CONSTRAINT_EXCLUSION_ON, true},
{"false", CONSTRAINT_EXCLUSION_OFF, true},
{"yes", CONSTRAINT_EXCLUSION_ON, true},
{"no", CONSTRAINT_EXCLUSION_OFF, true},
{"1", CONSTRAINT_EXCLUSION_ON, true},
{"0", CONSTRAINT_EXCLUSION_OFF, true},
{NULL, 0, false}
};
Merge synchronous_replication setting into synchronous_commit. This means one less thing to configure when setting up synchronous replication, and also avoids some ambiguity around what the behavior should be when the settings of these variables conflict. Fujii Masao, with additional hacking by me.
2011-04-04 22:13:01 +02:00
/*
* Although only "on", "off", and "local" are documented, we
* accept all the likely variants of "on" and "off".
*/
static const struct config_enum_entry synchronous_commit_options[] = {
Avoid assuming there will be only 3 states for synchronous_commit. Also avoid hardcoding the current default state by giving it the name "on" and replace with a meaningful name that reflects its behaviour. Coding only, no change in behaviour.
2011-04-05 00:23:13 +02:00
{"local", SYNCHRONOUS_COMMIT_LOCAL_FLUSH, false},
Merge synchronous_replication setting into synchronous_commit. This means one less thing to configure when setting up synchronous replication, and also avoids some ambiguity around what the behavior should be when the settings of these variables conflict. Fujii Masao, with additional hacking by me.
2011-04-04 22:13:01 +02:00
{"on", SYNCHRONOUS_COMMIT_ON, false},
{"off", SYNCHRONOUS_COMMIT_OFF, false},
{"true", SYNCHRONOUS_COMMIT_ON, true},
{"false", SYNCHRONOUS_COMMIT_OFF, true},
{"yes", SYNCHRONOUS_COMMIT_ON, true},
{"no", SYNCHRONOUS_COMMIT_OFF, true},
{"1", SYNCHRONOUS_COMMIT_ON, true},
{"0", SYNCHRONOUS_COMMIT_OFF, true},
{NULL, 0, false}
};
Convert wal_sync_method to guc enum.
2008-05-12 10:35:05 +02:00
/*
* Options for enum values stored in other modules
*/
Introduce wal_level GUC to explicitly control if information needed for archival or hot standby should be WAL-logged, instead of deducing that from other options like archive_mode. This replaces recovery_connections GUC in the primary, where it now has no effect, but it's still used in the standby to enable/disable hot standby. Remove the WAL-logging of "unlogged operations", like creating an index without WAL-logging and fsyncing it at the end. Instead, we keep a copy of the wal_mode setting and the settings that affect how much shared memory a hot standby server needs to track master transactions (max_connections, max_prepared_xacts, max_locks_per_xact) in pg_control. Whenever the settings change, at server restart, write a WAL record noting the new settings and update pg_control. This allows us to notice the change in those settings in the standby at the right moment, they used to be included in checkpoint records, but that meant that a changed value was not reflected in the standby until the first checkpoint after the change. Bump PG_CONTROL_VERSION and XLOG_PAGE_MAGIC. Whack XLOG_PAGE_MAGIC back to the sequence it used to follow, before hot standby and subsequent patches changed it to 0x9003.
2010-04-28 18:10:43 +02:00
extern const struct config_enum_entry wal_level_options[];
Convert wal_sync_method to guc enum.
2008-05-12 10:35:05 +02:00
extern const struct config_enum_entry sync_method_options[];
Whack some sense into the configuration-file-location patch. Refactor code into something reasonably understandable, cause use of the feature to not fail in standalone backends or in EXEC_BACKEND case, fix sloppy guc.c table entries, make the documentation minimally usable.
2004-10-08 03:36:36 +02:00
/*
* GUC option variables that are exported from this module
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
*/
Fix the assert_enabled issue properly. This eliminates the former ABI difference between USE_ASSERT_CHECKING and not: the assert_enabled variable is always there.
2006-01-08 22:24:37 +01:00
#ifdef USE_ASSERT_CHECKING
pgindent run. Make it all clean.
2001-03-22 05:01:46 +01:00
bool assert_enabled = true;
Fix the assert_enabled issue properly. This eliminates the former ABI difference between USE_ASSERT_CHECKING and not: the assert_enabled variable is always there.
2006-01-08 22:24:37 +01:00
#else
bool assert_enabled = false;
#endif
Rename: ! #show_parser_stats = false ! #show_planner_stats = false ! #show_executor_stats = false ! #show_statement_stats = false TO: ! #log_parser_stats = false ! #log_planner_stats = false ! #log_executor_stats = false ! #log_statement_stats = false
2002-11-15 01:47:22 +01:00
bool log_duration = false;
pgindent run. Make it all clean.
2001-03-22 05:01:46 +01:00
bool Debug_print_plan = false;
bool Debug_print_parse = false;
bool Debug_print_rewritten = false;
Cause the output from debug_print_parse, debug_print_rewritten, and debug_print_plan to appear at LOG message level, not DEBUG1 as historically. Make debug_pretty_print default to on. Also, cause plans generated via EXPLAIN to be subject to debug_print_plan. This is all to make debug_print_plan a reasonably comfortable substitute for the former behavior of EXPLAIN VERBOSE.
2008-08-19 20:30:04 +02:00
bool Debug_pretty_print = true;
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
Rename: ! #show_parser_stats = false ! #show_planner_stats = false ! #show_executor_stats = false ! #show_statement_stats = false TO: ! #log_parser_stats = false ! #log_planner_stats = false ! #log_executor_stats = false ! #log_statement_stats = false
2002-11-15 01:47:22 +01:00
bool log_parser_stats = false;
bool log_planner_stats = false;
bool log_executor_stats = false;
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
bool log_statement_stats = false; /* this is sort of all three
* above together */
Rename show_btree_build_stats to log_btree_build_stats
2002-11-15 02:26:09 +01:00
bool log_btree_build_stats = false;
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
Clean up local redeclarations of variables with DLLIMPORT, per report from Magnus that MSVC complains about this.
2006-10-19 20:32:48 +02:00
bool check_function_bodies = true;
bool default_with_oids = false;
pgindent run. Make it all clean.
2001-03-22 05:01:46 +01:00
bool SQL_inheritance = true;
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
Enable encrypted passwords by default.
2002-08-15 00:07:56 +02:00
bool Password_encryption = true;
Use MD5 for wire protocol encryption for >= 7.2 client/server. Allow pg_shadow to be MD5 encrypted. Add ENCRYPTED/UNENCRYPTED option to CREATE/ALTER user. Add password_encryption postgresql.conf option. Update wire protocol version to 2.1.
2001-08-15 20:42:16 +02:00
Change the default setting for log_min_error_statement to ERROR. Per recent discussion in which majority opinion was that this is a more widely useful setting than the previous default of PANIC.
2006-11-21 02:23:37 +01:00
int log_min_error_statement = ERROR;
Increase the default value of log_min_messages to WARNING, so that NOTICE-grade messages are not logged by default. Per pgsql-hackers discussion back on 21-Nov-2007.
2008-03-10 04:22:29 +01:00
int log_min_messages = WARNING;
server_min_messages and friends should have valid values at bootup.
2002-09-22 21:52:38 +02:00
int client_min_messages = NOTICE;
Have log_min_duration_statement = 0 always print duration/statement. Change log line to be "duration: ms query:" Indent multi-line queries with a tab in the server logs.
2003-10-08 05:49:38 +02:00
int log_min_duration_statement = -1;
Add GUC log_temp_files to log the use of temporary files. Bill Moran
2007-01-09 22:31:17 +01:00
int log_temp_files = -1;
Clean up some randomness associated with trace_recovery_messages: don't put the variable declaration in the middle of a bunch of externs, and do use extern where it should be used.
2010-06-17 19:44:40 +02:00
int trace_recovery_messages = LOG;
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
Add temp_file_limit GUC parameter to constrain temporary file space usage. The limit is enforced against the total amount of temp file space used by each session. Mark Kirkwood, reviewed by Cédric Villemain and Tatsuo Ishii
2011-07-17 20:19:31 +02:00
int temp_file_limit = -1;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
int num_temp_buffers = 1024;
Add temp_buffers GUC variable to allow users to determine the size of the local buffer arena for temporary table access.
2005-03-20 00:27:11 +01:00
Use GUC lexer for recovery.conf parsing. This eliminates some crufty, special-purpose code and, as a non-trivial side benefit, allows recovery.conf parameters to be unquoted. Dimitri Fontaine, with review and cleanup by Alvaro Herrera, Itagaki Takahiro, and me.
2010-12-03 14:44:15 +01:00
char *data_directory;
Adjust configuration-files GUC behavior as per my recent proposal. The vars are renamed to data_directory, config_file, hba_file, and ident_file, and are guaranteed to be set to accurate absolute paths during postmaster startup. This commit does not yet do anything about hiding path values from non-superusers.
2004-10-10 01:13:22 +02:00
char *ConfigFileName;
char *HbaFileName;
char *IdentFileName;
char *external_pid_file;
Whack some sense into the configuration-file-location patch. Refactor code into something reasonably understandable, cause use of the feature to not fail in standalone backends or in EXEC_BACKEND case, fix sloppy guc.c table entries, make the documentation minimally usable.
2004-10-08 03:36:36 +02:00
Make the temporary directory for pgstat files configurable by the GUC variable stats_temp_directory, instead of requiring the admin to mount/symlink the pg_stat_tmp directory manually. For now the config variable is PGC_POSTMASTER. Room for further improvment that would allow it to be changed on-the-fly.
2008-08-15 10:37:41 +02:00
char *pgstat_temp_directory;
Add support for an application_name parameter, which is displayed in pg_stat_activity and recorded in log entries. Dave Page, reviewed by Andres Freund
2009-11-29 00:38:08 +01:00
char *application_name;
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
int tcp_keepalives_idle;
int tcp_keepalives_interval;
int tcp_keepalives_count;
Add GUC variable to print original query to the server logs when there is an error, warning, etc. Gavin Sherry
2002-09-02 07:42:54 +02:00
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
/*
* These variables are all dummies that don't do anything, except in some
* cases provide the value for SHOW to display. The real state is elsewhere
* and is kept in sync by assign_hooks.
*/
Remove 'syslog' GUC variable, and add more logical 'log_destination' variable to control logoutput location on Unix and Win32. Magnus Hagander
2004-04-05 05:02:11 +02:00
static char *log_destination_string;
pgindent run for 8.2.
2006-10-04 02:30:14 +02:00
Fix syslog bug: if any messages are emitted to write_syslog before the facility has been set, the facility gets set to LOCAL0 and cannot be changed later. This seems reasonably plausible to happen, particularly at higher debug log levels, though I am not certain it explains Han Holl's recent report. Easiest fix is to teach the code how to change the value on-the-fly, which is nicer anyway. I made the settings PGC_SIGHUP to conform with log_destination.
2005-10-14 22:53:56 +02:00
static char *syslog_ident_str;
Backend support for autocommit removed, per recent discussions. The only remnant of this failed experiment is that the server will take SET AUTOCOMMIT TO ON. Still TODO: provide some client-side autocommit logic in libpq.
2003-05-14 05:26:03 +02:00
static bool phony_autocommit;
Add is_superuser parameter reporting, soon to be used by psql.
2003-06-27 21:08:38 +02:00
static bool session_auth_is_superuser;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
static double phony_random_seed;
static char *client_encoding_string;
static char *datestyle_string;
In the continuing saga of FE/BE protocol revisions, add reporting of initial values and runtime changes in selected parameters. This gets rid of the need for an initial 'select pg_client_encoding()' query in libpq, bringing us back to one message transmitted in each direction for a standard connection startup. To allow server version to be sent using the same GUC mechanism that handles other parameters, invent the concept of a never-settable GUC parameter: you can 'show server_version' but it's not settable by any GUC input source. Create 'lc_collate' and 'lc_ctype' never-settable parameters so that people can find out these settings without need for pg_controldata. (These side ideas were all discussed some time ago in pgsql-hackers, but not yet implemented.)
2003-04-25 21:45:10 +02:00
static char *locale_collate;
static char *locale_ctype;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
static char *server_encoding_string;
In the continuing saga of FE/BE protocol revisions, add reporting of initial values and runtime changes in selected parameters. This gets rid of the need for an initial 'select pg_client_encoding()' query in libpq, bringing us back to one message transmitted in each direction for a standard connection startup. To allow server version to be sent using the same GUC mechanism that handles other parameters, invent the concept of a never-settable GUC parameter: you can 'show server_version' but it's not settable by any GUC input source. Create 'lc_collate' and 'lc_ctype' never-settable parameters so that people can find out these settings without need for pg_controldata. (These side ideas were all discussed some time ago in pgsql-hackers, but not yet implemented.)
2003-04-25 21:45:10 +02:00
static char *server_version_string;
pgindent run for 8.2.
2006-10-04 02:30:14 +02:00
static int server_version_num;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
static char *timezone_string;
Switch over to using the src/timezone functions for formatting timestamps displayed in the postmaster log. This avoids Windows-specific problems with localized time zone names that are in the wrong encoding, and generally seems like a good idea to forestall other potential platform-dependent issues. To preserve the existing behavior that all backends will log in the same time zone, create a new GUC variable log_timezone that can only be changed on a system-wide basis, and reference log-related calculations to that zone instead of the TimeZone variable. This fixes the issue reported by Hiroshi Saito that timestamps printed by xlog.c startup could be improperly localized on Windows. We still need a simpler patch for that problem in the back branches, however.
2007-08-04 03:26:54 +02:00
static char *log_timezone_string;
Adjust initialization sequence for timezone_abbreviations so that it's handled just about like timezone; in particular, don't try to read anything during InitializeGUCOptions. Should solve current startup failure on Windows, and avoid wasted cycles if a nondefault setting is specified in postgresql.conf too. Possibly we need to think about a more general solution for handling 'expensive to set' GUC options.
2006-07-29 05:02:56 +02:00
static char *timezone_abbreviations_string;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
static char *XactIsoLevel_string;
static char *session_authorization_string;
Added new group of read-only GUC variables to allow simple access to certain compile-time options (FUNC_MAX_ARGS, INDEX_MAX_KEYS, NAMEDATALEN, BLCKSZ, HAVE_INT64_TIMESTAMP). Also added "category", "short_desc", and "extra_desc" to the pg_settings view. Per recent discussion here: http://archives.postgresql.org/pgsql-patches/2003-11/msg00363.php
2003-12-03 19:52:00 +01:00
static int max_function_args;
static int max_index_keys;
static int max_identifier_length;
static int block_size;
Add unchangeable GUC "variables" segment_size, wal_block_size, and wal_segment_size to make those configuration parameters available to clients, in the same way that block_size was previously exposed. Bernd Helmle, with comments from Abhijit Menon-Sen and some further tweaking by me.
2008-07-11 00:08:17 +02:00
static int segment_size;
static int wal_block_size;
static int wal_segment_size;
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
static bool integer_datetimes;
Implement prefetching via posix_fadvise() for bitmap index scans. A new GUC variable effective_io_concurrency controls how many concurrent block prefetch requests will be issued. (The best way to handle this for plain index scans is still under debate, so that part is not applied yet --- tgl) Greg Stark
2009-01-12 06:10:45 +01:00
static int effective_io_concurrency;
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
/* should be static, but commands/variable.c needs to get at this */
Add SET ROLE. This is a partial commit of Stephen Frost's recent patch; I'm still working on the has_role function and information_schema changes.
2005-07-26 00:12:34 +02:00
char *role_string;
Make the location of the Kerberos server key file run time configurable (rather than compile time). For libpq, even when Kerberos support is compiled in, the default user name should still fall back to geteuid() if it can't be determined via the Kerberos system. A couple of fixes for string type configuration parameters, now that there is one.
2000-08-25 12:00:35 +02:00
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
/*
Code review for recent GUC changes --- try to make it less obvious that these things were added at different times by different people ;-). Includes Aizaz Ahmed's patch to remove duplicate array in help_config.c.
2003-07-28 21:31:32 +02:00
* Displayable names for context types (enum GucContext)
*
* Note: these strings are deliberately not localized.
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
*/
pgindent run.
2003-08-04 02:43:34 +02:00
const char *const GucContext_Names[] =
Code review for recent GUC changes --- try to make it less obvious that these things were added at different times by different people ;-). Includes Aizaz Ahmed's patch to remove duplicate array in help_config.c.
2003-07-28 21:31:32 +02:00
{
pgindent run.
2003-08-04 02:43:34 +02:00
/* PGC_INTERNAL */ "internal",
/* PGC_POSTMASTER */ "postmaster",
/* PGC_SIGHUP */ "sighup",
/* PGC_BACKEND */ "backend",
/* PGC_SUSET */ "superuser",
/* PGC_USERSET */ "user"
Code review for recent GUC changes --- try to make it less obvious that these things were added at different times by different people ;-). Includes Aizaz Ahmed's patch to remove duplicate array in help_config.c.
2003-07-28 21:31:32 +02:00
};
/*
* Displayable names for source types (enum GucSource)
*
* Note: these strings are deliberately not localized.
pgindent run.
2003-08-04 02:43:34 +02:00
*/
const char *const GucSource_Names[] =
Code review for recent GUC changes --- try to make it less obvious that these things were added at different times by different people ;-). Includes Aizaz Ahmed's patch to remove duplicate array in help_config.c.
2003-07-28 21:31:32 +02:00
{
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
/* PGC_S_DEFAULT */ "default",
Split PGC_S_DEFAULT into two values, for true boot_val vs computed default. Failure to distinguish these cases is the real cause behind the recent reports of Windows builds crashing on 'infinity'::timestamp, which was directly due to failure to establish a value of timezone_abbreviations in postmaster child processes. The postmaster had the desired value, but write_one_nondefault_variable() didn't transmit it to backends. To fix that, invent a new value PGC_S_DYNAMIC_DEFAULT, and be sure to use that or PGC_S_ENV_VAR (as appropriate) for "default" settings that are computed during initialization. (We need both because there's at least one variable that could receive a value from either source.) This commit also fixes ProcessConfigFile's failure to restore the correct default value for certain GUC variables if they are set in postgresql.conf and then removed/commented out of the file. We have to recompute and reinstall the value for any GUC variable that could have received a value from PGC_S_DYNAMIC_DEFAULT or PGC_S_ENV_VAR sources, and there were a number of oversights. (That whole thing is a crock that needs to be redesigned, but not today.) However, I intentionally didn't make it work "exactly right" for the cases of timezone and log_timezone. The exactly right behavior would involve running select_default_timezone, which we'd have to do independently in each postgres process, causing the whole database to become entirely unresponsive for as much as several seconds. That didn't seem like a good idea, especially since the variable's removal from postgresql.conf might be just an accidental edit. Instead the behavior is to adopt the previously active setting as if it were default. Note that this patch creates an ABI break for extensions that use any of the PGC_S_XXX constants; they'll need to be recompiled.
2011-05-12 01:57:38 +02:00
/* PGC_S_DYNAMIC_DEFAULT */ "default",
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
/* PGC_S_ENV_VAR */ "environment variable",
/* PGC_S_FILE */ "configuration file",
/* PGC_S_ARGV */ "command line",
/* PGC_S_DATABASE */ "database",
/* PGC_S_USER */ "user",
Add new PGC_S_DATABASE_USER enum value to several places missed by my patch last week. Per note and patch from Jeff Davis.
2009-10-13 16:18:40 +02:00
/* PGC_S_DATABASE_USER */ "database user",
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
/* PGC_S_CLIENT */ "client",
/* PGC_S_OVERRIDE */ "override",
/* PGC_S_INTERACTIVE */ "interactive",
/* PGC_S_TEST */ "test",
/* PGC_S_SESSION */ "session"
Code review for recent GUC changes --- try to make it less obvious that these things were added at different times by different people ;-). Includes Aizaz Ahmed's patch to remove duplicate array in help_config.c.
2003-07-28 21:31:32 +02:00
};
/*
* Displayable names for the groupings defined in enum config_group
*/
const char *const config_group_names[] =
{
/* UNGROUPED */
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
gettext_noop("Ungrouped"),
Whack some sense into the configuration-file-location patch. Refactor code into something reasonably understandable, cause use of the feature to not fail in standalone backends or in EXEC_BACKEND case, fix sloppy guc.c table entries, make the documentation minimally usable.
2004-10-08 03:36:36 +02:00
/* FILE_LOCATIONS */
gettext_noop("File Locations"),
Code review for recent GUC changes --- try to make it less obvious that these things were added at different times by different people ;-). Includes Aizaz Ahmed's patch to remove duplicate array in help_config.c.
2003-07-28 21:31:32 +02:00
/* CONN_AUTH */
Message editing: remove gratuitous variations in message wording, standardize terms, add some clarifications, fix some untranslatable attempts at dynamic message building.
2003-09-25 08:58:07 +02:00
gettext_noop("Connections and Authentication"),
Code review for recent GUC changes --- try to make it less obvious that these things were added at different times by different people ;-). Includes Aizaz Ahmed's patch to remove duplicate array in help_config.c.
2003-07-28 21:31:32 +02:00
/* CONN_AUTH_SETTINGS */
Message editing: remove gratuitous variations in message wording, standardize terms, add some clarifications, fix some untranslatable attempts at dynamic message building.
2003-09-25 08:58:07 +02:00
gettext_noop("Connections and Authentication / Connection Settings"),
Code review for recent GUC changes --- try to make it less obvious that these things were added at different times by different people ;-). Includes Aizaz Ahmed's patch to remove duplicate array in help_config.c.
2003-07-28 21:31:32 +02:00
/* CONN_AUTH_SECURITY */
Message editing: remove gratuitous variations in message wording, standardize terms, add some clarifications, fix some untranslatable attempts at dynamic message building.
2003-09-25 08:58:07 +02:00
gettext_noop("Connections and Authentication / Security and Authentication"),
Code review for recent GUC changes --- try to make it less obvious that these things were added at different times by different people ;-). Includes Aizaz Ahmed's patch to remove duplicate array in help_config.c.
2003-07-28 21:31:32 +02:00
/* RESOURCES */
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
gettext_noop("Resource Usage"),
Code review for recent GUC changes --- try to make it less obvious that these things were added at different times by different people ;-). Includes Aizaz Ahmed's patch to remove duplicate array in help_config.c.
2003-07-28 21:31:32 +02:00
/* RESOURCES_MEM */
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
gettext_noop("Resource Usage / Memory"),
Add temp_file_limit GUC parameter to constrain temporary file space usage. The limit is enforced against the total amount of temp file space used by each session. Mark Kirkwood, reviewed by Cédric Villemain and Tatsuo Ishii
2011-07-17 20:19:31 +02:00
/* RESOURCES_DISK */
gettext_noop("Resource Usage / Disk"),
Code review for recent GUC changes --- try to make it less obvious that these things were added at different times by different people ;-). Includes Aizaz Ahmed's patch to remove duplicate array in help_config.c.
2003-07-28 21:31:32 +02:00
/* RESOURCES_KERNEL */
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
gettext_noop("Resource Usage / Kernel Resources"),
Add new GUC categories corresponding to sections in docs, and move description for vacuum_defer_cleanup_age to the correct category. Sections in postgresql.conf are also sorted in the same order with docs. Per gripe by Fujii Masao, suggestion by Heikki Linnakangas, and patch by me.
2010-06-15 09:52:11 +02:00
/* RESOURCES_VACUUM_DELAY */
gettext_noop("Resource Usage / Cost-Based Vacuum Delay"),
/* RESOURCES_BGWRITER */
gettext_noop("Resource Usage / Background Writer"),
/* RESOURCES_ASYNCHRONOUS */
gettext_noop("Resource Usage / Asynchronous Behavior"),
Code review for recent GUC changes --- try to make it less obvious that these things were added at different times by different people ;-). Includes Aizaz Ahmed's patch to remove duplicate array in help_config.c.
2003-07-28 21:31:32 +02:00
/* WAL */
Message editing: remove gratuitous variations in message wording, standardize terms, add some clarifications, fix some untranslatable attempts at dynamic message building.
2003-09-25 08:58:07 +02:00
gettext_noop("Write-Ahead Log"),
Code review for recent GUC changes --- try to make it less obvious that these things were added at different times by different people ;-). Includes Aizaz Ahmed's patch to remove duplicate array in help_config.c.
2003-07-28 21:31:32 +02:00
/* WAL_SETTINGS */
Message editing: remove gratuitous variations in message wording, standardize terms, add some clarifications, fix some untranslatable attempts at dynamic message building.
2003-09-25 08:58:07 +02:00
gettext_noop("Write-Ahead Log / Settings"),
Code review for recent GUC changes --- try to make it less obvious that these things were added at different times by different people ;-). Includes Aizaz Ahmed's patch to remove duplicate array in help_config.c.
2003-07-28 21:31:32 +02:00
/* WAL_CHECKPOINTS */
Message editing: remove gratuitous variations in message wording, standardize terms, add some clarifications, fix some untranslatable attempts at dynamic message building.
2003-09-25 08:58:07 +02:00
gettext_noop("Write-Ahead Log / Checkpoints"),
Add new GUC categories corresponding to sections in docs, and move description for vacuum_defer_cleanup_age to the correct category. Sections in postgresql.conf are also sorted in the same order with docs. Per gripe by Fujii Masao, suggestion by Heikki Linnakangas, and patch by me.
2010-06-15 09:52:11 +02:00
/* WAL_ARCHIVING */
gettext_noop("Write-Ahead Log / Archiving"),
Reclassify replication-related GUC variables as "master" and "standby". Per discussion, this structure seems more understandable than what was there before. Make config.sgml and postgresql.conf.sample agree. In passing do a bit of editorial work on the variable descriptions.
2011-07-07 21:10:32 +02:00
/* REPLICATION */
gettext_noop("Replication"),
Introduce sending servers as new category for replication params Fujii Masao
2011-07-19 09:59:55 +02:00
/* REPLICATION_SENDING */
gettext_noop("Replication / Sending Servers"),
Reclassify replication-related GUC variables as "master" and "standby". Per discussion, this structure seems more understandable than what was there before. Make config.sgml and postgresql.conf.sample agree. In passing do a bit of editorial work on the variable descriptions.
2011-07-07 21:10:32 +02:00
/* REPLICATION_MASTER */
gettext_noop("Replication / Master Server"),
/* REPLICATION_STANDBY */
gettext_noop("Replication / Standby Servers"),
Code review for recent GUC changes --- try to make it less obvious that these things were added at different times by different people ;-). Includes Aizaz Ahmed's patch to remove duplicate array in help_config.c.
2003-07-28 21:31:32 +02:00
/* QUERY_TUNING */
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
gettext_noop("Query Tuning"),
Code review for recent GUC changes --- try to make it less obvious that these things were added at different times by different people ;-). Includes Aizaz Ahmed's patch to remove duplicate array in help_config.c.
2003-07-28 21:31:32 +02:00
/* QUERY_TUNING_METHOD */
This patch adds a new GUC var, "default_with_oids", which follows the proposal for eventually deprecating OIDs on user tables that I posted earlier to pgsql-hackers. pg_dump now always specifies WITH OIDS or WITHOUT OIDS when dumping a table. The documentation has been updated. Neil Conway
2003-12-01 23:08:02 +01:00
gettext_noop("Query Tuning / Planner Method Configuration"),
Code review for recent GUC changes --- try to make it less obvious that these things were added at different times by different people ;-). Includes Aizaz Ahmed's patch to remove duplicate array in help_config.c.
2003-07-28 21:31:32 +02:00
/* QUERY_TUNING_COST */
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
gettext_noop("Query Tuning / Planner Cost Constants"),
Code review for recent GUC changes --- try to make it less obvious that these things were added at different times by different people ;-). Includes Aizaz Ahmed's patch to remove duplicate array in help_config.c.
2003-07-28 21:31:32 +02:00
/* QUERY_TUNING_GEQO */
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
gettext_noop("Query Tuning / Genetic Query Optimizer"),
Code review for recent GUC changes --- try to make it less obvious that these things were added at different times by different people ;-). Includes Aizaz Ahmed's patch to remove duplicate array in help_config.c.
2003-07-28 21:31:32 +02:00
/* QUERY_TUNING_OTHER */
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
gettext_noop("Query Tuning / Other Planner Options"),
Code review for recent GUC changes --- try to make it less obvious that these things were added at different times by different people ;-). Includes Aizaz Ahmed's patch to remove duplicate array in help_config.c.
2003-07-28 21:31:32 +02:00
/* LOGGING */
Message editing: remove gratuitous variations in message wording, standardize terms, add some clarifications, fix some untranslatable attempts at dynamic message building.
2003-09-25 08:58:07 +02:00
gettext_noop("Reporting and Logging"),
Remove 'syslog' GUC variable, and add more logical 'log_destination' variable to control logoutput location on Unix and Win32. Magnus Hagander
2004-04-05 05:02:11 +02:00
/* LOGGING_WHERE */
gettext_noop("Reporting and Logging / Where to Log"),
Code review for recent GUC changes --- try to make it less obvious that these things were added at different times by different people ;-). Includes Aizaz Ahmed's patch to remove duplicate array in help_config.c.
2003-07-28 21:31:32 +02:00
/* LOGGING_WHEN */
Message editing: remove gratuitous variations in message wording, standardize terms, add some clarifications, fix some untranslatable attempts at dynamic message building.
2003-09-25 08:58:07 +02:00
gettext_noop("Reporting and Logging / When to Log"),
Code review for recent GUC changes --- try to make it less obvious that these things were added at different times by different people ;-). Includes Aizaz Ahmed's patch to remove duplicate array in help_config.c.
2003-07-28 21:31:32 +02:00
/* LOGGING_WHAT */
Message editing: remove gratuitous variations in message wording, standardize terms, add some clarifications, fix some untranslatable attempts at dynamic message building.
2003-09-25 08:58:07 +02:00
gettext_noop("Reporting and Logging / What to Log"),
Code review for recent GUC changes --- try to make it less obvious that these things were added at different times by different people ;-). Includes Aizaz Ahmed's patch to remove duplicate array in help_config.c.
2003-07-28 21:31:32 +02:00
/* STATS */
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
gettext_noop("Statistics"),
Code review for recent GUC changes --- try to make it less obvious that these things were added at different times by different people ;-). Includes Aizaz Ahmed's patch to remove duplicate array in help_config.c.
2003-07-28 21:31:32 +02:00
/* STATS_MONITORING */
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
gettext_noop("Statistics / Monitoring"),
Code review for recent GUC changes --- try to make it less obvious that these things were added at different times by different people ;-). Includes Aizaz Ahmed's patch to remove duplicate array in help_config.c.
2003-07-28 21:31:32 +02:00
/* STATS_COLLECTOR */
Message editing: remove gratuitous variations in message wording, standardize terms, add some clarifications, fix some untranslatable attempts at dynamic message building.
2003-09-25 08:58:07 +02:00
gettext_noop("Statistics / Query and Index Statistics Collector"),
Integrate autovacuum functionality into the backend. There's still a few loose ends to be dealt with, but it seems to work. Alvaro Herrera, based on the contrib code by Matthew O'Connor.
2005-07-14 07:13:45 +02:00
/* AUTOVACUUM */
Make all command-line options of postmaster and postgres the same. See http://archives.postgresql.org/pgsql-hackers/2006-01/msg00151.php for the complete plan.
2006-01-05 11:07:46 +01:00
gettext_noop("Autovacuum"),
Code review for recent GUC changes --- try to make it less obvious that these things were added at different times by different people ;-). Includes Aizaz Ahmed's patch to remove duplicate array in help_config.c.
2003-07-28 21:31:32 +02:00
/* CLIENT_CONN */
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
gettext_noop("Client Connection Defaults"),
Code review for recent GUC changes --- try to make it less obvious that these things were added at different times by different people ;-). Includes Aizaz Ahmed's patch to remove duplicate array in help_config.c.
2003-07-28 21:31:32 +02:00
/* CLIENT_CONN_STATEMENT */
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
gettext_noop("Client Connection Defaults / Statement Behavior"),
Code review for recent GUC changes --- try to make it less obvious that these things were added at different times by different people ;-). Includes Aizaz Ahmed's patch to remove duplicate array in help_config.c.
2003-07-28 21:31:32 +02:00
/* CLIENT_CONN_LOCALE */
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
gettext_noop("Client Connection Defaults / Locale and Formatting"),
Code review for recent GUC changes --- try to make it less obvious that these things were added at different times by different people ;-). Includes Aizaz Ahmed's patch to remove duplicate array in help_config.c.
2003-07-28 21:31:32 +02:00
/* CLIENT_CONN_OTHER */
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
gettext_noop("Client Connection Defaults / Other Defaults"),
Code review for recent GUC changes --- try to make it less obvious that these things were added at different times by different people ;-). Includes Aizaz Ahmed's patch to remove duplicate array in help_config.c.
2003-07-28 21:31:32 +02:00
/* LOCK_MANAGEMENT */
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
gettext_noop("Lock Management"),
Code review for recent GUC changes --- try to make it less obvious that these things were added at different times by different people ;-). Includes Aizaz Ahmed's patch to remove duplicate array in help_config.c.
2003-07-28 21:31:32 +02:00
/* COMPAT_OPTIONS */
Message editing: remove gratuitous variations in message wording, standardize terms, add some clarifications, fix some untranslatable attempts at dynamic message building.
2003-09-25 08:58:07 +02:00
gettext_noop("Version and Platform Compatibility"),
Code review for recent GUC changes --- try to make it less obvious that these things were added at different times by different people ;-). Includes Aizaz Ahmed's patch to remove duplicate array in help_config.c.
2003-07-28 21:31:32 +02:00
/* COMPAT_OPTIONS_PREVIOUS */
Message editing: remove gratuitous variations in message wording, standardize terms, add some clarifications, fix some untranslatable attempts at dynamic message building.
2003-09-25 08:58:07 +02:00
gettext_noop("Version and Platform Compatibility / Previous PostgreSQL Versions"),
Code review for recent GUC changes --- try to make it less obvious that these things were added at different times by different people ;-). Includes Aizaz Ahmed's patch to remove duplicate array in help_config.c.
2003-07-28 21:31:32 +02:00
/* COMPAT_OPTIONS_CLIENT */
Message editing: remove gratuitous variations in message wording, standardize terms, add some clarifications, fix some untranslatable attempts at dynamic message building.
2003-09-25 08:58:07 +02:00
gettext_noop("Version and Platform Compatibility / Other Platforms and Clients"),
Add restart_after_crash GUC. Normally, we automatically restart after a backend crash, but in some cases when PostgreSQL is invoked by clusterware it may be desirable to suppress this behavior, so we provide an option which does this. Since no existing GUC group quite fits, create a new group called "error handling options" for this and the previously undocumented GUC exit_on_error, which is now documented. Review by Fujii Masao.
2010-07-20 02:47:53 +02:00
/* ERROR_HANDLING */
gettext_noop("Error Handling"),
Whack some sense into the configuration-file-location patch. Refactor code into something reasonably understandable, cause use of the feature to not fail in standalone backends or in EXEC_BACKEND case, fix sloppy guc.c table entries, make the documentation minimally usable.
2004-10-08 03:36:36 +02:00
/* PRESET_OPTIONS */
gettext_noop("Preset Options"),
/* CUSTOM_OPTIONS */
gettext_noop("Customized Options"),
Code review for recent GUC changes --- try to make it less obvious that these things were added at different times by different people ;-). Includes Aizaz Ahmed's patch to remove duplicate array in help_config.c.
2003-07-28 21:31:32 +02:00
/* DEVELOPER_OPTIONS */
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
gettext_noop("Developer Options"),
Code review for recent GUC changes --- try to make it less obvious that these things were added at different times by different people ;-). Includes Aizaz Ahmed's patch to remove duplicate array in help_config.c.
2003-07-28 21:31:32 +02:00
/* help_config wants this array to be null-terminated */
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL
};
Code review for recent GUC changes --- try to make it less obvious that these things were added at different times by different people ;-). Includes Aizaz Ahmed's patch to remove duplicate array in help_config.c.
2003-07-28 21:31:32 +02:00
/*
* Displayable names for GUC variable types (enum config_type)
*
* Note: these strings are deliberately not localized.
*/
pgindent run.
2003-08-04 02:43:34 +02:00
const char *const config_type_names[] =
Code review for recent GUC changes --- try to make it less obvious that these things were added at different times by different people ;-). Includes Aizaz Ahmed's patch to remove duplicate array in help_config.c.
2003-07-28 21:31:32 +02:00
{
pgindent run.
2003-08-04 02:43:34 +02:00
/* PGC_BOOL */ "bool",
/* PGC_INT */ "integer",
/* PGC_REAL */ "real",
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
/* PGC_STRING */ "string",
/* PGC_ENUM */ "enum"
Code review for recent GUC changes --- try to make it less obvious that these things were added at different times by different people ;-). Includes Aizaz Ahmed's patch to remove duplicate array in help_config.c.
2003-07-28 21:31:32 +02:00
};
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
Second pass over run-time configuration system. Adjust priorities on some option settings. Sort out SIGHUP vs BACKEND -- there is no total ordering here, so make explicit checks. Add comments explaining all of this. Removed permissions check on SHOW command. Add examine_subclass to the game, rename to SQL_inheritance to fit the official data model better. Adjust documentation. Standalone backend needs to reset all options before it starts. To facilitate that, have IsUnderPostmaster be set by the postmaster itself, don't wait for the magic -p switch. Also make sure that all environment variables and argv's survive init_ps_display(). Use strdup where necessary. Have initdb make configuration files (postgresql.conf, pg_hba.conf) mode 0600 -- having configuration files is no fun if you can't edit them.
2000-06-23 00:31:24 +02:00
/*
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
* Contents of GUC tables
*
* See src/backend/utils/misc/README for design notes.
*
Second pass over run-time configuration system. Adjust priorities on some option settings. Sort out SIGHUP vs BACKEND -- there is no total ordering here, so make explicit checks. Add comments explaining all of this. Removed permissions check on SHOW command. Add examine_subclass to the game, rename to SQL_inheritance to fit the official data model better. Adjust documentation. Standalone backend needs to reset all options before it starts. To facilitate that, have IsUnderPostmaster be set by the postmaster itself, don't wait for the magic -p switch. Also make sure that all environment variables and argv's survive init_ps_display(). Use strdup where necessary. Have initdb make configuration files (postgresql.conf, pg_hba.conf) mode 0600 -- having configuration files is no fun if you can't edit them.
2000-06-23 00:31:24 +02:00
* TO ADD AN OPTION:
*
* 1. Declare a global variable of type bool, int, double, or char*
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
* and make use of it.
Second pass over run-time configuration system. Adjust priorities on some option settings. Sort out SIGHUP vs BACKEND -- there is no total ordering here, so make explicit checks. Add comments explaining all of this. Removed permissions check on SHOW command. Add examine_subclass to the game, rename to SQL_inheritance to fit the official data model better. Adjust documentation. Standalone backend needs to reset all options before it starts. To facilitate that, have IsUnderPostmaster be set by the postmaster itself, don't wait for the magic -p switch. Also make sure that all environment variables and argv's survive init_ps_display(). Use strdup where necessary. Have initdb make configuration files (postgresql.conf, pg_hba.conf) mode 0600 -- having configuration files is no fun if you can't edit them.
2000-06-23 00:31:24 +02:00
*
* 2. Decide at what times it's safe to set the option. See guc.h for
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
* details.
Second pass over run-time configuration system. Adjust priorities on some option settings. Sort out SIGHUP vs BACKEND -- there is no total ordering here, so make explicit checks. Add comments explaining all of this. Removed permissions check on SHOW command. Add examine_subclass to the game, rename to SQL_inheritance to fit the official data model better. Adjust documentation. Standalone backend needs to reset all options before it starts. To facilitate that, have IsUnderPostmaster be set by the postmaster itself, don't wait for the magic -p switch. Also make sure that all environment variables and argv's survive init_ps_display(). Use strdup where necessary. Have initdb make configuration files (postgresql.conf, pg_hba.conf) mode 0600 -- having configuration files is no fun if you can't edit them.
2000-06-23 00:31:24 +02:00
*
* 3. Decide on a name, a default value, upper and lower bounds (if
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
* applicable), etc.
Second pass over run-time configuration system. Adjust priorities on some option settings. Sort out SIGHUP vs BACKEND -- there is no total ordering here, so make explicit checks. Add comments explaining all of this. Removed permissions check on SHOW command. Add examine_subclass to the game, rename to SQL_inheritance to fit the official data model better. Adjust documentation. Standalone backend needs to reset all options before it starts. To facilitate that, have IsUnderPostmaster be set by the postmaster itself, don't wait for the magic -p switch. Also make sure that all environment variables and argv's survive init_ps_display(). Use strdup where necessary. Have initdb make configuration files (postgresql.conf, pg_hba.conf) mode 0600 -- having configuration files is no fun if you can't edit them.
2000-06-23 00:31:24 +02:00
*
* 4. Add a record below.
*
Make the 'wal_debug' GUC variable a boolean (rather than an integer), and hide it behind #ifdef WAL_DEBUG blocks.
2004-01-06 18:26:23 +01:00
* 5. Add it to src/backend/utils/misc/postgresql.conf.sample, if
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
* appropriate.
Add all possible config file options.
2001-01-24 19:37:31 +01:00
*
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
* 6. Don't forget to document the option (at least in config.sgml).
Add all possible config file options.
2001-01-24 19:37:31 +01:00
*
Remove comment about modifying tab-complete.c for userset GUC. Simon Riggs
2007-06-18 12:02:57 +02:00
* 7. If it's a new GUC_LIST option you must edit pg_dumpall.c to ensure
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
* it is not single quoted at dump time.
Second pass over run-time configuration system. Adjust priorities on some option settings. Sort out SIGHUP vs BACKEND -- there is no total ordering here, so make explicit checks. Add comments explaining all of this. Removed permissions check on SHOW command. Add examine_subclass to the game, rename to SQL_inheritance to fit the official data model better. Adjust documentation. Standalone backend needs to reset all options before it starts. To facilitate that, have IsUnderPostmaster be set by the postmaster itself, don't wait for the magic -p switch. Also make sure that all environment variables and argv's survive init_ps_display(). Use strdup where necessary. Have initdb make configuration files (postgresql.conf, pg_hba.conf) mode 0600 -- having configuration files is no fun if you can't edit them.
2000-06-23 00:31:24 +02:00
*/
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
/******** option records follow ********/
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
static struct config_bool ConfigureNamesBool[] =
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
{
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"enable_seqscan", PGC_USERSET, QUERY_TUNING_METHOD,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Enables the planner's use of sequential-scan plans."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL
},
&enable_seqscan,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
true,
NULL, NULL, NULL
Another pgindent run. Fixes enum indenting, and improves #endif spacing. Also adds space for one-line comments.
2001-10-28 07:26:15 +01:00
},
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"enable_indexscan", PGC_USERSET, QUERY_TUNING_METHOD,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Enables the planner's use of index-scan plans."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL
},
&enable_indexscan,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
true,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
Rethink original decision to use AND/OR Expr nodes to represent bitmap logic operations during planning. Seems cleaner to create two new Path node types, instead --- this avoids duplication of cost-estimation code. Also, create an enable_bitmapscan GUC parameter to control use of bitmap plans.
2005-04-21 21:18:13 +02:00
{
{"enable_bitmapscan", PGC_USERSET, QUERY_TUNING_METHOD,
gettext_noop("Enables the planner's use of bitmap-scan plans."),
NULL
},
&enable_bitmapscan,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
true,
NULL, NULL, NULL
Rethink original decision to use AND/OR Expr nodes to represent bitmap logic operations during planning. Seems cleaner to create two new Path node types, instead --- this avoids duplication of cost-estimation code. Also, create an enable_bitmapscan GUC parameter to control use of bitmap plans.
2005-04-21 21:18:13 +02:00
},
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"enable_tidscan", PGC_USERSET, QUERY_TUNING_METHOD,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Enables the planner's use of TID scan plans."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL
},
&enable_tidscan,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
true,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"enable_sort", PGC_USERSET, QUERY_TUNING_METHOD,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Enables the planner's use of explicit sort steps."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL
},
&enable_sort,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
true,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
Finish implementation of hashed aggregation. Add enable_hashagg GUC parameter to allow it to be forced off for comparison purposes. Add ORDER BY clauses to a bunch of regression test queries that will otherwise produce randomly-ordered output in the new regime.
2002-11-21 01:42:20 +01:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"enable_hashagg", PGC_USERSET, QUERY_TUNING_METHOD,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Enables the planner's use of hashed aggregation plans."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL
},
&enable_hashagg,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
true,
NULL, NULL, NULL
Finish implementation of hashed aggregation. Add enable_hashagg GUC parameter to allow it to be forced off for comparison purposes. Add ORDER BY clauses to a bunch of regression test queries that will otherwise produce randomly-ordered output in the new regime.
2002-11-21 01:42:20 +01:00
},
Add an 'enable_material' GUC. The logic for determining whether to materialize has been significantly overhauled for 9.0. In case there should be any doubt about whether materialization is a win in any particular case, this should provide a convenient way of seeing what happens without it; but even with enable_material turned off, we still materialize in cases where it is required for correctness. Thanks to Tom Lane for the review.
2010-04-19 02:55:26 +02:00
{
{"enable_material", PGC_USERSET, QUERY_TUNING_METHOD,
gettext_noop("Enables the planner's use of materialization."),
NULL
},
&enable_material,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
true,
NULL, NULL, NULL
Add an 'enable_material' GUC. The logic for determining whether to materialize has been significantly overhauled for 9.0. In case there should be any doubt about whether materialization is a win in any particular case, this should provide a convenient way of seeing what happens without it; but even with enable_material turned off, we still materialize in cases where it is required for correctness. Thanks to Tom Lane for the review.
2010-04-19 02:55:26 +02:00
},
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"enable_nestloop", PGC_USERSET, QUERY_TUNING_METHOD,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Enables the planner's use of nested-loop join plans."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL
},
&enable_nestloop,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
true,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"enable_mergejoin", PGC_USERSET, QUERY_TUNING_METHOD,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Enables the planner's use of merge join plans."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL
},
&enable_mergejoin,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
true,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"enable_hashjoin", PGC_USERSET, QUERY_TUNING_METHOD,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Enables the planner's use of hash join plans."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL
},
&enable_hashjoin,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
true,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"geqo", PGC_USERSET, QUERY_TUNING_GEQO,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Enables genetic query optimization."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
gettext_noop("This algorithm attempts to do planning without "
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
"exhaustive searching.")
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
&enable_geqo,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
true,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
Add is_superuser parameter reporting, soon to be used by psql.
2003-06-27 21:08:38 +02:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
/* Not for general use --- used by SET SESSION AUTHORIZATION */
{"is_superuser", PGC_INTERNAL, UNGROUPED,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Shows whether the current user is a superuser."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL,
GUC_REPORT | GUC_NO_SHOW_ALL | GUC_NO_RESET_ALL | GUC_NOT_IN_SAMPLE | GUC_DISALLOW_IN_FILE
},
Add is_superuser parameter reporting, soon to be used by psql.
2003-06-27 21:08:38 +02:00
&session_auth_is_superuser,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
NULL, NULL, NULL
Add is_superuser parameter reporting, soon to be used by psql.
2003-06-27 21:08:38 +02:00
},
Add a boolean GUC parameter "bonjour" to control whether a Bonjour-enabled build actually attempts to advertise itself via Bonjour. Formerly it always did so, which meant that packagers had to decide for their users whether this behavior was wanted or not. The default is "off" to be on the safe side, though this represents a change in the default behavior of a Bonjour-enabled build. Per discussion.
2009-09-08 19:08:36 +02:00
{
{"bonjour", PGC_POSTMASTER, CONN_AUTH_SETTINGS,
gettext_noop("Enables advertising the server via Bonjour."),
NULL
},
&enable_bonjour,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
check_bonjour, NULL, NULL
Add a boolean GUC parameter "bonjour" to control whether a Bonjour-enabled build actually attempts to advertise itself via Bonjour. Formerly it always did so, which meant that packagers had to decide for their users whether this behavior was wanted or not. The default is "off" to be on the safe side, though this represents a change in the default behavior of a Bonjour-enabled build. Per discussion.
2009-09-08 19:08:36 +02:00
},
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"ssl", PGC_POSTMASTER, CONN_AUTH_SECURITY,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Enables SSL connections."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL
},
&EnableSSL,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
check_ssl, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"fsync", PGC_SIGHUP, WAL_SETTINGS,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Forces synchronization of updates to disk."),
Message editing: remove gratuitous variations in message wording, standardize terms, add some clarifications, fix some untranslatable attempts at dynamic message building.
2003-09-25 08:58:07 +02:00
gettext_noop("The server will use the fsync() system call in several places to make "
Re-run pgindent, fixing a problem where comment lines after a blank comment line where output as too long, and update typedefs for /lib directory. Also fix case where identifiers were used as variable names in the backend, but as typedefs in ecpg (favor the backend for indenting). Backpatch to 8.1.X.
2005-11-22 19:17:34 +01:00
"sure that updates are physically written to disk. This insures "
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
"that a database cluster will recover to a consistent state after "
Message editing: remove gratuitous variations in message wording, standardize terms, add some clarifications, fix some untranslatable attempts at dynamic message building.
2003-09-25 08:58:07 +02:00
"an operating system or hardware crash.")
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
&enableFsync,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
true,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
Add code to apply some simple sanity checks to the header fields of a page when it's read in, per pghackers discussion around 17-Feb. Add a GUC variable zero_damaged_pages that causes the response to be a WARNING followed by zeroing the page, rather than the normal ERROR; this is per Hiroshi's suggestion that there needs to be a way to get at the data in the rest of the table.
2003-03-28 21:17:13 +01:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"zero_damaged_pages", PGC_SUSET, DEVELOPER_OPTIONS,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Continues processing past damaged page headers."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
gettext_noop("Detection of a damaged page header normally causes PostgreSQL to "
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
"report an error, aborting the current transaction. Setting "
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
"zero_damaged_pages to true causes the system to instead report a "
"warning, zero out the damaged page, and continue processing. This "
Message editing: remove gratuitous variations in message wording, standardize terms, add some clarifications, fix some untranslatable attempts at dynamic message building.
2003-09-25 08:58:07 +02:00
"behavior will destroy data, namely all the rows on the damaged page."),
zero_damaged_pages must absolutely NOT be marked GUC_DISALLOW_IN_FILE, else it cannot be used to handle failures detected during WAL replay. Fortunately this flag isn't actually enforced yet, but get it right.
2003-11-07 22:27:38 +01:00
GUC_NOT_IN_SAMPLE
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
&zero_damaged_pages,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
NULL, NULL, NULL
Add code to apply some simple sanity checks to the header fields of a page when it's read in, per pghackers discussion around 17-Feb. Add a GUC variable zero_damaged_pages that causes the response to be a WARNING followed by zeroing the page, rather than the normal ERROR; this is per Hiroshi's suggestion that there needs to be a way to get at the data in the rest of the table.
2003-03-28 21:17:13 +01:00
},
Add GUC full_page_writes to control writing full pages to WAL.
2005-07-06 01:18:10 +02:00
{
{"full_page_writes", PGC_SIGHUP, WAL_SETTINGS,
gettext_noop("Writes full pages to WAL when first modified after a checkpoint."),
gettext_noop("A page write in process during an operating system crash might be "
Add space after description. Euler Taveira de Oliveira
2005-10-16 20:26:00 +02:00
"only partially written to disk. During recovery, the row changes "
Re-run pgindent, fixing a problem where comment lines after a blank comment line where output as too long, and update typedefs for /lib directory. Also fix case where identifiers were used as variable names in the backend, but as typedefs in ecpg (favor the backend for indenting). Backpatch to 8.1.X.
2005-11-22 19:17:34 +01:00
"stored in WAL are not enough to recover. This option writes "
Add GUC full_page_writes to control writing full pages to WAL.
2005-07-06 01:18:10 +02:00
"pages when first modified after a checkpoint to WAL so full recovery "
"is possible.")
},
&fullPageWrites,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
true,
NULL, NULL, NULL
Add GUC full_page_writes to control writing full pages to WAL.
2005-07-06 01:18:10 +02:00
},
Improve logging of checkpoints. Patch by Greg Smith, worked over by Heikki and a little bit by me.
2007-06-30 21:12:02 +02:00
{
{"log_checkpoints", PGC_SIGHUP, LOGGING_WHAT,
gettext_noop("Logs each checkpoint."),
NULL
},
&log_checkpoints,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
NULL, NULL, NULL
Improve logging of checkpoints. Patch by Greg Smith, worked over by Heikki and a little bit by me.
2007-06-30 21:12:02 +02:00
},
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"log_connections", PGC_BACKEND, LOGGING_WHAT,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Logs each successful connection."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL
},
&Log_connections,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
This patch brings up to date what I did last year (now unfortunately bitrotted) to allow the logging of the end of a session, enabled by the config setting "log_disconnections". Andrew Dunstan
2004-02-17 04:54:57 +01:00
{
{"log_disconnections", PGC_BACKEND, LOGGING_WHAT,
Message style revisions
2004-10-12 23:54:45 +02:00
gettext_noop("Logs end of a session, including duration."),
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
NULL
This patch brings up to date what I did last year (now unfortunately bitrotted) to allow the logging of the end of a session, enabled by the config setting "log_disconnections". Andrew Dunstan
2004-02-17 04:54:57 +01:00
},
&Log_disconnections,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
NULL, NULL, NULL
This patch brings up to date what I did last year (now unfortunately bitrotted) to allow the logging of the end of a session, enabled by the config setting "log_disconnections". Andrew Dunstan
2004-02-17 04:54:57 +01:00
},
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"debug_assertions", PGC_USERSET, DEVELOPER_OPTIONS,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Turns on various assertion checks."),
Message editing: remove gratuitous variations in message wording, standardize terms, add some clarifications, fix some untranslatable attempts at dynamic message building.
2003-09-25 08:58:07 +02:00
gettext_noop("This is a debugging aid."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
GUC_NOT_IN_SAMPLE
},
&assert_enabled,
Make all command-line options of postmaster and postgres the same. See http://archives.postgresql.org/pgsql-hackers/2006-01/msg00151.php for the complete plan.
2006-01-05 11:07:46 +01:00
#ifdef USE_ASSERT_CHECKING
true,
#else
false,
Add assert checking to GUC ("debug_assertions") Rename settings net_server to tcpip_socket, max_backends to max_connections Add --help and --version to postmaster, reformat help output
2000-07-12 19:38:53 +02:00
#endif
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
check_debug_assertions, NULL, NULL
Make all command-line options of postmaster and postgres the same. See http://archives.postgresql.org/pgsql-hackers/2006-01/msg00151.php for the complete plan.
2006-01-05 11:07:46 +01:00
},
Add restart_after_crash GUC. Normally, we automatically restart after a backend crash, but in some cases when PostgreSQL is invoked by clusterware it may be desirable to suppress this behavior, so we provide an option which does this. Since no existing GUC group quite fits, create a new group called "error handling options" for this and the previously undocumented GUC exit_on_error, which is now documented. Review by Fujii Masao.
2010-07-20 02:47:53 +02:00
Fix things so that an error occuring during standalone-backend processing in initdb will result in exit(1), allowing the initdb script to realize that there's something wrong.
2003-05-28 20:19:09 +02:00
{
Add restart_after_crash GUC. Normally, we automatically restart after a backend crash, but in some cases when PostgreSQL is invoked by clusterware it may be desirable to suppress this behavior, so we provide an option which does this. Since no existing GUC group quite fits, create a new group called "error handling options" for this and the previously undocumented GUC exit_on_error, which is now documented. Review by Fujii Masao.
2010-07-20 02:47:53 +02:00
{"exit_on_error", PGC_USERSET, ERROR_HANDLING_OPTIONS,
gettext_noop("Terminate session on any error."),
NULL
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
&ExitOnAnyError,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
NULL, NULL, NULL
Fix things so that an error occuring during standalone-backend processing in initdb will result in exit(1), allowing the initdb script to realize that there's something wrong.
2003-05-28 20:19:09 +02:00
},
Add restart_after_crash GUC. Normally, we automatically restart after a backend crash, but in some cases when PostgreSQL is invoked by clusterware it may be desirable to suppress this behavior, so we provide an option which does this. Since no existing GUC group quite fits, create a new group called "error handling options" for this and the previously undocumented GUC exit_on_error, which is now documented. Review by Fujii Masao.
2010-07-20 02:47:53 +02:00
{
{"restart_after_crash", PGC_SIGHUP, ERROR_HANDLING_OPTIONS,
Message style tweaks
2011-07-04 23:01:35 +02:00
gettext_noop("Reinitialize server after backend crash."),
Add restart_after_crash GUC. Normally, we automatically restart after a backend crash, but in some cases when PostgreSQL is invoked by clusterware it may be desirable to suppress this behavior, so we provide an option which does this. Since no existing GUC group quite fits, create a new group called "error handling options" for this and the previously undocumented GUC exit_on_error, which is now documented. Review by Fujii Masao.
2010-07-20 02:47:53 +02:00
NULL
},
&restart_after_crash,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
true,
NULL, NULL, NULL
Add restart_after_crash GUC. Normally, we automatically restart after a backend crash, but in some cases when PostgreSQL is invoked by clusterware it may be desirable to suppress this behavior, so we provide an option which does this. Since no existing GUC group quite fits, create a new group called "error handling options" for this and the previously undocumented GUC exit_on_error, which is now documented. Review by Fujii Masao.
2010-07-20 02:47:53 +02:00
},
Add log_duration to GUC/postgresql.conf. Rename debug_print_query to log_statement and rename show_query_stats to show_statement_stats.
2002-09-02 01:26:06 +02:00
{
Remove GUC USERLIMIT variable category, making the affected variables plain SUSET instead. Also delay processing of options received in client connection request until after we know if the user is a superuser, so that SUSET values can be set that way by legitimate superusers. Per recent discussion.
2004-11-14 20:35:35 +01:00
{"log_duration", PGC_SUSET, LOGGING_WHAT,
gettext_noop("Logs the duration of each completed SQL statement."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL
},
&log_duration,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"debug_print_parse", PGC_USERSET, LOGGING_WHAT,
Cause the output from debug_print_parse, debug_print_rewritten, and debug_print_plan to appear at LOG message level, not DEBUG1 as historically. Make debug_pretty_print default to on. Also, cause plans generated via EXPLAIN to be subject to debug_print_plan. This is all to make debug_print_plan a reasonably comfortable substitute for the former behavior of EXPLAIN VERBOSE.
2008-08-19 20:30:04 +02:00
gettext_noop("Logs each query's parse tree."),
Message editing: remove gratuitous variations in message wording, standardize terms, add some clarifications, fix some untranslatable attempts at dynamic message building.
2003-09-25 08:58:07 +02:00
NULL
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
&Debug_print_parse,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"debug_print_rewritten", PGC_USERSET, LOGGING_WHAT,
Cause the output from debug_print_parse, debug_print_rewritten, and debug_print_plan to appear at LOG message level, not DEBUG1 as historically. Make debug_pretty_print default to on. Also, cause plans generated via EXPLAIN to be subject to debug_print_plan. This is all to make debug_print_plan a reasonably comfortable substitute for the former behavior of EXPLAIN VERBOSE.
2008-08-19 20:30:04 +02:00
gettext_noop("Logs each query's rewritten parse tree."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL
},
&Debug_print_rewritten,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"debug_print_plan", PGC_USERSET, LOGGING_WHAT,
Cause the output from debug_print_parse, debug_print_rewritten, and debug_print_plan to appear at LOG message level, not DEBUG1 as historically. Make debug_pretty_print default to on. Also, cause plans generated via EXPLAIN to be subject to debug_print_plan. This is all to make debug_print_plan a reasonably comfortable substitute for the former behavior of EXPLAIN VERBOSE.
2008-08-19 20:30:04 +02:00
gettext_noop("Logs each query's execution plan."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL
},
&Debug_print_plan,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"debug_pretty_print", PGC_USERSET, LOGGING_WHAT,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Indents parse and plan tree displays."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL
},
&Debug_pretty_print,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
true,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
{
Remove GUC USERLIMIT variable category, making the affected variables plain SUSET instead. Also delay processing of options received in client connection request until after we know if the user is a superuser, so that SUSET values can be set that way by legitimate superusers. Per recent discussion.
2004-11-14 20:35:35 +01:00
{"log_parser_stats", PGC_SUSET, STATS_MONITORING,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Writes parser performance statistics to the server log."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL
},
&log_parser_stats,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
check_stage_log_stats, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
{
Remove GUC USERLIMIT variable category, making the affected variables plain SUSET instead. Also delay processing of options received in client connection request until after we know if the user is a superuser, so that SUSET values can be set that way by legitimate superusers. Per recent discussion.
2004-11-14 20:35:35 +01:00
{"log_planner_stats", PGC_SUSET, STATS_MONITORING,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Writes planner performance statistics to the server log."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL
},
&log_planner_stats,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
check_stage_log_stats, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
{
Remove GUC USERLIMIT variable category, making the affected variables plain SUSET instead. Also delay processing of options received in client connection request until after we know if the user is a superuser, so that SUSET values can be set that way by legitimate superusers. Per recent discussion.
2004-11-14 20:35:35 +01:00
{"log_executor_stats", PGC_SUSET, STATS_MONITORING,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Writes executor performance statistics to the server log."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL
},
&log_executor_stats,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
check_stage_log_stats, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
{
Remove GUC USERLIMIT variable category, making the affected variables plain SUSET instead. Also delay processing of options received in client connection request until after we know if the user is a superuser, so that SUSET values can be set that way by legitimate superusers. Per recent discussion.
2004-11-14 20:35:35 +01:00
{"log_statement_stats", PGC_SUSET, STATS_MONITORING,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Writes cumulative performance statistics to the server log."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL
},
&log_statement_stats,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
check_log_stats, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
#ifdef BTREE_BUILD_STATS
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"log_btree_build_stats", PGC_SUSET, DEVELOPER_OPTIONS,
Various improvements to the GUC description strings. Punctuate and capitalize the strings like sentences. Remove unnecessarily specific descriptions of the units used by GUC variables, since we now allow any reasonable unit to be specified.
2006-12-06 18:35:49 +01:00
gettext_noop("No description available."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL,
GUC_NOT_IN_SAMPLE
},
&log_btree_build_stats,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
#endif
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Simplify and rename some GUC variables, per various recent discussions: * stats_start_collector goes away; we always start the collector process, unless prevented by a problem with setting up the stats UDP socket. * stats_reset_on_server_start goes away; it seems useless in view of the availability of pg_stat_reset(). * stats_block_level and stats_row_level are merged into a single variable "track_counts", which controls all reports sent to the collector process. * stats_command_string is renamed to track_activities. * log_autovacuum is renamed to log_autovacuum_min_duration to better reflect its meaning. The log_autovacuum change is not a compatibility issue since it didn't exist before 8.3 anyway. The other changes need to be release-noted.
2007-09-24 05:12:23 +02:00
{"track_activities", PGC_SUSET, STATS_COLLECTOR,
gettext_noop("Collects information about executing commands."),
gettext_noop("Enables the collection of information on the currently "
"executing command of each session, along with "
"the time at which that command began execution.")
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
Simplify and rename some GUC variables, per various recent discussions: * stats_start_collector goes away; we always start the collector process, unless prevented by a problem with setting up the stats UDP socket. * stats_reset_on_server_start goes away; it seems useless in view of the availability of pg_stat_reset(). * stats_block_level and stats_row_level are merged into a single variable "track_counts", which controls all reports sent to the collector process. * stats_command_string is renamed to track_activities. * log_autovacuum is renamed to log_autovacuum_min_duration to better reflect its meaning. The log_autovacuum change is not a compatibility issue since it didn't exist before 8.3 anyway. The other changes need to be release-noted.
2007-09-24 05:12:23 +02:00
&pgstat_track_activities,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
true,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
{
Simplify and rename some GUC variables, per various recent discussions: * stats_start_collector goes away; we always start the collector process, unless prevented by a problem with setting up the stats UDP socket. * stats_reset_on_server_start goes away; it seems useless in view of the availability of pg_stat_reset(). * stats_block_level and stats_row_level are merged into a single variable "track_counts", which controls all reports sent to the collector process. * stats_command_string is renamed to track_activities. * log_autovacuum is renamed to log_autovacuum_min_duration to better reflect its meaning. The log_autovacuum change is not a compatibility issue since it didn't exist before 8.3 anyway. The other changes need to be release-noted.
2007-09-24 05:12:23 +02:00
{"track_counts", PGC_SUSET, STATS_COLLECTOR,
gettext_noop("Collects statistics on database activity."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL
},
Simplify and rename some GUC variables, per various recent discussions: * stats_start_collector goes away; we always start the collector process, unless prevented by a problem with setting up the stats UDP socket. * stats_reset_on_server_start goes away; it seems useless in view of the availability of pg_stat_reset(). * stats_block_level and stats_row_level are merged into a single variable "track_counts", which controls all reports sent to the collector process. * stats_command_string is renamed to track_activities. * log_autovacuum is renamed to log_autovacuum_min_duration to better reflect its meaning. The log_autovacuum change is not a compatibility issue since it didn't exist before 8.3 anyway. The other changes need to be release-noted.
2007-09-24 05:12:23 +02:00
&pgstat_track_counts,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
true,
NULL, NULL, NULL
Take the statistics collector out of the loop for monitoring backends' current commands; instead, store current-status information in shared memory. This substantially reduces the overhead of stats_command_string and also ensures that pg_stat_activity is fully up to date at all times. Per my recent proposal.
2006-06-19 03:51:22 +02:00
},
Add GUC update_process_title to control whether 'ps' display is updated for every command, default to on.
2006-06-28 00:16:44 +02:00
{
{"update_process_title", PGC_SUSET, STATS_COLLECTOR,
gettext_noop("Updates the process title to show the active SQL command."),
Remove embedded newline in string literal --- seems to make newer gcc versions very unhappy, and shouldn't be there anyway.
2006-06-28 01:25:28 +02:00
gettext_noop("Enables updating of the process title every time a new SQL command is received by the server.")
Add GUC update_process_title to control whether 'ps' display is updated for every command, default to on.
2006-06-28 00:16:44 +02:00
},
&update_process_title,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
true,
NULL, NULL, NULL
Add GUC update_process_title to control whether 'ps' display is updated for every command, default to on.
2006-06-28 00:16:44 +02:00
},
Integrate autovacuum functionality into the backend. There's still a few loose ends to be dealt with, but it seems to work. Alvaro Herrera, based on the contrib code by Matthew O'Connor.
2005-07-14 07:13:45 +02:00
{
{"autovacuum", PGC_SIGHUP, AUTOVACUUM,
Message corrections
2005-10-29 02:31:52 +02:00
gettext_noop("Starts the autovacuum subprocess."),
Integrate autovacuum functionality into the backend. There's still a few loose ends to be dealt with, but it seems to work. Alvaro Herrera, based on the contrib code by Matthew O'Connor.
2005-07-14 07:13:45 +02:00
NULL
},
&autovacuum_start_daemon,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
true,
NULL, NULL, NULL
Integrate autovacuum functionality into the backend. There's still a few loose ends to be dealt with, but it seems to work. Alvaro Herrera, based on the contrib code by Matthew O'Connor.
2005-07-14 07:13:45 +02:00
},
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"trace_notify", PGC_USERSET, DEVELOPER_OPTIONS,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Generates debugging output for LISTEN and NOTIFY."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL,
GUC_NOT_IN_SAMPLE
},
&Trace_notify,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
#ifdef LOCK_DEBUG
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"trace_locks", PGC_SUSET, DEVELOPER_OPTIONS,
Various improvements to the GUC description strings. Punctuate and capitalize the strings like sentences. Remove unnecessarily specific descriptions of the units used by GUC variables, since we now allow any reasonable unit to be specified.
2006-12-06 18:35:49 +01:00
gettext_noop("No description available."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL,
GUC_NOT_IN_SAMPLE
},
&Trace_locks,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"trace_userlocks", PGC_SUSET, DEVELOPER_OPTIONS,
Various improvements to the GUC description strings. Punctuate and capitalize the strings like sentences. Remove unnecessarily specific descriptions of the units used by GUC variables, since we now allow any reasonable unit to be specified.
2006-12-06 18:35:49 +01:00
gettext_noop("No description available."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL,
GUC_NOT_IN_SAMPLE
},
&Trace_userlocks,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"trace_lwlocks", PGC_SUSET, DEVELOPER_OPTIONS,
Various improvements to the GUC description strings. Punctuate and capitalize the strings like sentences. Remove unnecessarily specific descriptions of the units used by GUC variables, since we now allow any reasonable unit to be specified.
2006-12-06 18:35:49 +01:00
gettext_noop("No description available."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL,
GUC_NOT_IN_SAMPLE
},
&Trace_lwlocks,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"debug_deadlocks", PGC_SUSET, DEVELOPER_OPTIONS,
Various improvements to the GUC description strings. Punctuate and capitalize the strings like sentences. Remove unnecessarily specific descriptions of the units used by GUC variables, since we now allow any reasonable unit to be specified.
2006-12-06 18:35:49 +01:00
gettext_noop("No description available."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL,
GUC_NOT_IN_SAMPLE
},
&Debug_deadlocks,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
#endif
Add GUC log_lock_waits to log long wait times. Simon Riggs
2007-03-03 19:46:40 +01:00
{
Code review for log_lock_waits patch. Don't try to issue log messages from within a signal handler (this might be safe given the relatively narrow code range in which the interrupt is enabled, but it seems awfully risky); do issue more informative log messages that tell what is being waited for and the exact length of the wait; minor other code cleanup. Greg Stark and Tom Lane
2007-06-19 22:13:22 +02:00
{"log_lock_waits", PGC_SUSET, LOGGING_WHAT,
gettext_noop("Logs long lock waits."),
Add GUC log_lock_waits to log long wait times. Simon Riggs
2007-03-03 19:46:40 +01:00
NULL
},
&log_lock_waits,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
NULL, NULL, NULL
Add GUC log_lock_waits to log long wait times. Simon Riggs
2007-03-03 19:46:40 +01:00
},
Code review for log_lock_waits patch. Don't try to issue log messages from within a signal handler (this might be safe given the relatively narrow code range in which the interrupt is enabled, but it seems awfully risky); do issue more informative log messages that tell what is being waited for and the exact length of the wait; minor other code cleanup. Greg Stark and Tom Lane
2007-06-19 22:13:22 +02:00
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"log_hostname", PGC_SIGHUP, LOGGING_WHAT,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Logs the host name in the connection logs."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
gettext_noop("By default, connection logs only show the IP address "
"of the connecting host. If you want them to show the host name you "
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
"can turn this on, but depending on your host name resolution "
"setup it might impose a non-negligible performance penalty.")
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
&log_hostname,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"sql_inheritance", PGC_USERSET, COMPAT_OPTIONS_PREVIOUS,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Causes subtables to be included by default in various commands."),
Message editing: remove gratuitous variations in message wording, standardize terms, add some clarifications, fix some untranslatable attempts at dynamic message building.
2003-09-25 08:58:07 +02:00
NULL
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
&SQL_inheritance,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
true,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"password_encryption", PGC_USERSET, CONN_AUTH_SECURITY,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Encrypt passwords."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
gettext_noop("When a password is specified in CREATE USER or "
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
"ALTER USER without writing either ENCRYPTED or UNENCRYPTED, "
More message editing, some suggested by Alvaro Herrera
2003-09-29 02:05:25 +02:00
"this parameter determines whether the password is to be encrypted.")
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
&Password_encryption,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
true,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"transform_null_equals", PGC_USERSET, COMPAT_OPTIONS_CLIENT,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Treats \"expr=NULL\" as \"expr IS NULL\"."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
gettext_noop("When turned on, expressions of the form expr = NULL "
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
"(or NULL = expr) are treated as expr IS NULL, that is, they "
"return true if expr evaluates to the null value, and false "
"otherwise. The correct behavior of expr = NULL is to always "
Message editing: remove gratuitous variations in message wording, standardize terms, add some clarifications, fix some untranslatable attempts at dynamic message building.
2003-09-25 08:58:07 +02:00
"return null (unknown).")
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
&Transform_null_equals,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
Add db-local user names, per discussion on hackers.
2002-08-18 05:03:26 +02:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"db_user_namespace", PGC_SIGHUP, CONN_AUTH_SECURITY,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Enables per-database user names."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL
},
&Db_user_namespace,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
NULL, NULL, NULL
Add db-local user names, per discussion on hackers.
2002-08-18 05:03:26 +02:00
},
AUTOCOMMIT mode is now an available backend GUC variable; setting it to false provides more SQL-spec-compliant behavior than we had before. I am not sure that setting it false is actually a good idea yet; there is a lot of client-side code that will probably be broken by turning autocommit off. But it's a start. Loosely based on a patch by David Van Wie.
2002-08-31 00:18:07 +02:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
/* only here for backwards compatibility */
{"autocommit", PGC_USERSET, CLIENT_CONN_STATEMENT,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("This parameter doesn't do anything."),
gettext_noop("It's just here so that we won't choke on SET AUTOCOMMIT TO ON from 7.3-vintage clients."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
GUC_NO_SHOW_ALL | GUC_NOT_IN_SAMPLE
},
&phony_autocommit,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
true,
check_phony_autocommit, NULL, NULL
AUTOCOMMIT mode is now an available backend GUC variable; setting it to false provides more SQL-spec-compliant behavior than we had before. I am not sure that setting it false is actually a good idea yet; there is a lot of client-side code that will probably be broken by turning autocommit off. But it's a start. Loosely based on a patch by David Van Wie.
2002-08-31 00:18:07 +02:00
},
Read-only transactions, as defined in SQL.
2003-01-10 23:03:30 +01:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"default_transaction_read_only", PGC_USERSET, CLIENT_CONN_STATEMENT,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Sets the default read-only status of new transactions."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL
},
&DefaultXactReadOnly,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
NULL, NULL, NULL
Read-only transactions, as defined in SQL.
2003-01-10 23:03:30 +01:00
},
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"transaction_read_only", PGC_USERSET, CLIENT_CONN_STATEMENT,
Seems there are three GUC variables that are defined as "Shows ..." while you can actually set them with SET. This applied patch changes the wording from "Show" to "Set".
2003-12-01 04:55:21 +01:00
gettext_noop("Sets the current transaction's read-only status."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL,
GUC_NO_RESET_ALL | GUC_NOT_IN_SAMPLE | GUC_DISALLOW_IN_FILE
},
&XactReadOnly,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
check_transaction_read_only, NULL, NULL
Read-only transactions, as defined in SQL.
2003-01-10 23:03:30 +01:00
},
Implement genuine serializable isolation level. Until now, our Serializable mode has in fact been what's called Snapshot Isolation, which allows some anomalies that could not occur in any serialized ordering of the transactions. This patch fixes that using a method called Serializable Snapshot Isolation, based on research papers by Michael J. Cahill (see README-SSI for full references). In Serializable Snapshot Isolation, transactions run like they do in Snapshot Isolation, but a predicate lock manager observes the reads and writes performed and aborts transactions if it detects that an anomaly might occur. This method produces some false positives, ie. it sometimes aborts transactions even though there is no anomaly. To track reads we implement predicate locking, see storage/lmgr/predicate.c. Whenever a tuple is read, a predicate lock is acquired on the tuple. Shared memory is finite, so when a transaction takes many tuple-level locks on a page, the locks are promoted to a single page-level lock, and further to a single relation level lock if necessary. To lock key values with no matching tuple, a sequential scan always takes a relation-level lock, and an index scan acquires a page-level lock that covers the search key, whether or not there are any matching keys at the moment. A predicate lock doesn't conflict with any regular locks or with another predicate locks in the normal sense. They're only used by the predicate lock manager to detect the danger of anomalies. Only serializable transactions participate in predicate locking, so there should be no extra overhead for for other transactions. Predicate locks can't be released at commit, but must be remembered until all the transactions that overlapped with it have completed. That means that we need to remember an unbounded amount of predicate locks, so we apply a lossy but conservative method of tracking locks for committed transactions. If we run short of shared memory, we overflow to a new "pg_serial" SLRU pool. We don't currently allow Serializable transactions in Hot Standby mode. That would be hard, because even read-only transactions can cause anomalies that wouldn't otherwise occur. Serializable isolation mode now means the new fully serializable level. Repeatable Read gives you the old Snapshot Isolation level that we have always had. Kevin Grittner and Dan Ports, reviewed by Jeff Davis, Heikki Linnakangas and Anssi Kääriäinen
2011-02-07 22:46:51 +01:00
{
{"default_transaction_deferrable", PGC_USERSET, CLIENT_CONN_STATEMENT,
gettext_noop("Sets the default deferrable status of new transactions."),
NULL
},
&DefaultXactDeferrable,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
NULL, NULL, NULL
Implement genuine serializable isolation level. Until now, our Serializable mode has in fact been what's called Snapshot Isolation, which allows some anomalies that could not occur in any serialized ordering of the transactions. This patch fixes that using a method called Serializable Snapshot Isolation, based on research papers by Michael J. Cahill (see README-SSI for full references). In Serializable Snapshot Isolation, transactions run like they do in Snapshot Isolation, but a predicate lock manager observes the reads and writes performed and aborts transactions if it detects that an anomaly might occur. This method produces some false positives, ie. it sometimes aborts transactions even though there is no anomaly. To track reads we implement predicate locking, see storage/lmgr/predicate.c. Whenever a tuple is read, a predicate lock is acquired on the tuple. Shared memory is finite, so when a transaction takes many tuple-level locks on a page, the locks are promoted to a single page-level lock, and further to a single relation level lock if necessary. To lock key values with no matching tuple, a sequential scan always takes a relation-level lock, and an index scan acquires a page-level lock that covers the search key, whether or not there are any matching keys at the moment. A predicate lock doesn't conflict with any regular locks or with another predicate locks in the normal sense. They're only used by the predicate lock manager to detect the danger of anomalies. Only serializable transactions participate in predicate locking, so there should be no extra overhead for for other transactions. Predicate locks can't be released at commit, but must be remembered until all the transactions that overlapped with it have completed. That means that we need to remember an unbounded amount of predicate locks, so we apply a lossy but conservative method of tracking locks for committed transactions. If we run short of shared memory, we overflow to a new "pg_serial" SLRU pool. We don't currently allow Serializable transactions in Hot Standby mode. That would be hard, because even read-only transactions can cause anomalies that wouldn't otherwise occur. Serializable isolation mode now means the new fully serializable level. Repeatable Read gives you the old Snapshot Isolation level that we have always had. Kevin Grittner and Dan Ports, reviewed by Jeff Davis, Heikki Linnakangas and Anssi Kääriäinen
2011-02-07 22:46:51 +01:00
},
{
{"transaction_deferrable", PGC_USERSET, CLIENT_CONN_STATEMENT,
gettext_noop("Whether to defer a read-only serializable transaction until it can be executed with no possible serialization failures."),
NULL,
GUC_NO_RESET_ALL | GUC_NOT_IN_SAMPLE | GUC_DISALLOW_IN_FILE
},
&XactDeferrable,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
check_transaction_deferrable, NULL, NULL
Implement genuine serializable isolation level. Until now, our Serializable mode has in fact been what's called Snapshot Isolation, which allows some anomalies that could not occur in any serialized ordering of the transactions. This patch fixes that using a method called Serializable Snapshot Isolation, based on research papers by Michael J. Cahill (see README-SSI for full references). In Serializable Snapshot Isolation, transactions run like they do in Snapshot Isolation, but a predicate lock manager observes the reads and writes performed and aborts transactions if it detects that an anomaly might occur. This method produces some false positives, ie. it sometimes aborts transactions even though there is no anomaly. To track reads we implement predicate locking, see storage/lmgr/predicate.c. Whenever a tuple is read, a predicate lock is acquired on the tuple. Shared memory is finite, so when a transaction takes many tuple-level locks on a page, the locks are promoted to a single page-level lock, and further to a single relation level lock if necessary. To lock key values with no matching tuple, a sequential scan always takes a relation-level lock, and an index scan acquires a page-level lock that covers the search key, whether or not there are any matching keys at the moment. A predicate lock doesn't conflict with any regular locks or with another predicate locks in the normal sense. They're only used by the predicate lock manager to detect the danger of anomalies. Only serializable transactions participate in predicate locking, so there should be no extra overhead for for other transactions. Predicate locks can't be released at commit, but must be remembered until all the transactions that overlapped with it have completed. That means that we need to remember an unbounded amount of predicate locks, so we apply a lossy but conservative method of tracking locks for committed transactions. If we run short of shared memory, we overflow to a new "pg_serial" SLRU pool. We don't currently allow Serializable transactions in Hot Standby mode. That would be hard, because even read-only transactions can cause anomalies that wouldn't otherwise occur. Serializable isolation mode now means the new fully serializable level. Repeatable Read gives you the old Snapshot Isolation level that we have always had. Kevin Grittner and Dan Ports, reviewed by Jeff Davis, Heikki Linnakangas and Anssi Kääriäinen
2011-02-07 22:46:51 +01:00
},
Add GUC parameter check_function_bodies to control whether validation of function bodies is done at CREATE FUNCTION time. This is normally true but can be set false to avoid problems with forward references, wrong schema search path, etc. This is just the backend patch, still need to adjust pg_dump to make use of it.
2003-10-03 21:26:49 +02:00
{
{"check_function_bodies", PGC_USERSET, CLIENT_CONN_STATEMENT,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Check function bodies during CREATE FUNCTION."),
Add GUC parameter check_function_bodies to control whether validation of function bodies is done at CREATE FUNCTION time. This is normally true but can be set false to avoid problems with forward references, wrong schema search path, etc. This is just the backend patch, still need to adjust pg_dump to make use of it.
2003-10-03 21:26:49 +02:00
NULL
},
&check_function_bodies,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
true,
NULL, NULL, NULL
Add GUC parameter check_function_bodies to control whether validation of function bodies is done at CREATE FUNCTION time. This is normally true but can be set false to avoid problems with forward references, wrong schema search path, etc. This is just the backend patch, still need to adjust pg_dump to make use of it.
2003-10-03 21:26:49 +02:00
},
Make SQL arrays support null elements. This commit fixes the core array functionality, but I still need to make another pass looking at places that incidentally use arrays (such as ACL manipulation) to make sure they are null-safe. Contrib needs work too. I have not changed the behaviors that are still under discussion about array comparison and what to do with lower bounds.
2005-11-17 23:14:56 +01:00
{
{"array_nulls", PGC_USERSET, COMPAT_OPTIONS_PREVIOUS,
gettext_noop("Enable input of NULL elements in arrays."),
gettext_noop("When turned on, unquoted NULL in an array input "
Message style improvements
2006-10-06 19:14:01 +02:00
"value means a null value; "
Make SQL arrays support null elements. This commit fixes the core array functionality, but I still need to make another pass looking at places that incidentally use arrays (such as ACL manipulation) to make sure they are null-safe. Contrib needs work too. I have not changed the behaviors that are still under discussion about array comparison and what to do with lower bounds.
2005-11-17 23:14:56 +01:00
"otherwise it is taken literally.")
},
&Array_nulls,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
true,
NULL, NULL, NULL
Make SQL arrays support null elements. This commit fixes the core array functionality, but I still need to make another pass looking at places that incidentally use arrays (such as ACL manipulation) to make sure they are null-safe. Contrib needs work too. I have not changed the behaviors that are still under discussion about array comparison and what to do with lower bounds.
2005-11-17 23:14:56 +01:00
},
This patch adds a new GUC var, "default_with_oids", which follows the proposal for eventually deprecating OIDs on user tables that I posted earlier to pgsql-hackers. pg_dump now always specifies WITH OIDS or WITHOUT OIDS when dumping a table. The documentation has been updated. Neil Conway
2003-12-01 23:08:02 +01:00
{
{"default_with_oids", PGC_USERSET, COMPAT_OPTIONS_PREVIOUS,
Make default_with_oids default to false -- user-created tables will now no longer include OIDs, unless WITH OIDS is specified or the default_with_oids configuration parameter is enabled. Update the docs accordingly.
2005-03-13 10:36:31 +01:00
gettext_noop("Create new tables with OIDs by default."),
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
NULL
This patch adds a new GUC var, "default_with_oids", which follows the proposal for eventually deprecating OIDs on user tables that I posted earlier to pgsql-hackers. pg_dump now always specifies WITH OIDS or WITHOUT OIDS when dumping a table. The documentation has been updated. Neil Conway
2003-12-01 23:08:02 +01:00
},
&default_with_oids,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
NULL, NULL, NULL
This patch adds a new GUC var, "default_with_oids", which follows the proposal for eventually deprecating OIDs on user tables that I posted earlier to pgsql-hackers. pg_dump now always specifies WITH OIDS or WITHOUT OIDS when dumping a table. The documentation has been updated. Neil Conway
2003-12-01 23:08:02 +01:00
},
Added new group of read-only GUC variables to allow simple access to certain compile-time options (FUNC_MAX_ARGS, INDEX_MAX_KEYS, NAMEDATALEN, BLCKSZ, HAVE_INT64_TIMESTAMP). Also added "category", "short_desc", and "extra_desc" to the pg_settings view. Per recent discussion here: http://archives.postgresql.org/pgsql-patches/2003-11/msg00363.php
2003-12-03 19:52:00 +01:00
{
Provide for logfiles in machine readable CSV format. In consequence, rename redirect_stderr to logging_collector. Original patch from Arul Shaji, subsequently modified by Greg Smith, and then heavily modified by me.
2007-08-19 03:41:25 +02:00
{"logging_collector", PGC_POSTMASTER, LOGGING_WHERE,
gettext_noop("Start a subprocess to capture stderr output and/or csvlogs into log files."),
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
NULL
Added new group of read-only GUC variables to allow simple access to certain compile-time options (FUNC_MAX_ARGS, INDEX_MAX_KEYS, NAMEDATALEN, BLCKSZ, HAVE_INT64_TIMESTAMP). Also added "category", "short_desc", and "extra_desc" to the pg_settings view. Per recent discussion here: http://archives.postgresql.org/pgsql-patches/2003-11/msg00363.php
2003-12-03 19:52:00 +01:00
},
Provide for logfiles in machine readable CSV format. In consequence, rename redirect_stderr to logging_collector. Original patch from Arul Shaji, subsequently modified by Greg Smith, and then heavily modified by me.
2007-08-19 03:41:25 +02:00
&Logging_collector,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
NULL, NULL, NULL
Added new group of read-only GUC variables to allow simple access to certain compile-time options (FUNC_MAX_ARGS, INDEX_MAX_KEYS, NAMEDATALEN, BLCKSZ, HAVE_INT64_TIMESTAMP). Also added "category", "short_desc", and "extra_desc" to the pg_settings view. Per recent discussion here: http://archives.postgresql.org/pgsql-patches/2003-11/msg00363.php
2003-12-03 19:52:00 +01:00
},
Replace log_filename_prefix with more general log_filename parameter, to allow DBA to choose the form in which log filenames reflect the current time. Also allow for truncating instead of appending to pre-existing files --- this is convenient when the log filename pattern rewrites the same names cyclically. Per Ed L.
2004-08-31 06:53:44 +02:00
{
{"log_truncate_on_rotation", PGC_SIGHUP, LOGGING_WHERE,
Message style revisions
2004-10-12 23:54:45 +02:00
gettext_noop("Truncate existing log files of same name during log rotation."),
Replace log_filename_prefix with more general log_filename parameter, to allow DBA to choose the form in which log filenames reflect the current time. Also allow for truncating instead of appending to pre-existing files --- this is convenient when the log filename pattern rewrites the same names cyclically. Per Ed L.
2004-08-31 06:53:44 +02:00
NULL
},
&Log_truncate_on_rotation,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
NULL, NULL, NULL
Replace log_filename_prefix with more general log_filename parameter, to allow DBA to choose the form in which log filenames reflect the current time. Also allow for truncating instead of appending to pre-existing files --- this is convenient when the log filename pattern rewrites the same names cyclically. Per Ed L.
2004-08-31 06:53:44 +02:00
},
Added new group of read-only GUC variables to allow simple access to certain compile-time options (FUNC_MAX_ARGS, INDEX_MAX_KEYS, NAMEDATALEN, BLCKSZ, HAVE_INT64_TIMESTAMP). Also added "category", "short_desc", and "extra_desc" to the pg_settings view. Per recent discussion here: http://archives.postgresql.org/pgsql-patches/2003-11/msg00363.php
2003-12-03 19:52:00 +01:00
Add a trace_sort option to help with measuring resource usage of external sort operations. Per recent discussion. Simon Riggs and Tom Lane.
2005-10-04 00:55:56 +02:00
#ifdef TRACE_SORT
{
{"trace_sort", PGC_USERSET, DEVELOPER_OPTIONS,
gettext_noop("Emit information about resource usage in sorting."),
NULL,
GUC_NOT_IN_SAMPLE
},
&trace_sort,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
NULL, NULL, NULL
Add a trace_sort option to help with measuring resource usage of external sort operations. Per recent discussion. Simon Riggs and Tom Lane.
2005-10-04 00:55:56 +02:00
},
#endif
Arrange for large sequential scans to synchronize with each other, so that when multiple backends are scanning the same relation concurrently, each page is (ideally) read only once. Jeff Davis, with review by Heikki and Tom.
2007-06-08 20:23:53 +02:00
#ifdef TRACE_SYNCSCAN
/* this is undocumented because not exposed in a standard build */
{
{"trace_syncscan", PGC_USERSET, DEVELOPER_OPTIONS,
gettext_noop("Generate debugging output for synchronized scanning."),
NULL,
GUC_NOT_IN_SAMPLE
},
&trace_syncscan,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
NULL, NULL, NULL
Arrange for large sequential scans to synchronize with each other, so that when multiple backends are scanning the same relation concurrently, each page is (ideally) read only once. Jeff Davis, with review by Heikki and Tom.
2007-06-08 20:23:53 +02:00
},
#endif
Teach tuplesort.c about "top N" sorting, in which only the first N tuples need be returned. We keep a heap of the current best N tuples and sift-up new tuples into it as we scan the input. For M input tuples this means only about M*log(N) comparisons instead of M*log(M), not to mention a lot less workspace when N is small --- avoiding spill-to-disk for large M is actually the most attractive thing about it. Patch includes planner and executor support for invoking this facility in ORDER BY ... LIMIT queries. Greg Stark, with some editorialization by moi.
2007-05-04 03:13:45 +02:00
#ifdef DEBUG_BOUNDED_SORT
/* this is undocumented because not exposed in a standard build */
{
{
"optimize_bounded_sort", PGC_USERSET, QUERY_TUNING_METHOD,
gettext_noop("Enable bounded sorting using heap sort."),
NULL,
GUC_NOT_IN_SAMPLE
},
&optimize_bounded_sort,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
true,
NULL, NULL, NULL
Teach tuplesort.c about "top N" sorting, in which only the first N tuples need be returned. We keep a heap of the current best N tuples and sift-up new tuples into it as we scan the input. For M input tuples this means only about M*log(N) comparisons instead of M*log(M), not to mention a lot less workspace when N is small --- avoiding spill-to-disk for large M is actually the most attractive thing about it. Patch includes planner and executor support for invoking this facility in ORDER BY ... LIMIT queries. Greg Stark, with some editorialization by moi.
2007-05-04 03:13:45 +02:00
},
#endif
Make the 'wal_debug' GUC variable a boolean (rather than an integer), and hide it behind #ifdef WAL_DEBUG blocks.
2004-01-06 18:26:23 +01:00
#ifdef WAL_DEBUG
{
{"wal_debug", PGC_SUSET, DEVELOPER_OPTIONS,
gettext_noop("Emit WAL-related debugging output."),
NULL,
GUC_NOT_IN_SAMPLE
},
&XLOG_DEBUG,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
NULL, NULL, NULL
Make the 'wal_debug' GUC variable a boolean (rather than an integer), and hide it behind #ifdef WAL_DEBUG blocks.
2004-01-06 18:26:23 +01:00
},
#endif
Create a built-in log rotation program, so that we no longer have to recommend that people go get Apache's rotatelogs program. Additional benefits are that configuration is done through GUC, rather than externally, and that the postmaster can monitor the log rotator and restart it after failure (though we certainly hope that won't happen often). Andreas Pflug, some rework by Tom Lane.
2004-08-06 01:32:13 +02:00
{
Whack some sense into the configuration-file-location patch. Refactor code into something reasonably understandable, cause use of the feature to not fail in standalone backends or in EXEC_BACKEND case, fix sloppy guc.c table entries, make the documentation minimally usable.
2004-10-08 03:36:36 +02:00
{"integer_datetimes", PGC_INTERNAL, PRESET_OPTIONS,
Fix up grammar in variable description string.
2005-10-08 22:08:19 +02:00
gettext_noop("Datetimes are integer based."),
Create a built-in log rotation program, so that we no longer have to recommend that people go get Apache's rotatelogs program. Additional benefits are that configuration is done through GUC, rather than externally, and that the postmaster can monitor the log rotator and restart it after failure (though we certainly hope that won't happen often). Andreas Pflug, some rework by Tom Lane.
2004-08-06 01:32:13 +02:00
NULL,
Mark server_encoding and integer_datetimes as GUC_REPORT, per previous proposals by Oliver Jowett. Update documentation.
2004-08-16 04:12:29 +02:00
GUC_REPORT | GUC_NOT_IN_SAMPLE | GUC_DISALLOW_IN_FILE
Create a built-in log rotation program, so that we no longer have to recommend that people go get Apache's rotatelogs program. Additional benefits are that configuration is done through GUC, rather than externally, and that the postmaster can monitor the log rotator and restart it after failure (though we certainly hope that won't happen often). Andreas Pflug, some rework by Tom Lane.
2004-08-06 01:32:13 +02:00
},
&integer_datetimes,
#ifdef HAVE_INT64_TIMESTAMP
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
true,
Create a built-in log rotation program, so that we no longer have to recommend that people go get Apache's rotatelogs program. Additional benefits are that configuration is done through GUC, rather than externally, and that the postmaster can monitor the log rotator and restart it after failure (though we certainly hope that won't happen often). Andreas Pflug, some rework by Tom Lane.
2004-08-06 01:32:13 +02:00
#else
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
Create a built-in log rotation program, so that we no longer have to recommend that people go get Apache's rotatelogs program. Additional benefits are that configuration is done through GUC, rather than externally, and that the postmaster can monitor the log rotator and restart it after failure (though we certainly hope that won't happen often). Andreas Pflug, some rework by Tom Lane.
2004-08-06 01:32:13 +02:00
#endif
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
NULL, NULL, NULL
Create a built-in log rotation program, so that we no longer have to recommend that people go get Apache's rotatelogs program. Additional benefits are that configuration is done through GUC, rather than externally, and that the postmaster can monitor the log rotator and restart it after failure (though we certainly hope that won't happen often). Andreas Pflug, some rework by Tom Lane.
2004-08-06 01:32:13 +02:00
},
Allow kerberos name and username case sensitivity to be specified from postgresql.conf. --------------------------------------------------------------------------- Here's an updated version of the patch, with the following changes: 1) No longer uses "service name" as "application version". It's instead hardcoded as "postgres". It could be argued that this part should be backpatched to 8.0, but it doesn't make a big difference until you can start changing it with GUC / connection parameters. This change only affects kerberos 5, not 4. 2) Now downcases kerberos usernames when the client is running on win32. 3) Adds guc option for "krb_caseins_users" to make the server ignore case mismatch which is required by some KDCs such as Active Directory. Off by default, per discussion with Tom. This change only affects kerberos 5, not 4. 4) Updated so it doesn't conflict with the rendevouz/bonjour patch already in ;-) Magnus Hagander
2005-06-04 22:42:43 +02:00
{
Make it possible to change Kerberos/GSSAPI parameters without restarting the postmaster. They are only used in backend processes, so it's just a matter of re-labeling the GUCs.
2009-01-02 11:33:20 +01:00
{"krb_caseins_users", PGC_SIGHUP, CONN_AUTH_SECURITY,
Add support for GSSAPI authentication. Documentation still being written, will be committed later. Henry B. Hotz and Magnus Hagander
2007-07-10 15:14:22 +02:00
gettext_noop("Sets whether Kerberos and GSSAPI user names should be treated as case-insensitive."),
Allow kerberos name and username case sensitivity to be specified from postgresql.conf. --------------------------------------------------------------------------- Here's an updated version of the patch, with the following changes: 1) No longer uses "service name" as "application version". It's instead hardcoded as "postgres". It could be argued that this part should be backpatched to 8.0, but it doesn't make a big difference until you can start changing it with GUC / connection parameters. This change only affects kerberos 5, not 4. 2) Now downcases kerberos usernames when the client is running on win32. 3) Adds guc option for "krb_caseins_users" to make the server ignore case mismatch which is required by some KDCs such as Active Directory. Off by default, per discussion with Tom. This change only affects kerberos 5, not 4. 4) Updated so it doesn't conflict with the rendevouz/bonjour patch already in ;-) Magnus Hagander
2005-06-04 22:42:43 +02:00
NULL
},
&pg_krb_caseins_users,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
NULL, NULL, NULL
Allow kerberos name and username case sensitivity to be specified from postgresql.conf. --------------------------------------------------------------------------- Here's an updated version of the patch, with the following changes: 1) No longer uses "service name" as "application version". It's instead hardcoded as "postgres". It could be argued that this part should be backpatched to 8.0, but it doesn't make a big difference until you can start changing it with GUC / connection parameters. This change only affects kerberos 5, not 4. 2) Now downcases kerberos usernames when the client is running on win32. 3) Adds guc option for "krb_caseins_users" to make the server ignore case mismatch which is required by some KDCs such as Active Directory. Off by default, per discussion with Tom. This change only affects kerberos 5, not 4. 4) Updated so it doesn't conflict with the rendevouz/bonjour patch already in ;-) Magnus Hagander
2005-06-04 22:42:43 +02:00
},
Add E'' syntax so eventually normal strings can treat backslashes literally. Add GUC variables: "escape_string_warning" - warn about backslashes in non-E strings "escape_string_syntax" - supports E'' syntax? "standard_compliant_strings" - treats backslashes literally in '' Update code to use E'' when escapes are used.
2005-06-26 05:04:37 +02:00
{
{"escape_string_warning", PGC_USERSET, COMPAT_OPTIONS_PREVIOUS,
Code review for escape-strings patch. Sync psql and plpgsql lexers with main, avoid using a SQL-defined SQLSTATE for what is most definitely not a SQL-compatible error condition, fix documentation omissions, adhere to message style guidelines, don't use two GUC_REPORT variables when one is sufficient. Nothing done about pg_dump issues.
2005-06-26 21:16:07 +02:00
gettext_noop("Warn about backslash escapes in ordinary string literals."),
Add E'' syntax so eventually normal strings can treat backslashes literally. Add GUC variables: "escape_string_warning" - warn about backslashes in non-E strings "escape_string_syntax" - supports E'' syntax? "standard_compliant_strings" - treats backslashes literally in '' Update code to use E'' when escapes are used.
2005-06-26 05:04:37 +02:00
NULL
},
&escape_string_warning,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
true,
NULL, NULL, NULL
Add E'' syntax so eventually normal strings can treat backslashes literally. Add GUC variables: "escape_string_warning" - warn about backslashes in non-E strings "escape_string_syntax" - supports E'' syntax? "standard_compliant_strings" - treats backslashes literally in '' Update code to use E'' when escapes are used.
2005-06-26 05:04:37 +02:00
},
{
Enable standard_conforming_strings to be turned on. Kevin Grittner
2006-03-06 20:49:20 +01:00
{"standard_conforming_strings", PGC_USERSET, COMPAT_OPTIONS_PREVIOUS,
Various improvements to the GUC description strings. Punctuate and capitalize the strings like sentences. Remove unnecessarily specific descriptions of the units used by GUC variables, since we now allow any reasonable unit to be specified.
2006-12-06 18:35:49 +01:00
gettext_noop("Causes '...' strings to treat backslashes literally."),
Add E'' syntax so eventually normal strings can treat backslashes literally. Add GUC variables: "escape_string_warning" - warn about backslashes in non-E strings "escape_string_syntax" - supports E'' syntax? "standard_compliant_strings" - treats backslashes literally in '' Update code to use E'' when escapes are used.
2005-06-26 05:04:37 +02:00
NULL,
Enable standard_conforming_strings to be turned on. Kevin Grittner
2006-03-06 20:49:20 +01:00
GUC_REPORT
Add E'' syntax so eventually normal strings can treat backslashes literally. Add GUC variables: "escape_string_warning" - warn about backslashes in non-E strings "escape_string_syntax" - supports E'' syntax? "standard_compliant_strings" - treats backslashes literally in '' Update code to use E'' when escapes are used.
2005-06-26 05:04:37 +02:00
},
Change standard_compliant_strings to standard_conforming_strings.
2005-08-15 00:19:50 +02:00
&standard_conforming_strings,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
true,
NULL, NULL, NULL
Add E'' syntax so eventually normal strings can treat backslashes literally. Add GUC variables: "escape_string_warning" - warn about backslashes in non-E strings "escape_string_syntax" - supports E'' syntax? "standard_compliant_strings" - treats backslashes literally in '' Update code to use E'' when escapes are used.
2005-06-26 05:04:37 +02:00
},
Add a GUC variable "synchronize_seqscans" to allow clients to disable the new synchronized-scanning behavior, and make pg_dump disable sync scans so that it will reliably preserve row ordering. Per recent discussions.
2008-01-30 19:35:55 +01:00
{
{"synchronize_seqscans", PGC_USERSET, COMPAT_OPTIONS_PREVIOUS,
gettext_noop("Enable synchronized sequential scans."),
NULL
},
&synchronize_seqscans,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
true,
NULL, NULL, NULL
Add a GUC variable "synchronize_seqscans" to allow clients to disable the new synchronized-scanning behavior, and make pg_dump disable sync scans so that it will reliably preserve row ordering. Per recent discussions.
2008-01-30 19:35:55 +01:00
},
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
{
Add new GUC categories corresponding to sections in docs, and move description for vacuum_defer_cleanup_age to the correct category. Sections in postgresql.conf are also sorted in the same order with docs. Per gripe by Fujii Masao, suggestion by Heikki Linnakangas, and patch by me.
2010-06-15 09:52:11 +02:00
{"archive_mode", PGC_POSTMASTER, WAL_ARCHIVING,
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
gettext_noop("Allows archiving of WAL files using archive_command."),
NULL
},
&XLogArchiveMode,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
NULL, NULL, NULL
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
},
Minor improvements in backup and recovery: - create a separate archive_mode GUC, on which archive_command is dependent - %r option in recovery.conf sends last restartpoint to recovery command - %r used in pg_standby, updated README - minor other code cleanup in pg_standby - doc on Warm Standby now mentions pg_standby and %r - log_restartpoints recovery option emits LOG message at each restartpoint - end of recovery now displays last transaction end time, as requested by Warren Little; also shown at each restartpoint - restart archiver if needed to carry away WAL files at shutdown Simon Riggs
2007-09-27 00:36:30 +02:00
Allow read only connections during recovery, known as Hot Standby. Enabled by recovery_connections = on (default) and forcing archive recovery using a recovery.conf. Recovery processing now emulates the original transactions as they are replayed, providing full locking and MVCC behaviour for read only queries. Recovery must enter consistent state before connections are allowed, so there is a delay, typically short, before connections succeed. Replay of recovering transactions can conflict and in some cases deadlock with queries during recovery; these result in query cancellation after max_standby_delay seconds have expired. Infrastructure changes have minor effects on normal running, though introduce four new types of WAL record. New test mode "make standbycheck" allows regression tests of static command behaviour on a standby server while in recovery. Typical and extreme dynamic behaviours have been checked via code inspection and manual testing. Few port specific behaviours have been utilised, though primary testing has been on Linux only so far. This commit is the basic patch. Additional changes will follow in this release to enhance some aspects of behaviour, notably improved handling of conflicts, deadlock detection and query cancellation. Changes to VACUUM FULL are also required. Simon Riggs, with significant and lengthy review by Heikki Linnakangas, including streamlined redesign of snapshot creation and two-phase commit. Important contributions from Florian Pflug, Mark Kirkwood, Merlin Moncure, Greg Stark, Gianni Ciolli, Gabriele Bartolini, Hannu Krosing, Robert Haas, Tatsuo Ishii, Hiroyuki Yamada plus support and feedback from many other community members.
2009-12-19 02:32:45 +01:00
{
Reclassify replication-related GUC variables as "master" and "standby". Per discussion, this structure seems more understandable than what was there before. Make config.sgml and postgresql.conf.sample agree. In passing do a bit of editorial work on the variable descriptions.
2011-07-07 21:10:32 +02:00
{"hot_standby", PGC_POSTMASTER, REPLICATION_STANDBY,
Rename the parameter recovery_connections to hot_standby, to reduce possible confusion with streaming-replication settings. Also, change its default value to "off", because of concern about executing new and poorly-tested code during ordinary non-replicating operation. Per discussion. In passing do some minor editing of related documentation.
2010-04-29 23:36:19 +02:00
gettext_noop("Allows connections and queries during recovery."),
Allow read only connections during recovery, known as Hot Standby. Enabled by recovery_connections = on (default) and forcing archive recovery using a recovery.conf. Recovery processing now emulates the original transactions as they are replayed, providing full locking and MVCC behaviour for read only queries. Recovery must enter consistent state before connections are allowed, so there is a delay, typically short, before connections succeed. Replay of recovering transactions can conflict and in some cases deadlock with queries during recovery; these result in query cancellation after max_standby_delay seconds have expired. Infrastructure changes have minor effects on normal running, though introduce four new types of WAL record. New test mode "make standbycheck" allows regression tests of static command behaviour on a standby server while in recovery. Typical and extreme dynamic behaviours have been checked via code inspection and manual testing. Few port specific behaviours have been utilised, though primary testing has been on Linux only so far. This commit is the basic patch. Additional changes will follow in this release to enhance some aspects of behaviour, notably improved handling of conflicts, deadlock detection and query cancellation. Changes to VACUUM FULL are also required. Simon Riggs, with significant and lengthy review by Heikki Linnakangas, including streamlined redesign of snapshot creation and two-phase commit. Important contributions from Florian Pflug, Mark Kirkwood, Merlin Moncure, Greg Stark, Gianni Ciolli, Gabriele Bartolini, Hannu Krosing, Robert Haas, Tatsuo Ishii, Hiroyuki Yamada plus support and feedback from many other community members.
2009-12-19 02:32:45 +01:00
NULL
},
Rename the parameter recovery_connections to hot_standby, to reduce possible confusion with streaming-replication settings. Also, change its default value to "off", because of concern about executing new and poorly-tested code during ordinary non-replicating operation. Per discussion. In passing do some minor editing of related documentation.
2010-04-29 23:36:19 +02:00
&EnableHotStandby,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
NULL, NULL, NULL
Allow read only connections during recovery, known as Hot Standby. Enabled by recovery_connections = on (default) and forcing archive recovery using a recovery.conf. Recovery processing now emulates the original transactions as they are replayed, providing full locking and MVCC behaviour for read only queries. Recovery must enter consistent state before connections are allowed, so there is a delay, typically short, before connections succeed. Replay of recovering transactions can conflict and in some cases deadlock with queries during recovery; these result in query cancellation after max_standby_delay seconds have expired. Infrastructure changes have minor effects on normal running, though introduce four new types of WAL record. New test mode "make standbycheck" allows regression tests of static command behaviour on a standby server while in recovery. Typical and extreme dynamic behaviours have been checked via code inspection and manual testing. Few port specific behaviours have been utilised, though primary testing has been on Linux only so far. This commit is the basic patch. Additional changes will follow in this release to enhance some aspects of behaviour, notably improved handling of conflicts, deadlock detection and query cancellation. Changes to VACUUM FULL are also required. Simon Riggs, with significant and lengthy review by Heikki Linnakangas, including streamlined redesign of snapshot creation and two-phase commit. Important contributions from Florian Pflug, Mark Kirkwood, Merlin Moncure, Greg Stark, Gianni Ciolli, Gabriele Bartolini, Hannu Krosing, Robert Haas, Tatsuo Ishii, Hiroyuki Yamada plus support and feedback from many other community members.
2009-12-19 02:32:45 +01:00
},
Hot Standby feedback for avoidance of cleanup conflicts on standby. Standby optionally sends back information about oldestXmin of queries which is then checked and applied to the WALSender's proc->xmin. GetOldestXmin() is modified slightly to agree with GetSnapshotData(), so that all backends on primary include WALSender within their snapshots. Note this does nothing to change the snapshot xmin on either master or standby. Feedback piggybacks on the standby reply message. vacuum_defer_cleanup_age is no longer used on standby, though parameter still exists on primary, since some use cases still exist. Simon Riggs, review comments from Fujii Masao, Heikki Linnakangas, Robert Haas
2011-02-16 20:29:37 +01:00
{
Reclassify replication-related GUC variables as "master" and "standby". Per discussion, this structure seems more understandable than what was there before. Make config.sgml and postgresql.conf.sample agree. In passing do a bit of editorial work on the variable descriptions.
2011-07-07 21:10:32 +02:00
{"hot_standby_feedback", PGC_SIGHUP, REPLICATION_STANDBY,
Message style tweaks
2011-07-04 23:01:35 +02:00
gettext_noop("Allows feedback from a hot standby to the primary that will avoid query conflicts."),
Hot Standby feedback for avoidance of cleanup conflicts on standby. Standby optionally sends back information about oldestXmin of queries which is then checked and applied to the WALSender's proc->xmin. GetOldestXmin() is modified slightly to agree with GetSnapshotData(), so that all backends on primary include WALSender within their snapshots. Note this does nothing to change the snapshot xmin on either master or standby. Feedback piggybacks on the standby reply message. vacuum_defer_cleanup_age is no longer used on standby, though parameter still exists on primary, since some use cases still exist. Simon Riggs, review comments from Fujii Masao, Heikki Linnakangas, Robert Haas
2011-02-16 20:29:37 +01:00
NULL
},
&hot_standby_feedback,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
NULL, NULL, NULL
Hot Standby feedback for avoidance of cleanup conflicts on standby. Standby optionally sends back information about oldestXmin of queries which is then checked and applied to the WALSender's proc->xmin. GetOldestXmin() is modified slightly to agree with GetSnapshotData(), so that all backends on primary include WALSender within their snapshots. Note this does nothing to change the snapshot xmin on either master or standby. Feedback piggybacks on the standby reply message. vacuum_defer_cleanup_age is no longer used on standby, though parameter still exists on primary, since some use cases still exist. Simon Riggs, review comments from Fujii Masao, Heikki Linnakangas, Robert Haas
2011-02-16 20:29:37 +01:00
},
Make all command-line options of postmaster and postgres the same. See http://archives.postgresql.org/pgsql-hackers/2006-01/msg00151.php for the complete plan.
2006-01-05 11:07:46 +01:00
{
Change allow_system_table_mods to PGC_POSTMASTER, restoring previous behavior.
2006-01-09 11:05:31 +01:00
{"allow_system_table_mods", PGC_POSTMASTER, DEVELOPER_OPTIONS,
pgindent run for 8.2.
2006-10-04 02:30:14 +02:00
gettext_noop("Allows modifications of the structure of system tables."),
NULL,
GUC_NOT_IN_SAMPLE
Make all command-line options of postmaster and postgres the same. See http://archives.postgresql.org/pgsql-hackers/2006-01/msg00151.php for the complete plan.
2006-01-05 11:07:46 +01:00
},
&allowSystemTableMods,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
NULL, NULL, NULL
Make all command-line options of postmaster and postgres the same. See http://archives.postgresql.org/pgsql-hackers/2006-01/msg00151.php for the complete plan.
2006-01-05 11:07:46 +01:00
},
{
{"ignore_system_indexes", PGC_BACKEND, DEVELOPER_OPTIONS,
pgindent run for 8.2.
2006-10-04 02:30:14 +02:00
gettext_noop("Disables reading from system indexes."),
gettext_noop("It does not prevent updating the indexes, so it is safe "
"to use. The worst consequence is slowness."),
GUC_NOT_IN_SAMPLE
Make all command-line options of postmaster and postgres the same. See http://archives.postgresql.org/pgsql-hackers/2006-01/msg00151.php for the complete plan.
2006-01-05 11:07:46 +01:00
},
&IgnoreSystemIndexes,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
NULL, NULL, NULL
Make all command-line options of postmaster and postgres the same. See http://archives.postgresql.org/pgsql-hackers/2006-01/msg00151.php for the complete plan.
2006-01-05 11:07:46 +01:00
},
Add large object access control. A new system catalog pg_largeobject_metadata manages ownership and access privileges of large objects. KaiGai Kohei, reviewed by Jaime Casanova.
2009-12-11 04:34:57 +01:00
{
{"lo_compat_privileges", PGC_SUSET, COMPAT_OPTIONS_PREVIOUS,
Message tuning
2010-03-21 01:17:59 +01:00
gettext_noop("Enables backward compatibility mode for privilege checks on large objects."),
More cleanups for the recent large object permissions patch. Rewrite or adjust various comments for clarity. Remove one bogus comment that doesn't reflect what the code actually does. Improve the description of the lo_compat_privileges option.
2009-12-21 02:34:11 +01:00
gettext_noop("Skips privilege checks when reading or modifying large objects, "
pgindent run for 9.0
2010-02-26 03:01:40 +01:00
"for compatibility with PostgreSQL releases prior to 9.0.")
Add large object access control. A new system catalog pg_largeobject_metadata manages ownership and access privileges of large objects. KaiGai Kohei, reviewed by Jaime Casanova.
2009-12-11 04:34:57 +01:00
},
&lo_compat_privileges,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
NULL, NULL, NULL
Add large object access control. A new system catalog pg_largeobject_metadata manages ownership and access privileges of large objects. KaiGai Kohei, reviewed by Jaime Casanova.
2009-12-11 04:34:57 +01:00
},
Add options to force quoting of all identifiers. I've added a quote_all_identifiers GUC which affects the behavior of the backend, and a --quote-all-identifiers argument to pg_dump and pg_dumpall which sets the GUC and also affects the quoting done internally by those applications. Design by Tom Lane; review by Alex Hunsaker; in response to bug #5488 filed by Hartmut Goebel.
2010-07-22 03:22:35 +02:00
{
{"quote_all_identifiers", PGC_USERSET, COMPAT_OPTIONS_PREVIOUS,
gettext_noop("When generating SQL fragments, quote all identifiers."),
NULL,
},
&quote_all_identifiers,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
false,
NULL, NULL, NULL
Add options to force quoting of all identifiers. I've added a quote_all_identifiers GUC which affects the behavior of the backend, and a --quote-all-identifiers argument to pg_dump and pg_dumpall which sets the GUC and also affects the quoting done internally by those applications. Design by Tom Lane; review by Alex Hunsaker; in response to bug #5488 filed by Hartmut Goebel.
2010-07-22 03:22:35 +02:00
},
Add add_missing_from GUC variable. Nigel J. Andrews
2003-06-12 00:13:22 +02:00
/* End-of-list marker */
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
{NULL, 0, 0, NULL, NULL}, NULL, false, NULL, NULL, NULL
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
}
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
};
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
static struct config_int ConfigureNamesInt[] =
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
{
Implement archive_timeout feature to force xlog file switches to occur no more than N seconds apart. This allows a simple, if not very high performance, means of guaranteeing that a PITR archive is no more than N seconds behind real time. Also make pg_current_xlog_location return the WAL Write pointer, add pg_current_xlog_insert_location to return the Insert pointer, and fix pg_xlogfile_name_offset to return its results as a two-element record instead of a smashed-together string, as per recent discussion. Simon Riggs
2006-08-18 01:04:10 +02:00
{
Add new GUC categories corresponding to sections in docs, and move description for vacuum_defer_cleanup_age to the correct category. Sections in postgresql.conf are also sorted in the same order with docs. Per gripe by Fujii Masao, suggestion by Heikki Linnakangas, and patch by me.
2010-06-15 09:52:11 +02:00
{"archive_timeout", PGC_SIGHUP, WAL_ARCHIVING,
pgindent run for 8.2.
2006-10-04 02:30:14 +02:00
gettext_noop("Forces a switch to the next xlog file if a "
"new file has not been started within N seconds."),
NULL,
GUC_UNIT_S
Implement archive_timeout feature to force xlog file switches to occur no more than N seconds apart. This allows a simple, if not very high performance, means of guaranteeing that a PITR archive is no more than N seconds behind real time. Also make pg_current_xlog_location return the WAL Write pointer, add pg_current_xlog_insert_location to return the Insert pointer, and fix pg_xlogfile_name_offset to return its results as a two-element record instead of a smashed-together string, as per recent discussion. Simon Riggs
2006-08-18 01:04:10 +02:00
},
&XLogArchiveTimeout,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
0, 0, INT_MAX,
NULL, NULL, NULL
Implement archive_timeout feature to force xlog file switches to occur no more than N seconds apart. This allows a simple, if not very high performance, means of guaranteeing that a PITR archive is no more than N seconds behind real time. Also make pg_current_xlog_location return the WAL Write pointer, add pg_current_xlog_insert_location to return the Insert pointer, and fix pg_xlogfile_name_offset to return its results as a two-element record instead of a smashed-together string, as per recent discussion. Simon Riggs
2006-08-18 01:04:10 +02:00
},
Make all command-line options of postmaster and postgres the same. See http://archives.postgresql.org/pgsql-hackers/2006-01/msg00151.php for the complete plan.
2006-01-05 11:07:46 +01:00
{
{"post_auth_delay", PGC_BACKEND, DEVELOPER_OPTIONS,
pgindent run for 8.2.
2006-10-04 02:30:14 +02:00
gettext_noop("Waits N seconds on connection startup after authentication."),
gettext_noop("This allows attaching a debugger to the process."),
GUC_NOT_IN_SAMPLE | GUC_UNIT_S
Make all command-line options of postmaster and postgres the same. See http://archives.postgresql.org/pgsql-hackers/2006-01/msg00151.php for the complete plan.
2006-01-05 11:07:46 +01:00
},
&PostAuthDelay,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
0, 0, INT_MAX,
NULL, NULL, NULL
Make all command-line options of postmaster and postgres the same. See http://archives.postgresql.org/pgsql-hackers/2006-01/msg00151.php for the complete plan.
2006-01-05 11:07:46 +01:00
},
Instead of having a configure-time DEFAULT_ATTSTATTARGET, store -1 in attstattarget to indicate 'use the default'. The default is now a GUC variable default_statistics_target, and so may be changed on the fly. Along the way we gain the ability to have pg_dump dump the per-column statistics target when it's not the default. Patch by Neil Conway, with some kibitzing from Tom Lane.
2002-07-31 19:19:54 +02:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"default_statistics_target", PGC_USERSET, QUERY_TUNING_OTHER,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Sets the default statistics target."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
gettext_noop("This applies to table columns that have not had a "
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
"column-specific target set via ALTER TABLE SET STATISTICS.")
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
&default_statistics_target,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
100, 1, 10000,
NULL, NULL, NULL
Instead of having a configure-time DEFAULT_ATTSTATTARGET, store -1 in attstattarget to indicate 'use the default'. The default is now a GUC variable default_statistics_target, and so may be changed on the fly. Along the way we gain the ability to have pg_dump dump the per-column statistics target when it's not the default. Patch by Neil Conway, with some kibitzing from Tom Lane.
2002-07-31 19:19:54 +02:00
},
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"from_collapse_limit", PGC_USERSET, QUERY_TUNING_OTHER,
Various improvements to the GUC description strings. Punctuate and capitalize the strings like sentences. Remove unnecessarily specific descriptions of the units used by GUC variables, since we now allow any reasonable unit to be specified.
2006-12-06 18:35:49 +01:00
gettext_noop("Sets the FROM-list size beyond which subqueries "
"are not collapsed."),
Message editing: remove gratuitous variations in message wording, standardize terms, add some clarifications, fix some untranslatable attempts at dynamic message building.
2003-09-25 08:58:07 +02:00
gettext_noop("The planner will merge subqueries into upper "
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
"queries if the resulting FROM list would have no more than "
Message editing: remove gratuitous variations in message wording, standardize terms, add some clarifications, fix some untranslatable attempts at dynamic message building.
2003-09-25 08:58:07 +02:00
"this many items.")
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
&from_collapse_limit,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
8, 1, INT_MAX,
NULL, NULL, NULL
Allow the planner to collapse explicit inner JOINs together, rather than necessarily following the JOIN syntax to develop the query plan. The old behavior is still available by setting GUC variable JOIN_COLLAPSE_LIMIT to 1. Also create a GUC variable FROM_COLLAPSE_LIMIT to control the similar decision about when to collapse sub-SELECT lists into their parent lists. (This behavior existed already, but the limit was always GEQO_THRESHOLD/2; now it's separately adjustable.)
2003-01-26 00:10:30 +01:00
},
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"join_collapse_limit", PGC_USERSET, QUERY_TUNING_OTHER,
Various improvements to the GUC description strings. Punctuate and capitalize the strings like sentences. Remove unnecessarily specific descriptions of the units used by GUC variables, since we now allow any reasonable unit to be specified.
2006-12-06 18:35:49 +01:00
gettext_noop("Sets the FROM-list size beyond which JOIN "
"constructs are not flattened."),
Teach planner how to rearrange join order for some classes of OUTER JOIN. Per my recent proposal. I ended up basing the implementation on the existing mechanism for enforcing valid join orders of IN joins --- the rules for valid outer-join orders are somewhat similar.
2005-12-20 03:30:36 +01:00
gettext_noop("The planner will flatten explicit JOIN "
Various improvements to the GUC description strings. Punctuate and capitalize the strings like sentences. Remove unnecessarily specific descriptions of the units used by GUC variables, since we now allow any reasonable unit to be specified.
2006-12-06 18:35:49 +01:00
"constructs into lists of FROM items whenever a "
"list of no more than this many items would result.")
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
&join_collapse_limit,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
8, 1, INT_MAX,
NULL, NULL, NULL
Allow the planner to collapse explicit inner JOINs together, rather than necessarily following the JOIN syntax to develop the query plan. The old behavior is still available by setting GUC variable JOIN_COLLAPSE_LIMIT to 1. Also create a GUC variable FROM_COLLAPSE_LIMIT to control the similar decision about when to collapse sub-SELECT lists into their parent lists. (This behavior existed already, but the limit was always GEQO_THRESHOLD/2; now it's separately adjustable.)
2003-01-26 00:10:30 +01:00
},
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"geqo_threshold", PGC_USERSET, QUERY_TUNING_GEQO,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Sets the threshold of FROM items beyond which GEQO is used."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL
},
&geqo_threshold,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
12, 2, INT_MAX,
NULL, NULL, NULL
Another pgindent run. Fixes enum indenting, and improves #endif spacing. Also adds space for one-line comments.
2001-10-28 07:26:15 +01:00
},
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Revise GEQO planner to make use of some heuristic knowledge about SQL, namely that it's good to join where there are join clauses rather than where there are not. Also enable it to generate bushy plans at need, so that it doesn't fail in the presence of multiple IN clauses containing sub-joins. These changes appear to improve the behavior enough that we can substantially reduce the default pool size and generations count, thereby decreasing the runtime, and yet get as good or better plans as we were getting in 7.4. Consequently, adjust the default GEQO parameters. I also modified the way geqo_effort is used so that it affects both population size and number of generations; it's now useful as a single control to adjust the GEQO runtime-vs-plan-quality tradeoff. Bump geqo_threshold to 12, since even with these changes GEQO seems to be slower than the regular planner at 11 relations.
2004-01-24 00:54:21 +01:00
{"geqo_effort", PGC_USERSET, QUERY_TUNING_GEQO,
gettext_noop("GEQO: effort is used to set the default for other GEQO parameters."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL
},
Revise GEQO planner to make use of some heuristic knowledge about SQL, namely that it's good to join where there are join clauses rather than where there are not. Also enable it to generate bushy plans at need, so that it doesn't fail in the presence of multiple IN clauses containing sub-joins. These changes appear to improve the behavior enough that we can substantially reduce the default pool size and generations count, thereby decreasing the runtime, and yet get as good or better plans as we were getting in 7.4. Consequently, adjust the default GEQO parameters. I also modified the way geqo_effort is used so that it affects both population size and number of generations; it's now useful as a single control to adjust the GEQO runtime-vs-plan-quality tradeoff. Bump geqo_threshold to 12, since even with these changes GEQO seems to be slower than the regular planner at 11 relations.
2004-01-24 00:54:21 +01:00
&Geqo_effort,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
DEFAULT_GEQO_EFFORT, MIN_GEQO_EFFORT, MAX_GEQO_EFFORT,
NULL, NULL, NULL
Revise GEQO planner to make use of some heuristic knowledge about SQL, namely that it's good to join where there are join clauses rather than where there are not. Also enable it to generate bushy plans at need, so that it doesn't fail in the presence of multiple IN clauses containing sub-joins. These changes appear to improve the behavior enough that we can substantially reduce the default pool size and generations count, thereby decreasing the runtime, and yet get as good or better plans as we were getting in 7.4. Consequently, adjust the default GEQO parameters. I also modified the way geqo_effort is used so that it affects both population size and number of generations; it's now useful as a single control to adjust the GEQO runtime-vs-plan-quality tradeoff. Bump geqo_threshold to 12, since even with these changes GEQO seems to be slower than the regular planner at 11 relations.
2004-01-24 00:54:21 +01:00
},
{
{"geqo_pool_size", PGC_USERSET, QUERY_TUNING_GEQO,
gettext_noop("GEQO: number of individuals in the population."),
gettext_noop("Zero selects a suitable default value.")
},
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
&Geqo_pool_size,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
0, 0, INT_MAX,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"geqo_generations", PGC_USERSET, QUERY_TUNING_GEQO,
Repair error apparently introduced in the initial coding of GUC: the default value for geqo_effort is supposed to be 40, not 1. The actual 'genetic' component of the GEQO algorithm has been practically disabled since 7.1 because of this mistake. Improve documentation while at it.
2004-01-22 00:33:34 +01:00
gettext_noop("GEQO: number of iterations of the algorithm."),
Revise GEQO planner to make use of some heuristic knowledge about SQL, namely that it's good to join where there are join clauses rather than where there are not. Also enable it to generate bushy plans at need, so that it doesn't fail in the presence of multiple IN clauses containing sub-joins. These changes appear to improve the behavior enough that we can substantially reduce the default pool size and generations count, thereby decreasing the runtime, and yet get as good or better plans as we were getting in 7.4. Consequently, adjust the default GEQO parameters. I also modified the way geqo_effort is used so that it affects both population size and number of generations; it's now useful as a single control to adjust the GEQO runtime-vs-plan-quality tradeoff. Bump geqo_threshold to 12, since even with these changes GEQO seems to be slower than the regular planner at 11 relations.
2004-01-24 00:54:21 +01:00
gettext_noop("Zero selects a suitable default value.")
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
&Geqo_generations,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
0, 0, INT_MAX,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
{
Make deadlock_timeout PGC_SUSET rather than PGC_SIGHUP. This allows deadlock_timeout to be reduced for transactions that are particularly likely to be involved in a deadlock, thus detecting it more quickly. It is also potentially useful as a poor-man's deadlock priority mechanism: a transaction with a high deadlock_timeout is less likely to be chosen as the victim than one with a low deadlock_timeout. Since that could be used to game the system, we make this PGC_SUSET rather than PGC_USERSET. At some point, it might be worth thinking about a more explicit priority mechanism, since using this is far from fool-proof. But let's see whether there's enough use case to justify the additional work before we go down that route. Noah Misch, reviewed by Shigeru Hanada
2011-06-22 04:32:30 +02:00
/* This is PGC_SUSET to prevent hiding from log_lock_waits. */
{"deadlock_timeout", PGC_SUSET, LOCK_MANAGEMENT,
Revert addition of units to GUC descriptions; doesn't affect postgresql.conf.
2009-04-08 01:27:34 +02:00
gettext_noop("Sets the time to wait on a lock before checking for deadlock."),
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
NULL,
GUC_UNIT_MS
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
&DeadlockTimeout,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
1000, 1, INT_MAX,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
Replace max_standby_delay with two parameters, max_standby_archive_delay and max_standby_streaming_delay, and revise the implementation to avoid assuming that timestamps found in WAL records can meaningfully be compared to clock time on the standby server. Instead, the delay limits are compared to the elapsed time since we last obtained a new WAL segment from archive or since we were last "caught up" to WAL data arriving via streaming replication. This avoids problems with clock skew between primary and standby, as well as other corner cases that the original coding would misbehave in, such as the primary server having significant idle time between transactions. Per my complaint some time ago and considerable ensuing discussion. Do some desultory editing on the hot standby documentation, too.
2010-07-03 22:43:58 +02:00
{
Reclassify replication-related GUC variables as "master" and "standby". Per discussion, this structure seems more understandable than what was there before. Make config.sgml and postgresql.conf.sample agree. In passing do a bit of editorial work on the variable descriptions.
2011-07-07 21:10:32 +02:00
{"max_standby_archive_delay", PGC_SIGHUP, REPLICATION_STANDBY,
Replace max_standby_delay with two parameters, max_standby_archive_delay and max_standby_streaming_delay, and revise the implementation to avoid assuming that timestamps found in WAL records can meaningfully be compared to clock time on the standby server. Instead, the delay limits are compared to the elapsed time since we last obtained a new WAL segment from archive or since we were last "caught up" to WAL data arriving via streaming replication. This avoids problems with clock skew between primary and standby, as well as other corner cases that the original coding would misbehave in, such as the primary server having significant idle time between transactions. Per my complaint some time ago and considerable ensuing discussion. Do some desultory editing on the hot standby documentation, too.
2010-07-03 22:43:58 +02:00
gettext_noop("Sets the maximum delay before canceling queries when a hot standby server is processing archived WAL data."),
NULL,
GUC_UNIT_MS
},
&max_standby_archive_delay,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
30 * 1000, -1, INT_MAX,
NULL, NULL, NULL
Replace max_standby_delay with two parameters, max_standby_archive_delay and max_standby_streaming_delay, and revise the implementation to avoid assuming that timestamps found in WAL records can meaningfully be compared to clock time on the standby server. Instead, the delay limits are compared to the elapsed time since we last obtained a new WAL segment from archive or since we were last "caught up" to WAL data arriving via streaming replication. This avoids problems with clock skew between primary and standby, as well as other corner cases that the original coding would misbehave in, such as the primary server having significant idle time between transactions. Per my complaint some time ago and considerable ensuing discussion. Do some desultory editing on the hot standby documentation, too.
2010-07-03 22:43:58 +02:00
},
{
Reclassify replication-related GUC variables as "master" and "standby". Per discussion, this structure seems more understandable than what was there before. Make config.sgml and postgresql.conf.sample agree. In passing do a bit of editorial work on the variable descriptions.
2011-07-07 21:10:32 +02:00
{"max_standby_streaming_delay", PGC_SIGHUP, REPLICATION_STANDBY,
Replace max_standby_delay with two parameters, max_standby_archive_delay and max_standby_streaming_delay, and revise the implementation to avoid assuming that timestamps found in WAL records can meaningfully be compared to clock time on the standby server. Instead, the delay limits are compared to the elapsed time since we last obtained a new WAL segment from archive or since we were last "caught up" to WAL data arriving via streaming replication. This avoids problems with clock skew between primary and standby, as well as other corner cases that the original coding would misbehave in, such as the primary server having significant idle time between transactions. Per my complaint some time ago and considerable ensuing discussion. Do some desultory editing on the hot standby documentation, too.
2010-07-03 22:43:58 +02:00
gettext_noop("Sets the maximum delay before canceling queries when a hot standby server is processing streamed WAL data."),
NULL,
GUC_UNIT_MS
},
&max_standby_streaming_delay,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
30 * 1000, -1, INT_MAX,
NULL, NULL, NULL
Replace max_standby_delay with two parameters, max_standby_archive_delay and max_standby_streaming_delay, and revise the implementation to avoid assuming that timestamps found in WAL records can meaningfully be compared to clock time on the standby server. Instead, the delay limits are compared to the elapsed time since we last obtained a new WAL segment from archive or since we were last "caught up" to WAL data arriving via streaming replication. This avoids problems with clock skew between primary and standby, as well as other corner cases that the original coding would misbehave in, such as the primary server having significant idle time between transactions. Per my complaint some time ago and considerable ensuing discussion. Do some desultory editing on the hot standby documentation, too.
2010-07-03 22:43:58 +02:00
},
Send status updates back from standby server to master, indicating how far the standby has written, flushed, and applied the WAL. At the moment, this is for informational purposes only, the values are only shown in pg_stat_replication system view, but in the future they will also be needed for synchronous replication. Extracted from Simon riggs' synchronous replication patch by Robert Haas, with some tweaking by me.
2011-02-10 20:00:29 +01:00
{
Reclassify replication-related GUC variables as "master" and "standby". Per discussion, this structure seems more understandable than what was there before. Make config.sgml and postgresql.conf.sample agree. In passing do a bit of editorial work on the variable descriptions.
2011-07-07 21:10:32 +02:00
{"wal_receiver_status_interval", PGC_SIGHUP, REPLICATION_STANDBY,
Message style tweaks
2011-07-04 23:01:35 +02:00
gettext_noop("Sets the maximum interval between WAL receiver status reports to the primary."),
Send status updates back from standby server to master, indicating how far the standby has written, flushed, and applied the WAL. At the moment, this is for informational purposes only, the values are only shown in pg_stat_replication system view, but in the future they will also be needed for synchronous replication. Extracted from Simon riggs' synchronous replication patch by Robert Haas, with some tweaking by me.
2011-02-10 20:00:29 +01:00
NULL,
GUC_UNIT_S
},
&wal_receiver_status_interval,
pgindent run before PG 9.1 beta 1.
2011-04-10 17:42:00 +02:00
10, 0, INT_MAX / 1000,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
NULL, NULL, NULL
Send status updates back from standby server to master, indicating how far the standby has written, flushed, and applied the WAL. At the moment, this is for informational purposes only, the values are only shown in pg_stat_replication system view, but in the future they will also be needed for synchronous replication. Extracted from Simon riggs' synchronous replication patch by Robert Haas, with some tweaking by me.
2011-02-10 20:00:29 +01:00
},
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"max_connections", PGC_POSTMASTER, CONN_AUTH_SETTINGS,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Sets the maximum number of concurrent connections."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL
},
Add a multi-worker capability to autovacuum. This allows multiple worker processes to be running simultaneously. Also, now autovacuum processes do not count towards the max_connections limit; they are counted separately from regular processes, and are limited by the new GUC variable autovacuum_max_workers. The launcher now has intelligence to launch workers on each database every autovacuum_naptime seconds, limited only on the max amount of worker slots available. Also, the global worker I/O utilization is limited by the vacuum cost-based delay feature. Workers are "balanced" so that the total I/O consumption does not exceed the established limit. This part of the patch was contributed by ITAGAKI Takahiro. Per discussion.
2007-04-16 20:30:04 +02:00
&MaxConnections,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
100, 1, MAX_BACKENDS,
check_maxconnections, assign_maxconnections, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
Install infrastructure for shared-memory free space map. Doesn't actually do anything yet, but it has the necessary connections to initialization and so forth. Make some gestures towards allowing number of blocks in a relation to be BlockNumber, ie, unsigned int, rather than signed int. (I doubt I got all the places that are sloppy about it, yet.) On the way, replace the hardwired NLOCKS_PER_XACT fudge factor with a GUC variable.
2001-06-28 01:31:40 +02:00
This patch reserves the last superuser_reserved_connections slots for connections by the superuser only. This patch replaces the last patch I sent a couple of days ago. It closes a connection that has not been authorised by a superuser if it would leave less than the GUC variable ReservedBackends (superuser_reserved_connections in postgres.conf) backend process slots free in the SISeg. This differs to the first patch which only reserved the last ReservedBackends slots in the procState array. This has made the free slot test more expensive due to the use of a lock. After thinking about a comment on the first patch I've also made it a fatal error if the number of reserved slots is not less than the maximum number of connections. Nigel J. Andrews
2002-08-29 23:02:12 +02:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"superuser_reserved_connections", PGC_POSTMASTER, CONN_AUTH_SETTINGS,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Sets the number of connection slots reserved for superusers."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL
},
&ReservedBackends,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
3, 0, MAX_BACKENDS,
NULL, NULL, NULL
This patch reserves the last superuser_reserved_connections slots for connections by the superuser only. This patch replaces the last patch I sent a couple of days ago. It closes a connection that has not been authorised by a superuser if it would leave less than the GUC variable ReservedBackends (superuser_reserved_connections in postgres.conf) backend process slots free in the SISeg. This differs to the first patch which only reserved the last ReservedBackends slots in the procState array. This has made the free slot test more expensive due to the use of a lock. After thinking about a comment on the first patch I've also made it a fatal error if the number of reserved slots is not less than the maximum number of connections. Nigel J. Andrews
2002-08-29 23:02:12 +02:00
},
Include the backend ID in the relpath of temporary relations. This allows us to reliably remove all leftover temporary relation files on cluster startup without reference to system catalogs or WAL; therefore, we no longer include temporary relations in XLOG_XACT_COMMIT and XLOG_XACT_ABORT WAL records. Since these changes require including a backend ID in each SharedInvalSmgrMsg, the size of the SharedInvalidationMessage.id field has been reduced from two bytes to one, and the maximum number of connections has been reduced from INT_MAX / 4 to 2^23-1. It would be possible to remove these restrictions by increasing the size of SharedInvalidationMessage by 4 bytes, but right now that doesn't seem like a good trade-off. Review by Jaime Casanova and Tom Lane.
2010-08-13 22:10:54 +02:00
/*
* We sometimes multiply the number of shared buffers by two without
* checking for overflow, so we mustn't allow more than INT_MAX / 2.
*/
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"shared_buffers", PGC_POSTMASTER, RESOURCES_MEM,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Sets the number of shared memory buffers used by the server."),
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
NULL,
GUC_UNIT_BLOCKS
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
&NBuffers,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
1024, 16, INT_MAX / 2,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
Install infrastructure for shared-memory free space map. Doesn't actually do anything yet, but it has the necessary connections to initialization and so forth. Make some gestures towards allowing number of blocks in a relation to be BlockNumber, ie, unsigned int, rather than signed int. (I doubt I got all the places that are sloppy about it, yet.) On the way, replace the hardwired NLOCKS_PER_XACT fudge factor with a GUC variable.
2001-06-28 01:31:40 +02:00
Add temp_buffers GUC variable to allow users to determine the size of the local buffer arena for temporary table access.
2005-03-20 00:27:11 +01:00
{
{"temp_buffers", PGC_USERSET, RESOURCES_MEM,
gettext_noop("Sets the maximum number of temporary buffers used by each session."),
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
NULL,
GUC_UNIT_BLOCKS
Add temp_buffers GUC variable to allow users to determine the size of the local buffer arena for temporary table access.
2005-03-20 00:27:11 +01:00
},
&num_temp_buffers,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
1024, 100, INT_MAX / 2,
check_temp_buffers, NULL, NULL
Add temp_buffers GUC variable to allow users to determine the size of the local buffer arena for temporary table access.
2005-03-20 00:27:11 +01:00
},
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"port", PGC_POSTMASTER, CONN_AUTH_SETTINGS,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Sets the TCP port the server listens on."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL
},
&PostPortNumber,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
DEF_PGPORT, 1, 65535,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"unix_socket_permissions", PGC_POSTMASTER, CONN_AUTH_SETTINGS,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Sets the access permissions of the Unix-domain socket."),
Message editing: remove gratuitous variations in message wording, standardize terms, add some clarifications, fix some untranslatable attempts at dynamic message building.
2003-09-25 08:58:07 +02:00
gettext_noop("Unix-domain sockets use the usual Unix file system "
Add a log_file_mode GUC that allows control of the file permissions set on log files created by the syslogger process. In passing, make unix_file_permissions display its value in octal, same as log_file_mode now does. Martin Pihlak
2010-07-17 00:25:51 +02:00
"permission set. The parameter value is expected "
"to be a numeric mode specification in the form "
"accepted by the chmod and umask system calls. "
"(To use the customary octal format the number must "
"start with a 0 (zero).)")
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
&Unix_socket_permissions,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
0777, 0000, 0777,
NULL, NULL, show_unix_socket_permissions
Add a log_file_mode GUC that allows control of the file permissions set on log files created by the syslogger process. In passing, make unix_file_permissions display its value in octal, same as log_file_mode now does. Martin Pihlak
2010-07-17 00:25:51 +02:00
},
{
{"log_file_mode", PGC_SIGHUP, LOGGING_WHERE,
gettext_noop("Sets the file permissions for log files."),
gettext_noop("The parameter value is expected "
"to be a numeric mode specification in the form "
"accepted by the chmod and umask system calls. "
"(To use the customary octal format the number must "
"start with a 0 (zero).)")
},
&Log_file_mode,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
0600, 0000, 0777,
NULL, NULL, show_log_file_mode
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
Install infrastructure for shared-memory free space map. Doesn't actually do anything yet, but it has the necessary connections to initialization and so forth. Make some gestures towards allowing number of blocks in a relation to be BlockNumber, ie, unsigned int, rather than signed int. (I doubt I got all the places that are sloppy about it, yet.) On the way, replace the hardwired NLOCKS_PER_XACT fudge factor with a GUC variable.
2001-06-28 01:31:40 +02:00
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Rename SortMem and VacuumMem to work_mem and maintenance_work_mem. Make btree index creation and initial validation of foreign-key constraints use maintenance_work_mem rather than work_mem as their memory limit. Add some code to guc.c to allow these variables to be referenced by their old names in SHOW and SET commands, for backwards compatibility.
2004-02-03 18:34:04 +01:00
{"work_mem", PGC_USERSET, RESOURCES_MEM,
gettext_noop("Sets the maximum memory to be used for query workspaces."),
Wording cleanup for error messages. Also change can't -> cannot. Standard English uses "may", "can", and "might" in different ways: may - permission, "You may borrow my rake." can - ability, "I can lift that log." might - possibility, "It might rain today." Unfortunately, in conversational English, their use is often mixed, as in, "You may use this variable to do X", when in fact, "can" is a better choice. Similarly, "It may crash" is better stated, "It might crash".
2007-02-01 20:10:30 +01:00
gettext_noop("This much memory can be used by each internal "
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
"sort operation and hash table before switching to "
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
"temporary disk files."),
GUC_UNIT_KB
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
Rename SortMem and VacuumMem to work_mem and maintenance_work_mem. Make btree index creation and initial validation of foreign-key constraints use maintenance_work_mem rather than work_mem as their memory limit. Add some code to guc.c to allow these variables to be referenced by their old names in SHOW and SET commands, for backwards compatibility.
2004-02-03 18:34:04 +01:00
&work_mem,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
1024, 64, MAX_KILOBYTES,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Rename SortMem and VacuumMem to work_mem and maintenance_work_mem. Make btree index creation and initial validation of foreign-key constraints use maintenance_work_mem rather than work_mem as their memory limit. Add some code to guc.c to allow these variables to be referenced by their old names in SHOW and SET commands, for backwards compatibility.
2004-02-03 18:34:04 +01:00
{"maintenance_work_mem", PGC_USERSET, RESOURCES_MEM,
Revert addition of units to GUC descriptions; doesn't affect postgresql.conf.
2009-04-08 01:27:34 +02:00
gettext_noop("Sets the maximum memory to be used for maintenance operations."),
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
gettext_noop("This includes operations such as VACUUM and CREATE INDEX."),
GUC_UNIT_KB
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
Rename SortMem and VacuumMem to work_mem and maintenance_work_mem. Make btree index creation and initial validation of foreign-key constraints use maintenance_work_mem rather than work_mem as their memory limit. Add some code to guc.c to allow these variables to be referenced by their old names in SHOW and SET commands, for backwards compatibility.
2004-02-03 18:34:04 +01:00
&maintenance_work_mem,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
16384, 1024, MAX_KILOBYTES,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
Give VACUUM its own GUC parameter for memory usage, rather than piggybacking on SortMem. Add documentation for some recently-added GUC parameters that had so far escaped it.
2001-09-21 05:32:36 +02:00
Add a comment about the initial default value of max_stack_depth.
2010-08-06 16:51:33 +02:00
/*
* We use the hopefully-safely-small value of 100kB as the compiled-in
* default for max_stack_depth. InitializeGUCOptions will increase it if
* possible, depending on the actual platform-specific stack limit.
*/
Replace max_expr_depth parameter with a max_stack_depth parameter that is measured in kilobytes and checked against actual physical execution stack depth, as per my proposal of 30-Dec. This gives us a fairly bulletproof defense against crashing due to runaway recursive functions.
2004-03-24 23:40:29 +01:00
{
{"max_stack_depth", PGC_SUSET, RESOURCES_MEM,
gettext_noop("Sets the maximum stack depth, in kilobytes."),
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
NULL,
GUC_UNIT_KB
Replace max_expr_depth parameter with a max_stack_depth parameter that is measured in kilobytes and checked against actual physical execution stack depth, as per my proposal of 30-Dec. This gives us a fairly bulletproof defense against crashing due to runaway recursive functions.
2004-03-24 23:40:29 +01:00
},
&max_stack_depth,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
100, 100, MAX_KILOBYTES,
check_max_stack_depth, assign_max_stack_depth, NULL
Replace max_expr_depth parameter with a max_stack_depth parameter that is measured in kilobytes and checked against actual physical execution stack depth, as per my proposal of 30-Dec. This gives us a fairly bulletproof defense against crashing due to runaway recursive functions.
2004-03-24 23:40:29 +01:00
},
Add temp_file_limit GUC parameter to constrain temporary file space usage. The limit is enforced against the total amount of temp file space used by each session. Mark Kirkwood, reviewed by Cédric Villemain and Tatsuo Ishii
2011-07-17 20:19:31 +02:00
{
{"temp_file_limit", PGC_SUSET, RESOURCES_DISK,
gettext_noop("Limits the total size of all temp files used by each session."),
gettext_noop("-1 means no limit."),
GUC_UNIT_KB
},
&temp_file_limit,
-1, -1, INT_MAX,
NULL, NULL, NULL
},
Cost based vacuum delay feature. Jan
2004-02-06 20:36:18 +01:00
{
Add new GUC categories corresponding to sections in docs, and move description for vacuum_defer_cleanup_age to the correct category. Sections in postgresql.conf are also sorted in the same order with docs. Per gripe by Fujii Masao, suggestion by Heikki Linnakangas, and patch by me.
2010-06-15 09:52:11 +02:00
{"vacuum_cost_page_hit", PGC_USERSET, RESOURCES_VACUUM_DELAY,
Cost based vacuum delay feature. Jan
2004-02-06 20:36:18 +01:00
gettext_noop("Vacuum cost for a page found in the buffer cache."),
NULL
},
&VacuumCostPageHit,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
1, 0, 10000,
NULL, NULL, NULL
Cost based vacuum delay feature. Jan
2004-02-06 20:36:18 +01:00
},
{
Add new GUC categories corresponding to sections in docs, and move description for vacuum_defer_cleanup_age to the correct category. Sections in postgresql.conf are also sorted in the same order with docs. Per gripe by Fujii Masao, suggestion by Heikki Linnakangas, and patch by me.
2010-06-15 09:52:11 +02:00
{"vacuum_cost_page_miss", PGC_USERSET, RESOURCES_VACUUM_DELAY,
Cost based vacuum delay feature. Jan
2004-02-06 20:36:18 +01:00
gettext_noop("Vacuum cost for a page not found in the buffer cache."),
NULL
},
&VacuumCostPageMiss,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
10, 0, 10000,
NULL, NULL, NULL
Cost based vacuum delay feature. Jan
2004-02-06 20:36:18 +01:00
},
{
Add new GUC categories corresponding to sections in docs, and move description for vacuum_defer_cleanup_age to the correct category. Sections in postgresql.conf are also sorted in the same order with docs. Per gripe by Fujii Masao, suggestion by Heikki Linnakangas, and patch by me.
2010-06-15 09:52:11 +02:00
{"vacuum_cost_page_dirty", PGC_USERSET, RESOURCES_VACUUM_DELAY,
Cost based vacuum delay feature. Jan
2004-02-06 20:36:18 +01:00
gettext_noop("Vacuum cost for a page dirtied by vacuum."),
NULL
},
&VacuumCostPageDirty,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
20, 0, 10000,
NULL, NULL, NULL
Cost based vacuum delay feature. Jan
2004-02-06 20:36:18 +01:00
},
{
Add new GUC categories corresponding to sections in docs, and move description for vacuum_defer_cleanup_age to the correct category. Sections in postgresql.conf are also sorted in the same order with docs. Per gripe by Fujii Masao, suggestion by Heikki Linnakangas, and patch by me.
2010-06-15 09:52:11 +02:00
{"vacuum_cost_limit", PGC_USERSET, RESOURCES_VACUUM_DELAY,
Cost based vacuum delay feature. Jan
2004-02-06 20:36:18 +01:00
gettext_noop("Vacuum cost amount available before napping."),
NULL
},
&VacuumCostLimit,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
200, 1, 10000,
NULL, NULL, NULL
Cost based vacuum delay feature. Jan
2004-02-06 20:36:18 +01:00
},
{
Add new GUC categories corresponding to sections in docs, and move description for vacuum_defer_cleanup_age to the correct category. Sections in postgresql.conf are also sorted in the same order with docs. Per gripe by Fujii Masao, suggestion by Heikki Linnakangas, and patch by me.
2010-06-15 09:52:11 +02:00
{"vacuum_cost_delay", PGC_USERSET, RESOURCES_VACUUM_DELAY,
Rename vacuum_cost_naptime to vacuum_cost_delay, with agreement from Jan.
2004-08-06 06:15:09 +02:00
gettext_noop("Vacuum cost delay in milliseconds."),
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
NULL,
GUC_UNIT_MS
Cost based vacuum delay feature. Jan
2004-02-06 20:36:18 +01:00
},
Rename vacuum_cost_naptime to vacuum_cost_delay, with agreement from Jan.
2004-08-06 06:15:09 +02:00
&VacuumCostDelay,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
0, 0, 100,
NULL, NULL, NULL
Cost based vacuum delay feature. Jan
2004-02-06 20:36:18 +01:00
},
Autovacuum loose end mop-up. Provide autovacuum-specific vacuum cost delay and limit, both as global GUCs and as table-specific entries in pg_autovacuum. stats_reset_on_server_start is now OFF by default, but a reset is forced if we did WAL replay. XID-wrap vacuums do not ANALYZE, but do FREEZE if it's a template database. Alvaro Herrera
2005-08-11 23:11:50 +02:00
{
{"autovacuum_vacuum_cost_delay", PGC_SIGHUP, AUTOVACUUM,
gettext_noop("Vacuum cost delay in milliseconds, for autovacuum."),
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
NULL,
GUC_UNIT_MS
Autovacuum loose end mop-up. Provide autovacuum-specific vacuum cost delay and limit, both as global GUCs and as table-specific entries in pg_autovacuum. stats_reset_on_server_start is now OFF by default, but a reset is forced if we did WAL replay. XID-wrap vacuums do not ANALYZE, but do FREEZE if it's a template database. Alvaro Herrera
2005-08-11 23:11:50 +02:00
},
&autovacuum_vac_cost_delay,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
20, -1, 100,
NULL, NULL, NULL
Autovacuum loose end mop-up. Provide autovacuum-specific vacuum cost delay and limit, both as global GUCs and as table-specific entries in pg_autovacuum. stats_reset_on_server_start is now OFF by default, but a reset is forced if we did WAL replay. XID-wrap vacuums do not ANALYZE, but do FREEZE if it's a template database. Alvaro Herrera
2005-08-11 23:11:50 +02:00
},
{
{"autovacuum_vacuum_cost_limit", PGC_SIGHUP, AUTOVACUUM,
gettext_noop("Vacuum cost amount available before napping, for autovacuum."),
NULL
},
&autovacuum_vac_cost_limit,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
-1, -1, 10000,
NULL, NULL, NULL
Autovacuum loose end mop-up. Provide autovacuum-specific vacuum cost delay and limit, both as global GUCs and as table-specific entries in pg_autovacuum. stats_reset_on_server_start is now OFF by default, but a reset is forced if we did WAL replay. XID-wrap vacuums do not ANALYZE, but do FREEZE if it's a template database. Alvaro Herrera
2005-08-11 23:11:50 +02:00
},
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Do a direct probe during postmaster startup to determine the maximum number of openable files and the number already opened. This eliminates depending on sysconf(_SC_OPEN_MAX), and allows much saner behavior on platforms where open-file slots are used up by semaphores.
2004-02-23 21:45:59 +01:00
{"max_files_per_process", PGC_POSTMASTER, RESOURCES_KERNEL,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Sets the maximum number of simultaneously open files for each server process."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL
},
&max_files_per_process,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
1000, 25, INT_MAX,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
Create a GUC parameter max_files_per_process that is a configurable upper limit on what we will believe from sysconf(_SC_OPEN_MAX). The default value is 1000, so that under ordinary conditions it won't affect the behavior. But on platforms where the kernel promises far more than it can deliver, this can be used to prevent running out of file descriptors. See numerous past discussions, eg, pgsql-hackers around 23-Dec-2000.
2001-09-30 20:57:45 +02:00
Allow read only connections during recovery, known as Hot Standby. Enabled by recovery_connections = on (default) and forcing archive recovery using a recovery.conf. Recovery processing now emulates the original transactions as they are replayed, providing full locking and MVCC behaviour for read only queries. Recovery must enter consistent state before connections are allowed, so there is a delay, typically short, before connections succeed. Replay of recovering transactions can conflict and in some cases deadlock with queries during recovery; these result in query cancellation after max_standby_delay seconds have expired. Infrastructure changes have minor effects on normal running, though introduce four new types of WAL record. New test mode "make standbycheck" allows regression tests of static command behaviour on a standby server while in recovery. Typical and extreme dynamic behaviours have been checked via code inspection and manual testing. Few port specific behaviours have been utilised, though primary testing has been on Linux only so far. This commit is the basic patch. Additional changes will follow in this release to enhance some aspects of behaviour, notably improved handling of conflicts, deadlock detection and query cancellation. Changes to VACUUM FULL are also required. Simon Riggs, with significant and lengthy review by Heikki Linnakangas, including streamlined redesign of snapshot creation and two-phase commit. Important contributions from Florian Pflug, Mark Kirkwood, Merlin Moncure, Greg Stark, Gianni Ciolli, Gabriele Bartolini, Hannu Krosing, Robert Haas, Tatsuo Ishii, Hiroyuki Yamada plus support and feedback from many other community members.
2009-12-19 02:32:45 +01:00
/*
* See also CheckRequiredParameterValues() if this parameter changes
*/
Two-phase commit. Original patch by Heikki Linnakangas, with additional hacking by Alvaro Herrera and Tom Lane.
2005-06-18 00:32:51 +02:00
{
Add new GUC categories corresponding to sections in docs, and move description for vacuum_defer_cleanup_age to the correct category. Sections in postgresql.conf are also sorted in the same order with docs. Per gripe by Fujii Masao, suggestion by Heikki Linnakangas, and patch by me.
2010-06-15 09:52:11 +02:00
{"max_prepared_transactions", PGC_POSTMASTER, RESOURCES_MEM,
Two-phase commit. Original patch by Heikki Linnakangas, with additional hacking by Alvaro Herrera and Tom Lane.
2005-06-18 00:32:51 +02:00
gettext_noop("Sets the maximum number of simultaneously prepared transactions."),
NULL
},
&max_prepared_xacts,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
0, 0, MAX_BACKENDS,
NULL, NULL, NULL
Two-phase commit. Original patch by Heikki Linnakangas, with additional hacking by Alvaro Herrera and Tom Lane.
2005-06-18 00:32:51 +02:00
},
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
#ifdef LOCK_DEBUG
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"trace_lock_oidmin", PGC_SUSET, DEVELOPER_OPTIONS,
Fix capitalization and punctuation of two more GUC description strings.
2007-02-14 04:08:44 +01:00
gettext_noop("No description available."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL,
GUC_NOT_IN_SAMPLE
},
&Trace_lock_oidmin,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
FirstNormalObjectId, 0, INT_MAX,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"trace_lock_table", PGC_SUSET, DEVELOPER_OPTIONS,
Fix capitalization and punctuation of two more GUC description strings.
2007-02-14 04:08:44 +01:00
gettext_noop("No description available."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL,
GUC_NOT_IN_SAMPLE
},
&Trace_lock_table,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
0, 0, INT_MAX,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
#endif
Add SET statement_timeout capability. Timeout is in ms. A value of zero turns off the timer.
2002-07-13 03:02:14 +02:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"statement_timeout", PGC_USERSET, CLIENT_CONN_STATEMENT,
Revert addition of units to GUC descriptions; doesn't affect postgresql.conf.
2009-04-08 01:27:34 +02:00
gettext_noop("Sets the maximum allowed duration of any statement."),
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
gettext_noop("A value of 0 turns off the timeout."),
GUC_UNIT_MS
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
&StatementTimeout,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
0, 0, INT_MAX,
NULL, NULL, NULL
Add SET statement_timeout capability. Timeout is in ms. A value of zero turns off the timer.
2002-07-13 03:02:14 +02:00
},
Fix recently-understood problems with handling of XID freezing, particularly in PITR scenarios. We now WAL-log the replacement of old XIDs with FrozenTransactionId, so that such replacement is guaranteed to propagate to PITR slave databases. Also, rather than relying on hint-bit updates to be preserved, pg_clog is not truncated until all instances of an XID are known to have been replaced by FrozenTransactionId. Add new GUC variables and pg_autovacuum columns to allow management of the freezing policy, so that users can trade off the size of pg_clog against the amount of freezing work done. Revise the already-existing code that forces autovacuum of tables approaching the wraparound point to make it more bulletproof; also, revise the autovacuum logic so that anti-wraparound vacuuming is done per-table rather than per-database. initdb forced because of changes in pg_class, pg_database, and pg_autovacuum catalogs. Heikki Linnakangas, Simon Riggs, and Tom Lane.
2006-11-05 23:42:10 +01:00
{
{"vacuum_freeze_min_age", PGC_USERSET, CLIENT_CONN_STATEMENT,
gettext_noop("Minimum age at which VACUUM should freeze a table row."),
NULL
},
&vacuum_freeze_min_age,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
50000000, 0, 1000000000,
NULL, NULL, NULL
Add vacuum_freeze_table_age GUC option, to control when VACUUM should ignore the visibility map and scan the whole table, to advance relfrozenxid.
2009-01-16 14:27:24 +01:00
},
{
{"vacuum_freeze_table_age", PGC_USERSET, CLIENT_CONN_STATEMENT,
gettext_noop("Age at which VACUUM should scan whole table to freeze tuples."),
NULL
},
&vacuum_freeze_table_age,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
150000000, 0, 2000000000,
NULL, NULL, NULL
Fix recently-understood problems with handling of XID freezing, particularly in PITR scenarios. We now WAL-log the replacement of old XIDs with FrozenTransactionId, so that such replacement is guaranteed to propagate to PITR slave databases. Also, rather than relying on hint-bit updates to be preserved, pg_clog is not truncated until all instances of an XID are known to have been replaced by FrozenTransactionId. Add new GUC variables and pg_autovacuum columns to allow management of the freezing policy, so that users can trade off the size of pg_clog against the amount of freezing work done. Revise the already-existing code that forces autovacuum of tables approaching the wraparound point to make it more bulletproof; also, revise the autovacuum logic so that anti-wraparound vacuuming is done per-table rather than per-database. initdb forced because of changes in pg_class, pg_database, and pg_autovacuum catalogs. Heikki Linnakangas, Simon Riggs, and Tom Lane.
2006-11-05 23:42:10 +01:00
},
Allow read only connections during recovery, known as Hot Standby. Enabled by recovery_connections = on (default) and forcing archive recovery using a recovery.conf. Recovery processing now emulates the original transactions as they are replayed, providing full locking and MVCC behaviour for read only queries. Recovery must enter consistent state before connections are allowed, so there is a delay, typically short, before connections succeed. Replay of recovering transactions can conflict and in some cases deadlock with queries during recovery; these result in query cancellation after max_standby_delay seconds have expired. Infrastructure changes have minor effects on normal running, though introduce four new types of WAL record. New test mode "make standbycheck" allows regression tests of static command behaviour on a standby server while in recovery. Typical and extreme dynamic behaviours have been checked via code inspection and manual testing. Few port specific behaviours have been utilised, though primary testing has been on Linux only so far. This commit is the basic patch. Additional changes will follow in this release to enhance some aspects of behaviour, notably improved handling of conflicts, deadlock detection and query cancellation. Changes to VACUUM FULL are also required. Simon Riggs, with significant and lengthy review by Heikki Linnakangas, including streamlined redesign of snapshot creation and two-phase commit. Important contributions from Florian Pflug, Mark Kirkwood, Merlin Moncure, Greg Stark, Gianni Ciolli, Gabriele Bartolini, Hannu Krosing, Robert Haas, Tatsuo Ishii, Hiroyuki Yamada plus support and feedback from many other community members.
2009-12-19 02:32:45 +01:00
{
Reclassify replication-related GUC variables as "master" and "standby". Per discussion, this structure seems more understandable than what was there before. Make config.sgml and postgresql.conf.sample agree. In passing do a bit of editorial work on the variable descriptions.
2011-07-07 21:10:32 +02:00
{"vacuum_defer_cleanup_age", PGC_SIGHUP, REPLICATION_MASTER,
Make vacuum_defer_cleanup_age be PGC_SIGHUP level, since it's not sensible to have different values in different processes of the primary server. Also put it into the "Streaming Replication" GUC category; it doesn't belong in "Standby Servers" because you use it on the master not the standby. In passing also correct guc.c's idea of wal_keep_segments' category.
2010-07-03 23:23:58 +02:00
gettext_noop("Number of transactions by which VACUUM and HOT cleanup should be deferred, if any."),
Allow read only connections during recovery, known as Hot Standby. Enabled by recovery_connections = on (default) and forcing archive recovery using a recovery.conf. Recovery processing now emulates the original transactions as they are replayed, providing full locking and MVCC behaviour for read only queries. Recovery must enter consistent state before connections are allowed, so there is a delay, typically short, before connections succeed. Replay of recovering transactions can conflict and in some cases deadlock with queries during recovery; these result in query cancellation after max_standby_delay seconds have expired. Infrastructure changes have minor effects on normal running, though introduce four new types of WAL record. New test mode "make standbycheck" allows regression tests of static command behaviour on a standby server while in recovery. Typical and extreme dynamic behaviours have been checked via code inspection and manual testing. Few port specific behaviours have been utilised, though primary testing has been on Linux only so far. This commit is the basic patch. Additional changes will follow in this release to enhance some aspects of behaviour, notably improved handling of conflicts, deadlock detection and query cancellation. Changes to VACUUM FULL are also required. Simon Riggs, with significant and lengthy review by Heikki Linnakangas, including streamlined redesign of snapshot creation and two-phase commit. Important contributions from Florian Pflug, Mark Kirkwood, Merlin Moncure, Greg Stark, Gianni Ciolli, Gabriele Bartolini, Hannu Krosing, Robert Haas, Tatsuo Ishii, Hiroyuki Yamada plus support and feedback from many other community members.
2009-12-19 02:32:45 +01:00
NULL
},
&vacuum_defer_cleanup_age,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
0, 0, 1000000,
NULL, NULL, NULL
Allow read only connections during recovery, known as Hot Standby. Enabled by recovery_connections = on (default) and forcing archive recovery using a recovery.conf. Recovery processing now emulates the original transactions as they are replayed, providing full locking and MVCC behaviour for read only queries. Recovery must enter consistent state before connections are allowed, so there is a delay, typically short, before connections succeed. Replay of recovering transactions can conflict and in some cases deadlock with queries during recovery; these result in query cancellation after max_standby_delay seconds have expired. Infrastructure changes have minor effects on normal running, though introduce four new types of WAL record. New test mode "make standbycheck" allows regression tests of static command behaviour on a standby server while in recovery. Typical and extreme dynamic behaviours have been checked via code inspection and manual testing. Few port specific behaviours have been utilised, though primary testing has been on Linux only so far. This commit is the basic patch. Additional changes will follow in this release to enhance some aspects of behaviour, notably improved handling of conflicts, deadlock detection and query cancellation. Changes to VACUUM FULL are also required. Simon Riggs, with significant and lengthy review by Heikki Linnakangas, including streamlined redesign of snapshot creation and two-phase commit. Important contributions from Florian Pflug, Mark Kirkwood, Merlin Moncure, Greg Stark, Gianni Ciolli, Gabriele Bartolini, Hannu Krosing, Robert Haas, Tatsuo Ishii, Hiroyuki Yamada plus support and feedback from many other community members.
2009-12-19 02:32:45 +01:00
},
/*
* See also CheckRequiredParameterValues() if this parameter changes
*/
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"max_locks_per_transaction", PGC_POSTMASTER, LOCK_MANAGEMENT,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Sets the maximum number of locks per transaction."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
gettext_noop("The shared lock table is sized on the assumption that "
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
"at most max_locks_per_transaction * max_connections distinct "
Limit NBuffers and some related values to INT_MAX / BLCKSZ, to prevent arithmetic overflow during initial sizing calculations. This is not water-tight but it should avoid the grossest sorts of failures.
2004-10-16 21:08:38 +02:00
"objects will need to be locked at any one time.")
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
&max_locks_per_xact,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
64, 10, INT_MAX,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
Add runtime configuration options to control permission bits and group owner of unix socket.
2000-11-01 22:14:03 +01:00
Implement genuine serializable isolation level. Until now, our Serializable mode has in fact been what's called Snapshot Isolation, which allows some anomalies that could not occur in any serialized ordering of the transactions. This patch fixes that using a method called Serializable Snapshot Isolation, based on research papers by Michael J. Cahill (see README-SSI for full references). In Serializable Snapshot Isolation, transactions run like they do in Snapshot Isolation, but a predicate lock manager observes the reads and writes performed and aborts transactions if it detects that an anomaly might occur. This method produces some false positives, ie. it sometimes aborts transactions even though there is no anomaly. To track reads we implement predicate locking, see storage/lmgr/predicate.c. Whenever a tuple is read, a predicate lock is acquired on the tuple. Shared memory is finite, so when a transaction takes many tuple-level locks on a page, the locks are promoted to a single page-level lock, and further to a single relation level lock if necessary. To lock key values with no matching tuple, a sequential scan always takes a relation-level lock, and an index scan acquires a page-level lock that covers the search key, whether or not there are any matching keys at the moment. A predicate lock doesn't conflict with any regular locks or with another predicate locks in the normal sense. They're only used by the predicate lock manager to detect the danger of anomalies. Only serializable transactions participate in predicate locking, so there should be no extra overhead for for other transactions. Predicate locks can't be released at commit, but must be remembered until all the transactions that overlapped with it have completed. That means that we need to remember an unbounded amount of predicate locks, so we apply a lossy but conservative method of tracking locks for committed transactions. If we run short of shared memory, we overflow to a new "pg_serial" SLRU pool. We don't currently allow Serializable transactions in Hot Standby mode. That would be hard, because even read-only transactions can cause anomalies that wouldn't otherwise occur. Serializable isolation mode now means the new fully serializable level. Repeatable Read gives you the old Snapshot Isolation level that we have always had. Kevin Grittner and Dan Ports, reviewed by Jeff Davis, Heikki Linnakangas and Anssi Kääriäinen
2011-02-07 22:46:51 +01:00
{
Rename max_predicate_locks_per_transaction. The new name, max_pred_locks_per_transaction, is shorter. Kevin Grittner, per discussion.
2011-02-15 14:00:04 +01:00
{"max_pred_locks_per_transaction", PGC_POSTMASTER, LOCK_MANAGEMENT,
Implement genuine serializable isolation level. Until now, our Serializable mode has in fact been what's called Snapshot Isolation, which allows some anomalies that could not occur in any serialized ordering of the transactions. This patch fixes that using a method called Serializable Snapshot Isolation, based on research papers by Michael J. Cahill (see README-SSI for full references). In Serializable Snapshot Isolation, transactions run like they do in Snapshot Isolation, but a predicate lock manager observes the reads and writes performed and aborts transactions if it detects that an anomaly might occur. This method produces some false positives, ie. it sometimes aborts transactions even though there is no anomaly. To track reads we implement predicate locking, see storage/lmgr/predicate.c. Whenever a tuple is read, a predicate lock is acquired on the tuple. Shared memory is finite, so when a transaction takes many tuple-level locks on a page, the locks are promoted to a single page-level lock, and further to a single relation level lock if necessary. To lock key values with no matching tuple, a sequential scan always takes a relation-level lock, and an index scan acquires a page-level lock that covers the search key, whether or not there are any matching keys at the moment. A predicate lock doesn't conflict with any regular locks or with another predicate locks in the normal sense. They're only used by the predicate lock manager to detect the danger of anomalies. Only serializable transactions participate in predicate locking, so there should be no extra overhead for for other transactions. Predicate locks can't be released at commit, but must be remembered until all the transactions that overlapped with it have completed. That means that we need to remember an unbounded amount of predicate locks, so we apply a lossy but conservative method of tracking locks for committed transactions. If we run short of shared memory, we overflow to a new "pg_serial" SLRU pool. We don't currently allow Serializable transactions in Hot Standby mode. That would be hard, because even read-only transactions can cause anomalies that wouldn't otherwise occur. Serializable isolation mode now means the new fully serializable level. Repeatable Read gives you the old Snapshot Isolation level that we have always had. Kevin Grittner and Dan Ports, reviewed by Jeff Davis, Heikki Linnakangas and Anssi Kääriäinen
2011-02-07 22:46:51 +01:00
gettext_noop("Sets the maximum number of predicate locks per transaction."),
gettext_noop("The shared predicate lock table is sized on the assumption that "
pgindent run before PG 9.1 beta 1.
2011-04-10 17:42:00 +02:00
"at most max_pred_locks_per_transaction * max_connections distinct "
Implement genuine serializable isolation level. Until now, our Serializable mode has in fact been what's called Snapshot Isolation, which allows some anomalies that could not occur in any serialized ordering of the transactions. This patch fixes that using a method called Serializable Snapshot Isolation, based on research papers by Michael J. Cahill (see README-SSI for full references). In Serializable Snapshot Isolation, transactions run like they do in Snapshot Isolation, but a predicate lock manager observes the reads and writes performed and aborts transactions if it detects that an anomaly might occur. This method produces some false positives, ie. it sometimes aborts transactions even though there is no anomaly. To track reads we implement predicate locking, see storage/lmgr/predicate.c. Whenever a tuple is read, a predicate lock is acquired on the tuple. Shared memory is finite, so when a transaction takes many tuple-level locks on a page, the locks are promoted to a single page-level lock, and further to a single relation level lock if necessary. To lock key values with no matching tuple, a sequential scan always takes a relation-level lock, and an index scan acquires a page-level lock that covers the search key, whether or not there are any matching keys at the moment. A predicate lock doesn't conflict with any regular locks or with another predicate locks in the normal sense. They're only used by the predicate lock manager to detect the danger of anomalies. Only serializable transactions participate in predicate locking, so there should be no extra overhead for for other transactions. Predicate locks can't be released at commit, but must be remembered until all the transactions that overlapped with it have completed. That means that we need to remember an unbounded amount of predicate locks, so we apply a lossy but conservative method of tracking locks for committed transactions. If we run short of shared memory, we overflow to a new "pg_serial" SLRU pool. We don't currently allow Serializable transactions in Hot Standby mode. That would be hard, because even read-only transactions can cause anomalies that wouldn't otherwise occur. Serializable isolation mode now means the new fully serializable level. Repeatable Read gives you the old Snapshot Isolation level that we have always had. Kevin Grittner and Dan Ports, reviewed by Jeff Davis, Heikki Linnakangas and Anssi Kääriäinen
2011-02-07 22:46:51 +01:00
"objects will need to be locked at any one time.")
},
&max_predicate_locks_per_xact,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
64, 10, INT_MAX,
NULL, NULL, NULL
Implement genuine serializable isolation level. Until now, our Serializable mode has in fact been what's called Snapshot Isolation, which allows some anomalies that could not occur in any serialized ordering of the transactions. This patch fixes that using a method called Serializable Snapshot Isolation, based on research papers by Michael J. Cahill (see README-SSI for full references). In Serializable Snapshot Isolation, transactions run like they do in Snapshot Isolation, but a predicate lock manager observes the reads and writes performed and aborts transactions if it detects that an anomaly might occur. This method produces some false positives, ie. it sometimes aborts transactions even though there is no anomaly. To track reads we implement predicate locking, see storage/lmgr/predicate.c. Whenever a tuple is read, a predicate lock is acquired on the tuple. Shared memory is finite, so when a transaction takes many tuple-level locks on a page, the locks are promoted to a single page-level lock, and further to a single relation level lock if necessary. To lock key values with no matching tuple, a sequential scan always takes a relation-level lock, and an index scan acquires a page-level lock that covers the search key, whether or not there are any matching keys at the moment. A predicate lock doesn't conflict with any regular locks or with another predicate locks in the normal sense. They're only used by the predicate lock manager to detect the danger of anomalies. Only serializable transactions participate in predicate locking, so there should be no extra overhead for for other transactions. Predicate locks can't be released at commit, but must be remembered until all the transactions that overlapped with it have completed. That means that we need to remember an unbounded amount of predicate locks, so we apply a lossy but conservative method of tracking locks for committed transactions. If we run short of shared memory, we overflow to a new "pg_serial" SLRU pool. We don't currently allow Serializable transactions in Hot Standby mode. That would be hard, because even read-only transactions can cause anomalies that wouldn't otherwise occur. Serializable isolation mode now means the new fully serializable level. Repeatable Read gives you the old Snapshot Isolation level that we have always had. Kevin Grittner and Dan Ports, reviewed by Jeff Davis, Heikki Linnakangas and Anssi Kääriäinen
2011-02-07 22:46:51 +01:00
},
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"authentication_timeout", PGC_SIGHUP, CONN_AUTH_SECURITY,
Revert addition of units to GUC descriptions; doesn't affect postgresql.conf.
2009-04-08 01:27:34 +02:00
gettext_noop("Sets the maximum allowed time to complete client authentication."),
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
NULL,
GUC_UNIT_S
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
&AuthenticationTimeout,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
60, 1, 600,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
Add an overall timeout on the client authentication cycle, so that a hung client or lost connection can't indefinitely block a postmaster child (not to mention the possibility of deliberate DoS attacks). Timeout is controlled by new authentication_timeout GUC variable, which I set to 60 seconds by default ... does that seem reasonable?
2001-09-21 19:06:12 +02:00
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
/* Not for general use */
{"pre_auth_delay", PGC_SIGHUP, DEVELOPER_OPTIONS,
Various improvements to the GUC description strings. Punctuate and capitalize the strings like sentences. Remove unnecessarily specific descriptions of the units used by GUC variables, since we now allow any reasonable unit to be specified.
2006-12-06 18:35:49 +01:00
gettext_noop("Waits N seconds on connection startup before authentication."),
gettext_noop("This allows attaching a debugger to the process."),
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
GUC_NOT_IN_SAMPLE | GUC_UNIT_S
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
&PreAuthDelay,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
0, 0, 60,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
Add an overall timeout on the client authentication cycle, so that a hung client or lost connection can't indefinitely block a postmaster child (not to mention the possibility of deliberate DoS attacks). Timeout is controlled by new authentication_timeout GUC variable, which I set to 60 seconds by default ... does that seem reasonable?
2001-09-21 19:06:12 +02:00
Change the logic to decide when to delete old WAL segments, so that it doesn't take into account how far the WAL senders are. This way a hung WAL sender doesn't prevent old WAL segments from being recycled/removed in the primary, ultimately causing the disk to fill up. Instead add standby_keep_segments setting to control how many old WAL segments are kept in the primary. This also makes it more reliable to use streaming replication without WAL archiving, assuming that you set standby_keep_segments high enough.
2010-04-12 11:52:29 +02:00
{
Introduce sending servers as new category for replication params Fujii Masao
2011-07-19 09:59:55 +02:00
{"wal_keep_segments", PGC_SIGHUP, REPLICATION_SENDING,
Fix message punctuation
2010-04-22 21:40:03 +02:00
gettext_noop("Sets the number of WAL files held for standby servers."),
Change the logic to decide when to delete old WAL segments, so that it doesn't take into account how far the WAL senders are. This way a hung WAL sender doesn't prevent old WAL segments from being recycled/removed in the primary, ultimately causing the disk to fill up. Instead add standby_keep_segments setting to control how many old WAL segments are kept in the primary. This also makes it more reliable to use streaming replication without WAL archiving, assuming that you set standby_keep_segments high enough.
2010-04-12 11:52:29 +02:00
NULL
},
Rename standby_keep_segments to wal_keep_segments. Also, make the name of the GUC and the name of the backing variable match. Alnong the way, clean up a couple of slight typographical errors in the related docs.
2010-04-20 13:15:06 +02:00
&wal_keep_segments,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
0, 0, INT_MAX,
NULL, NULL, NULL
Change the logic to decide when to delete old WAL segments, so that it doesn't take into account how far the WAL senders are. This way a hung WAL sender doesn't prevent old WAL segments from being recycled/removed in the primary, ultimately causing the disk to fill up. Instead add standby_keep_segments setting to control how many old WAL segments are kept in the primary. This also makes it more reliable to use streaming replication without WAL archiving, assuming that you set standby_keep_segments high enough.
2010-04-12 11:52:29 +02:00
},
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"checkpoint_segments", PGC_SIGHUP, WAL_CHECKPOINTS,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Sets the maximum distance in log segments between automatic WAL checkpoints."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL
},
&CheckPointSegments,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
3, 1, INT_MAX,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
XLOG (and related) changes: * Store two past checkpoint locations, not just one, in pg_control. On startup, we fall back to the older checkpoint if the newer one is unreadable. Also, a physical copy of the newest checkpoint record is kept in pg_control for possible use in disaster recovery (ie, complete loss of pg_xlog). Also add a version number for pg_control itself. Remove archdir from pg_control; it ought to be a GUC parameter, not a special case (not that it's implemented yet anyway). * Suppress successive checkpoint records when nothing has been entered in the WAL log since the last one. This is not so much to avoid I/O as to make it actually useful to keep track of the last two checkpoints. If the things are right next to each other then there's not a lot of redundancy gained... * Change CRC scheme to a true 64-bit CRC, not a pair of 32-bit CRCs on alternate bytes. Polynomial borrowed from ECMA DLT1 standard. * Fix XLOG record length handling so that it will work at BLCKSZ = 32k. * Change XID allocation to work more like OID allocation. (This is of dubious necessity, but I think it's a good idea anyway.) * Fix a number of minor bugs, such as off-by-one logic for XLOG file wraparound at the 4 gig mark. * Add documentation and clean up some coding infelicities; move file format declarations out to include files where planned contrib utilities can get at them. * Checkpoint will now occur every CHECKPOINT_SEGMENTS log segments or every CHECKPOINT_TIMEOUT seconds, whichever comes first. It is also possible to force a checkpoint by sending SIGUSR1 to the postmaster (undocumented feature...) * Defend against kill -9 postmaster by storing shmem block's key and ID in postmaster.pid lockfile, and checking at startup to ensure that no processes are still connected to old shmem block (if it still exists). * Switch backends to accept SIGQUIT rather than SIGUSR1 for emergency stop, for symmetry with postmaster and xlog utilities. Clean up signal handling in bootstrap.c so that xlog utilities launched by postmaster will react to signals better. * Standalone bootstrap now grabs lockfile in target directory, as added insurance against running it in parallel with live postmaster.
2001-03-13 02:17:06 +01:00
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"checkpoint_timeout", PGC_SIGHUP, WAL_CHECKPOINTS,
Revert addition of units to GUC descriptions; doesn't affect postgresql.conf.
2009-04-08 01:27:34 +02:00
gettext_noop("Sets the maximum time between automatic WAL checkpoints."),
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
NULL,
GUC_UNIT_S
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
&CheckPointTimeout,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
300, 30, 3600,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
Auto checkpoint creation.
2000-11-09 12:26:00 +01:00
Add checkpoint_warning to warn of excessive checkpoints caused by too few WAL files.
2002-11-15 03:44:57 +01:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"checkpoint_warning", PGC_SIGHUP, WAL_CHECKPOINTS,
Various improvements to the GUC description strings. Punctuate and capitalize the strings like sentences. Remove unnecessarily specific descriptions of the units used by GUC variables, since we now allow any reasonable unit to be specified.
2006-12-06 18:35:49 +01:00
gettext_noop("Enables warnings if checkpoint segments are filled more "
"frequently than this."),
More message editing, some suggested by Alvaro Herrera
2003-09-29 02:05:25 +02:00
gettext_noop("Write a message to the server log if checkpoints "
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
"caused by the filling of checkpoint segment files happens more "
Add units to the default postgresql.conf. For the most part, this should match what SHOW displays as default value, to make the user experience uniform.
2006-09-22 19:41:21 +02:00
"frequently than this number of seconds. Zero turns off the warning."),
GUC_UNIT_S
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
&CheckPointWarning,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
30, 0, INT_MAX,
NULL, NULL, NULL
Add checkpoint_warning to warn of excessive checkpoints caused by too few WAL files.
2002-11-15 03:44:57 +01:00
},
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"wal_buffers", PGC_POSTMASTER, WAL_SETTINGS,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Sets the number of disk-page buffers in shared memory for WAL."),
The attached patch changes units of the some default values in postgresql.conf. - shared_buffers = 32000kB => 32MB - temp_buffers = 8000kB => 8MB - wal_buffers = 8 => 64kB The code of initdb was a bit modified to write MB-unit values. Values greater than 8000kB are rounded out to MB. GUC_UNIT_XBLOCKS is added for wal_buffers. It is like GUC_UNIT_BLOCKS, but uses XLOG_BLCKSZ instead of BLCKSZ. Also, I cleaned up the test of GUC_UNIT_* flags in preparation to add more unit flags in less bits. ITAGAKI Takahiro
2006-10-03 23:11:55 +02:00
NULL,
GUC_UNIT_XBLOCKS
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
&XLOGbuffers,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
-1, -1, INT_MAX,
check_wal_buffers, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
Auto checkpoint creation.
2000-11-09 12:26:00 +01:00
Create a new dedicated Postgres process, "wal writer", which exists to write and fsync WAL at convenient intervals. For the moment it just tries to offload this work from backends, but soon it will be responsible for guaranteeing a maximum delay before asynchronously-committed transactions will be flushed to disk. This is a portion of Simon Riggs' async-commit patch, committed to CVS separately because a background WAL writer seems like it might be a good idea independently of the async-commit feature. I rebased walwriter.c on bgwriter.c because it seemed like a more appropriate way of handling signals; while the startup/shutdown logic in postmaster.c is more like autovac because we want walwriter to quit before we start the shutdown checkpoint.
2007-07-24 06:54:09 +02:00
{
{"wal_writer_delay", PGC_SIGHUP, WAL_SETTINGS,
Revert addition of units to GUC descriptions; doesn't affect postgresql.conf.
2009-04-08 01:27:34 +02:00
gettext_noop("WAL writer sleep time between WAL flushes."),
Create a new dedicated Postgres process, "wal writer", which exists to write and fsync WAL at convenient intervals. For the moment it just tries to offload this work from backends, but soon it will be responsible for guaranteeing a maximum delay before asynchronously-committed transactions will be flushed to disk. This is a portion of Simon Riggs' async-commit patch, committed to CVS separately because a background WAL writer seems like it might be a good idea independently of the async-commit feature. I rebased walwriter.c on bgwriter.c because it seemed like a more appropriate way of handling signals; while the startup/shutdown logic in postmaster.c is more like autovac because we want walwriter to quit before we start the shutdown checkpoint.
2007-07-24 06:54:09 +02:00
NULL,
GUC_UNIT_MS
},
&WalWriterDelay,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
200, 1, 10000,
NULL, NULL, NULL
Create a new dedicated Postgres process, "wal writer", which exists to write and fsync WAL at convenient intervals. For the moment it just tries to offload this work from backends, but soon it will be responsible for guaranteeing a maximum delay before asynchronously-committed transactions will be flushed to disk. This is a portion of Simon Riggs' async-commit patch, committed to CVS separately because a background WAL writer seems like it might be a good idea independently of the async-commit feature. I rebased walwriter.c on bgwriter.c because it seemed like a more appropriate way of handling signals; while the startup/shutdown logic in postmaster.c is more like autovac because we want walwriter to quit before we start the shutdown checkpoint.
2007-07-24 06:54:09 +02:00
},
Introduce Streaming Replication. This includes two new kinds of postmaster processes, walsenders and walreceiver. Walreceiver is responsible for connecting to the primary server and streaming WAL to disk, while walsender runs in the primary server and streams WAL from disk to the client. Documentation still needs work, but the basics are there. We will probably pull the replication section to a new chapter later on, as well as the sections describing file-based replication. But let's do that as a separate patch, so that it's easier to see what has been added/changed. This patch also adds a new section to the chapter about FE/BE protocol, documenting the protocol used by walsender/walreceivxer. Bump catalog version because of two new functions, pg_last_xlog_receive_location() and pg_last_xlog_replay_location(), for monitoring the progress of replication. Fujii Masao, with additional hacking by me
2010-01-15 10:19:10 +01:00
{
/* see max_connections */
Introduce sending servers as new category for replication params Fujii Masao
2011-07-19 09:59:55 +02:00
{"max_wal_senders", PGC_POSTMASTER, REPLICATION_SENDING,
Introduce Streaming Replication. This includes two new kinds of postmaster processes, walsenders and walreceiver. Walreceiver is responsible for connecting to the primary server and streaming WAL to disk, while walsender runs in the primary server and streams WAL from disk to the client. Documentation still needs work, but the basics are there. We will probably pull the replication section to a new chapter later on, as well as the sections describing file-based replication. But let's do that as a separate patch, so that it's easier to see what has been added/changed. This patch also adds a new section to the chapter about FE/BE protocol, documenting the protocol used by walsender/walreceivxer. Bump catalog version because of two new functions, pg_last_xlog_receive_location() and pg_last_xlog_replay_location(), for monitoring the progress of replication. Fujii Masao, with additional hacking by me
2010-01-15 10:19:10 +01:00
gettext_noop("Sets the maximum number of simultaneously running WAL sender processes."),
NULL
},
Refer to max_wal_senders in a more consistent fashion. The error message now makes explicit reference to the GUC that must be changed to fix the problem, using wording suggested by Tom Lane. Along the way, rename the GUC from MaxWalSenders to max_wal_senders for consistency and grep-ability.
2010-04-01 02:43:29 +02:00
&max_wal_senders,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
0, 0, MAX_BACKENDS,
NULL, NULL, NULL
Introduce Streaming Replication. This includes two new kinds of postmaster processes, walsenders and walreceiver. Walreceiver is responsible for connecting to the primary server and streaming WAL to disk, while walsender runs in the primary server and streams WAL from disk to the client. Documentation still needs work, but the basics are there. We will probably pull the replication section to a new chapter later on, as well as the sections describing file-based replication. But let's do that as a separate patch, so that it's easier to see what has been added/changed. This patch also adds a new section to the chapter about FE/BE protocol, documenting the protocol used by walsender/walreceivxer. Bump catalog version because of two new functions, pg_last_xlog_receive_location() and pg_last_xlog_replay_location(), for monitoring the progress of replication. Fujii Masao, with additional hacking by me
2010-01-15 10:19:10 +01:00
},
Automatically terminate replication connections that are idle for more than replication_timeout (a new GUC) milliseconds. The TCP timeout is often too long, you want the master to notice a dead connection much sooner. People complained about that in 9.0 too, but with synchronous replication it's even more important to notice dead connections promptly. Fujii Masao and Heikki Linnakangas
2011-03-30 09:10:32 +02:00
{
Introduce sending servers as new category for replication params Fujii Masao
2011-07-19 09:59:55 +02:00
{"replication_timeout", PGC_SIGHUP, REPLICATION_SENDING,
Automatically terminate replication connections that are idle for more than replication_timeout (a new GUC) milliseconds. The TCP timeout is often too long, you want the master to notice a dead connection much sooner. People complained about that in 9.0 too, but with synchronous replication it's even more important to notice dead connections promptly. Fujii Masao and Heikki Linnakangas
2011-03-30 09:10:32 +02:00
gettext_noop("Sets the maximum time to wait for WAL replication."),
NULL,
GUC_UNIT_MS
},
&replication_timeout,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
60 * 1000, 0, INT_MAX,
NULL, NULL, NULL
Automatically terminate replication connections that are idle for more than replication_timeout (a new GUC) milliseconds. The TCP timeout is often too long, you want the master to notice a dead connection much sooner. People complained about that in 9.0 too, but with synchronous replication it's even more important to notice dead connections promptly. Fujii Masao and Heikki Linnakangas
2011-03-30 09:10:32 +02:00
},
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Support an optional asynchronous commit mode, in which we don't flush WAL before reporting a transaction committed. Data consistency is still guaranteed (unlike setting fsync = off), but a crash may lose the effects of the last few transactions. Patch by Simon, some editorialization by Tom.
2007-08-02 00:45:09 +02:00
{"commit_delay", PGC_USERSET, WAL_SETTINGS,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Sets the delay in microseconds between transaction commit and "
"flushing WAL to disk."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL
},
&CommitDelay,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
0, 0, 100000,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
XLOG stuff for sequences. CommitDelay in guc.c
2000-11-30 02:47:33 +01:00
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Support an optional asynchronous commit mode, in which we don't flush WAL before reporting a transaction committed. Data consistency is still guaranteed (unlike setting fsync = off), but a crash may lose the effects of the last few transactions. Patch by Simon, some editorialization by Tom.
2007-08-02 00:45:09 +02:00
{"commit_siblings", PGC_USERSET, WAL_SETTINGS,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Sets the minimum concurrent open transactions before performing "
"commit_delay."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL
},
&CommitSiblings,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
5, 0, 1000,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
Implement COMMIT_SIBLINGS parameter to allow pre-commit delay to occur only if at least N other backends currently have open transactions. This is not a great deal of intelligence about whether a delay might be profitable ... but it beats no intelligence at all. Note that the default COMMIT_DELAY is still zero --- this new code does nothing unless that setting is changed. Also, mark ENABLEFSYNC as a system-wide setting. It's no longer safe to allow that to be set per-backend, since we may be relying on some other backend's fsync to have synced the WAL log.
2001-02-26 01:50:08 +01:00
Add extra_float_digits GUC parameter to allow adjustment of displayed precision for float4, float8, and geometric types. Set it in pg_dump so that float data can be dumped/reloaded exactly (at least on platforms where the float I/O support is properly implemented). Initial patch by Pedro Ferreira, some additional work by Tom Lane.
2002-11-08 18:37:52 +01:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"extra_float_digits", PGC_USERSET, CLIENT_CONN_LOCALE,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Sets the number of digits displayed for floating-point values."),
Message editing: remove gratuitous variations in message wording, standardize terms, add some clarifications, fix some untranslatable attempts at dynamic message building.
2003-09-25 08:58:07 +02:00
gettext_noop("This affects real, double precision, and geometric data types. "
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
"The parameter value is added to the standard number of digits "
Message editing: remove gratuitous variations in message wording, standardize terms, add some clarifications, fix some untranslatable attempts at dynamic message building.
2003-09-25 08:58:07 +02:00
"(FLT_DIG or DBL_DIG as appropriate).")
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
&extra_float_digits,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
0, -15, 3,
NULL, NULL, NULL
Add extra_float_digits GUC parameter to allow adjustment of displayed precision for float4, float8, and geometric types. Set it in pg_dump so that float data can be dumped/reloaded exactly (at least on platforms where the float I/O support is properly implemented). Initial patch by Pedro Ferreira, some additional work by Tom Lane.
2002-11-08 18:37:52 +01:00
},
Add log_min_duration_statement. Christopher Kings-Lynne
2003-06-11 20:01:14 +02:00
{
Remove GUC USERLIMIT variable category, making the affected variables plain SUSET instead. Also delay processing of options received in client connection request until after we know if the user is a superuser, so that SUSET values can be set that way by legitimate superusers. Per recent discussion.
2004-11-14 20:35:35 +01:00
{"log_min_duration_statement", PGC_SUSET, LOGGING_WHEN,
Revert addition of units to GUC descriptions; doesn't affect postgresql.conf.
2009-04-08 01:27:34 +02:00
gettext_noop("Sets the minimum execution time above which "
"statements will be logged."),
Simplify and rename some GUC variables, per various recent discussions: * stats_start_collector goes away; we always start the collector process, unless prevented by a problem with setting up the stats UDP socket. * stats_reset_on_server_start goes away; it seems useless in view of the availability of pg_stat_reset(). * stats_block_level and stats_row_level are merged into a single variable "track_counts", which controls all reports sent to the collector process. * stats_command_string is renamed to track_activities. * log_autovacuum is renamed to log_autovacuum_min_duration to better reflect its meaning. The log_autovacuum change is not a compatibility issue since it didn't exist before 8.3 anyway. The other changes need to be release-noted.
2007-09-24 05:12:23 +02:00
gettext_noop("Zero prints all queries. -1 turns this feature off."),
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
GUC_UNIT_MS
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
&log_min_duration_statement,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
-1, -1, INT_MAX,
NULL, NULL, NULL
Add log_min_duration_statement. Christopher Kings-Lynne
2003-06-11 20:01:14 +02:00
},
Enable configurable log of autovacuum actions. Initial patch from Simon Riggs, additional code and docs by me. Per discussion.
2007-04-18 18:44:18 +02:00
{
Simplify and rename some GUC variables, per various recent discussions: * stats_start_collector goes away; we always start the collector process, unless prevented by a problem with setting up the stats UDP socket. * stats_reset_on_server_start goes away; it seems useless in view of the availability of pg_stat_reset(). * stats_block_level and stats_row_level are merged into a single variable "track_counts", which controls all reports sent to the collector process. * stats_command_string is renamed to track_activities. * log_autovacuum is renamed to log_autovacuum_min_duration to better reflect its meaning. The log_autovacuum change is not a compatibility issue since it didn't exist before 8.3 anyway. The other changes need to be release-noted.
2007-09-24 05:12:23 +02:00
{"log_autovacuum_min_duration", PGC_SIGHUP, LOGGING_WHAT,
Revert addition of units to GUC descriptions; doesn't affect postgresql.conf.
2009-04-08 01:27:34 +02:00
gettext_noop("Sets the minimum execution time above which "
Simplify and rename some GUC variables, per various recent discussions: * stats_start_collector goes away; we always start the collector process, unless prevented by a problem with setting up the stats UDP socket. * stats_reset_on_server_start goes away; it seems useless in view of the availability of pg_stat_reset(). * stats_block_level and stats_row_level are merged into a single variable "track_counts", which controls all reports sent to the collector process. * stats_command_string is renamed to track_activities. * log_autovacuum is renamed to log_autovacuum_min_duration to better reflect its meaning. The log_autovacuum change is not a compatibility issue since it didn't exist before 8.3 anyway. The other changes need to be release-noted.
2007-09-24 05:12:23 +02:00
"autovacuum actions will be logged."),
gettext_noop("Zero prints all actions. -1 turns autovacuum logging off."),
Enable configurable log of autovacuum actions. Initial patch from Simon Riggs, additional code and docs by me. Per discussion.
2007-04-18 18:44:18 +02:00
GUC_UNIT_MS
},
Simplify and rename some GUC variables, per various recent discussions: * stats_start_collector goes away; we always start the collector process, unless prevented by a problem with setting up the stats UDP socket. * stats_reset_on_server_start goes away; it seems useless in view of the availability of pg_stat_reset(). * stats_block_level and stats_row_level are merged into a single variable "track_counts", which controls all reports sent to the collector process. * stats_command_string is renamed to track_activities. * log_autovacuum is renamed to log_autovacuum_min_duration to better reflect its meaning. The log_autovacuum change is not a compatibility issue since it didn't exist before 8.3 anyway. The other changes need to be release-noted.
2007-09-24 05:12:23 +02:00
&Log_autovacuum_min_duration,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
-1, -1, INT_MAX,
NULL, NULL, NULL
Enable configurable log of autovacuum actions. Initial patch from Simon Riggs, additional code and docs by me. Per discussion.
2007-04-18 18:44:18 +02:00
},
Background writer process This first part of the background writer does no syncing at all. It's only purpose is to keep the LRU heads clean so that regular backends seldom to never have to call write(). Jan
2003-11-19 16:55:08 +01:00
{
Add new GUC categories corresponding to sections in docs, and move description for vacuum_defer_cleanup_age to the correct category. Sections in postgresql.conf are also sorted in the same order with docs. Per gripe by Fujii Masao, suggestion by Heikki Linnakangas, and patch by me.
2010-06-15 09:52:11 +02:00
{"bgwriter_delay", PGC_SIGHUP, RESOURCES_BGWRITER,
Revert addition of units to GUC descriptions; doesn't affect postgresql.conf.
2009-04-08 01:27:34 +02:00
gettext_noop("Background writer sleep time between rounds."),
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
NULL,
GUC_UNIT_MS
Background writer process This first part of the background writer does no syncing at all. It's only purpose is to keep the LRU heads clean so that regular backends seldom to never have to call write(). Jan
2003-11-19 16:55:08 +01:00
},
&BgWriterDelay,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
200, 10, 10000,
NULL, NULL, NULL
Background writer process This first part of the background writer does no syncing at all. It's only purpose is to keep the LRU heads clean so that regular backends seldom to never have to call write(). Jan
2003-11-19 16:55:08 +01:00
},
{
Add new GUC categories corresponding to sections in docs, and move description for vacuum_defer_cleanup_age to the correct category. Sections in postgresql.conf are also sorted in the same order with docs. Per gripe by Fujii Masao, suggestion by Heikki Linnakangas, and patch by me.
2010-06-15 09:52:11 +02:00
{"bgwriter_lru_maxpages", PGC_SIGHUP, RESOURCES_BGWRITER,
Various improvements to the GUC description strings. Punctuate and capitalize the strings like sentences. Remove unnecessarily specific descriptions of the units used by GUC variables, since we now allow any reasonable unit to be specified.
2006-12-06 18:35:49 +01:00
gettext_noop("Background writer maximum number of LRU pages to flush per round."),
Background writer process This first part of the background writer does no syncing at all. It's only purpose is to keep the LRU heads clean so that regular backends seldom to never have to call write(). Jan
2003-11-19 16:55:08 +01:00
NULL
},
Replace the BufMgrLock with separate locks on the lookup hashtable and the freelist, plus per-buffer spinlocks that protect access to individual shared buffer headers. This requires abandoning a global freelist (since the freelist is a global contention point), which shoots down ARC and 2Q as well as plain LRU management. Adopt a clock sweep algorithm instead. Preliminary results show substantial improvement in multi-backend situations.
2005-03-04 21:21:07 +01:00
&bgwriter_lru_maxpages,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
100, 0, 1000,
NULL, NULL, NULL
Background writer process This first part of the background writer does no syncing at all. It's only purpose is to keep the LRU heads clean so that regular backends seldom to never have to call write(). Jan
2003-11-19 16:55:08 +01:00
},
Implement prefetching via posix_fadvise() for bitmap index scans. A new GUC variable effective_io_concurrency controls how many concurrent block prefetch requests will be issued. (The best way to handle this for plain index scans is still under debate, so that part is not applied yet --- tgl) Greg Stark
2009-01-12 06:10:45 +01:00
{
Give a better error message when trying to change "effective_io_concurrency" on systems without posix_fadvise().
2009-04-02 21:57:19 +02:00
{"effective_io_concurrency",
#ifdef USE_PREFETCH
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
PGC_USERSET,
Give a better error message when trying to change "effective_io_concurrency" on systems without posix_fadvise().
2009-04-02 21:57:19 +02:00
#else
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
PGC_INTERNAL,
Give a better error message when trying to change "effective_io_concurrency" on systems without posix_fadvise().
2009-04-02 21:57:19 +02:00
#endif
Add new GUC categories corresponding to sections in docs, and move description for vacuum_defer_cleanup_age to the correct category. Sections in postgresql.conf are also sorted in the same order with docs. Per gripe by Fujii Masao, suggestion by Heikki Linnakangas, and patch by me.
2010-06-15 09:52:11 +02:00
RESOURCES_ASYNCHRONOUS,
Implement prefetching via posix_fadvise() for bitmap index scans. A new GUC variable effective_io_concurrency controls how many concurrent block prefetch requests will be issued. (The best way to handle this for plain index scans is still under debate, so that part is not applied yet --- tgl) Greg Stark
2009-01-12 06:10:45 +01:00
gettext_noop("Number of simultaneous requests that can be handled efficiently by the disk subsystem."),
gettext_noop("For RAID arrays, this should be approximately the number of drive spindles in the array.")
},
&effective_io_concurrency,
#ifdef USE_PREFETCH
1, 0, 1000,
#else
0, 0, 0,
#endif
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
check_effective_io_concurrency, assign_effective_io_concurrency, NULL
Implement prefetching via posix_fadvise() for bitmap index scans. A new GUC variable effective_io_concurrency controls how many concurrent block prefetch requests will be issued. (The best way to handle this for plain index scans is still under debate, so that part is not applied yet --- tgl) Greg Stark
2009-01-12 06:10:45 +01:00
},
Create a built-in log rotation program, so that we no longer have to recommend that people go get Apache's rotatelogs program. Additional benefits are that configuration is done through GUC, rather than externally, and that the postmaster can monitor the log rotator and restart it after failure (though we certainly hope that won't happen often). Andreas Pflug, some rework by Tom Lane.
2004-08-06 01:32:13 +02:00
{
{"log_rotation_age", PGC_SIGHUP, LOGGING_WHERE,
Various improvements to the GUC description strings. Punctuate and capitalize the strings like sentences. Remove unnecessarily specific descriptions of the units used by GUC variables, since we now allow any reasonable unit to be specified.
2006-12-06 18:35:49 +01:00
gettext_noop("Automatic log file rotation will occur after N minutes."),
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
NULL,
GUC_UNIT_MIN
Create a built-in log rotation program, so that we no longer have to recommend that people go get Apache's rotatelogs program. Additional benefits are that configuration is done through GUC, rather than externally, and that the postmaster can monitor the log rotator and restart it after failure (though we certainly hope that won't happen often). Andreas Pflug, some rework by Tom Lane.
2004-08-06 01:32:13 +02:00
},
&Log_RotationAge,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
HOURS_PER_DAY * MINS_PER_HOUR, 0, INT_MAX / MINS_PER_HOUR,
NULL, NULL, NULL
Create a built-in log rotation program, so that we no longer have to recommend that people go get Apache's rotatelogs program. Additional benefits are that configuration is done through GUC, rather than externally, and that the postmaster can monitor the log rotator and restart it after failure (though we certainly hope that won't happen often). Andreas Pflug, some rework by Tom Lane.
2004-08-06 01:32:13 +02:00
},
{
{"log_rotation_size", PGC_SIGHUP, LOGGING_WHERE,
Various improvements to the GUC description strings. Punctuate and capitalize the strings like sentences. Remove unnecessarily specific descriptions of the units used by GUC variables, since we now allow any reasonable unit to be specified.
2006-12-06 18:35:49 +01:00
gettext_noop("Automatic log file rotation will occur after N kilobytes."),
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
NULL,
GUC_UNIT_KB
Create a built-in log rotation program, so that we no longer have to recommend that people go get Apache's rotatelogs program. Additional benefits are that configuration is done through GUC, rather than externally, and that the postmaster can monitor the log rotator and restart it after failure (though we certainly hope that won't happen often). Andreas Pflug, some rework by Tom Lane.
2004-08-06 01:32:13 +02:00
},
&Log_RotationSize,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
10 * 1024, 0, INT_MAX / 1024,
NULL, NULL, NULL
Create a built-in log rotation program, so that we no longer have to recommend that people go get Apache's rotatelogs program. Additional benefits are that configuration is done through GUC, rather than externally, and that the postmaster can monitor the log rotator and restart it after failure (though we certainly hope that won't happen often). Andreas Pflug, some rework by Tom Lane.
2004-08-06 01:32:13 +02:00
},
Added new group of read-only GUC variables to allow simple access to certain compile-time options (FUNC_MAX_ARGS, INDEX_MAX_KEYS, NAMEDATALEN, BLCKSZ, HAVE_INT64_TIMESTAMP). Also added "category", "short_desc", and "extra_desc" to the pg_settings view. Per recent discussion here: http://archives.postgresql.org/pgsql-patches/2003-11/msg00363.php
2003-12-03 19:52:00 +01:00
{
Whack some sense into the configuration-file-location patch. Refactor code into something reasonably understandable, cause use of the feature to not fail in standalone backends or in EXEC_BACKEND case, fix sloppy guc.c table entries, make the documentation minimally usable.
2004-10-08 03:36:36 +02:00
{"max_function_args", PGC_INTERNAL, PRESET_OPTIONS,
Message style revisions
2004-10-12 23:54:45 +02:00
gettext_noop("Shows the maximum number of function arguments."),
Added new group of read-only GUC variables to allow simple access to certain compile-time options (FUNC_MAX_ARGS, INDEX_MAX_KEYS, NAMEDATALEN, BLCKSZ, HAVE_INT64_TIMESTAMP). Also added "category", "short_desc", and "extra_desc" to the pg_settings view. Per recent discussion here: http://archives.postgresql.org/pgsql-patches/2003-11/msg00363.php
2003-12-03 19:52:00 +01:00
NULL,
GUC_NOT_IN_SAMPLE | GUC_DISALLOW_IN_FILE
},
&max_function_args,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
FUNC_MAX_ARGS, FUNC_MAX_ARGS, FUNC_MAX_ARGS,
NULL, NULL, NULL
Added new group of read-only GUC variables to allow simple access to certain compile-time options (FUNC_MAX_ARGS, INDEX_MAX_KEYS, NAMEDATALEN, BLCKSZ, HAVE_INT64_TIMESTAMP). Also added "category", "short_desc", and "extra_desc" to the pg_settings view. Per recent discussion here: http://archives.postgresql.org/pgsql-patches/2003-11/msg00363.php
2003-12-03 19:52:00 +01:00
},
{
Whack some sense into the configuration-file-location patch. Refactor code into something reasonably understandable, cause use of the feature to not fail in standalone backends or in EXEC_BACKEND case, fix sloppy guc.c table entries, make the documentation minimally usable.
2004-10-08 03:36:36 +02:00
{"max_index_keys", PGC_INTERNAL, PRESET_OPTIONS,
Message style revisions
2004-10-12 23:54:45 +02:00
gettext_noop("Shows the maximum number of index keys."),
Added new group of read-only GUC variables to allow simple access to certain compile-time options (FUNC_MAX_ARGS, INDEX_MAX_KEYS, NAMEDATALEN, BLCKSZ, HAVE_INT64_TIMESTAMP). Also added "category", "short_desc", and "extra_desc" to the pg_settings view. Per recent discussion here: http://archives.postgresql.org/pgsql-patches/2003-11/msg00363.php
2003-12-03 19:52:00 +01:00
NULL,
GUC_NOT_IN_SAMPLE | GUC_DISALLOW_IN_FILE
},
&max_index_keys,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
INDEX_MAX_KEYS, INDEX_MAX_KEYS, INDEX_MAX_KEYS,
NULL, NULL, NULL
Added new group of read-only GUC variables to allow simple access to certain compile-time options (FUNC_MAX_ARGS, INDEX_MAX_KEYS, NAMEDATALEN, BLCKSZ, HAVE_INT64_TIMESTAMP). Also added "category", "short_desc", and "extra_desc" to the pg_settings view. Per recent discussion here: http://archives.postgresql.org/pgsql-patches/2003-11/msg00363.php
2003-12-03 19:52:00 +01:00
},
{
Whack some sense into the configuration-file-location patch. Refactor code into something reasonably understandable, cause use of the feature to not fail in standalone backends or in EXEC_BACKEND case, fix sloppy guc.c table entries, make the documentation minimally usable.
2004-10-08 03:36:36 +02:00
{"max_identifier_length", PGC_INTERNAL, PRESET_OPTIONS,
Various improvements to the GUC description strings. Punctuate and capitalize the strings like sentences. Remove unnecessarily specific descriptions of the units used by GUC variables, since we now allow any reasonable unit to be specified.
2006-12-06 18:35:49 +01:00
gettext_noop("Shows the maximum identifier length."),
Added new group of read-only GUC variables to allow simple access to certain compile-time options (FUNC_MAX_ARGS, INDEX_MAX_KEYS, NAMEDATALEN, BLCKSZ, HAVE_INT64_TIMESTAMP). Also added "category", "short_desc", and "extra_desc" to the pg_settings view. Per recent discussion here: http://archives.postgresql.org/pgsql-patches/2003-11/msg00363.php
2003-12-03 19:52:00 +01:00
NULL,
GUC_NOT_IN_SAMPLE | GUC_DISALLOW_IN_FILE
},
&max_identifier_length,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
NAMEDATALEN - 1, NAMEDATALEN - 1, NAMEDATALEN - 1,
NULL, NULL, NULL
Added new group of read-only GUC variables to allow simple access to certain compile-time options (FUNC_MAX_ARGS, INDEX_MAX_KEYS, NAMEDATALEN, BLCKSZ, HAVE_INT64_TIMESTAMP). Also added "category", "short_desc", and "extra_desc" to the pg_settings view. Per recent discussion here: http://archives.postgresql.org/pgsql-patches/2003-11/msg00363.php
2003-12-03 19:52:00 +01:00
},
{
Whack some sense into the configuration-file-location patch. Refactor code into something reasonably understandable, cause use of the feature to not fail in standalone backends or in EXEC_BACKEND case, fix sloppy guc.c table entries, make the documentation minimally usable.
2004-10-08 03:36:36 +02:00
{"block_size", PGC_INTERNAL, PRESET_OPTIONS,
Various improvements to the GUC description strings. Punctuate and capitalize the strings like sentences. Remove unnecessarily specific descriptions of the units used by GUC variables, since we now allow any reasonable unit to be specified.
2006-12-06 18:35:49 +01:00
gettext_noop("Shows the size of a disk block."),
Added new group of read-only GUC variables to allow simple access to certain compile-time options (FUNC_MAX_ARGS, INDEX_MAX_KEYS, NAMEDATALEN, BLCKSZ, HAVE_INT64_TIMESTAMP). Also added "category", "short_desc", and "extra_desc" to the pg_settings view. Per recent discussion here: http://archives.postgresql.org/pgsql-patches/2003-11/msg00363.php
2003-12-03 19:52:00 +01:00
NULL,
GUC_NOT_IN_SAMPLE | GUC_DISALLOW_IN_FILE
},
&block_size,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
BLCKSZ, BLCKSZ, BLCKSZ,
NULL, NULL, NULL
Added new group of read-only GUC variables to allow simple access to certain compile-time options (FUNC_MAX_ARGS, INDEX_MAX_KEYS, NAMEDATALEN, BLCKSZ, HAVE_INT64_TIMESTAMP). Also added "category", "short_desc", and "extra_desc" to the pg_settings view. Per recent discussion here: http://archives.postgresql.org/pgsql-patches/2003-11/msg00363.php
2003-12-03 19:52:00 +01:00
},
Add unchangeable GUC "variables" segment_size, wal_block_size, and wal_segment_size to make those configuration parameters available to clients, in the same way that block_size was previously exposed. Bernd Helmle, with comments from Abhijit Menon-Sen and some further tweaking by me.
2008-07-11 00:08:17 +02:00
{
{"segment_size", PGC_INTERNAL, PRESET_OPTIONS,
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
gettext_noop("Shows the number of pages per disk file."),
NULL,
GUC_UNIT_BLOCKS | GUC_NOT_IN_SAMPLE | GUC_DISALLOW_IN_FILE
Add unchangeable GUC "variables" segment_size, wal_block_size, and wal_segment_size to make those configuration parameters available to clients, in the same way that block_size was previously exposed. Bernd Helmle, with comments from Abhijit Menon-Sen and some further tweaking by me.
2008-07-11 00:08:17 +02:00
},
&segment_size,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
RELSEG_SIZE, RELSEG_SIZE, RELSEG_SIZE,
NULL, NULL, NULL
Add unchangeable GUC "variables" segment_size, wal_block_size, and wal_segment_size to make those configuration parameters available to clients, in the same way that block_size was previously exposed. Bernd Helmle, with comments from Abhijit Menon-Sen and some further tweaking by me.
2008-07-11 00:08:17 +02:00
},
{
{"wal_block_size", PGC_INTERNAL, PRESET_OPTIONS,
gettext_noop("Shows the block size in the write ahead log."),
NULL,
GUC_NOT_IN_SAMPLE | GUC_DISALLOW_IN_FILE
},
&wal_block_size,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
XLOG_BLCKSZ, XLOG_BLCKSZ, XLOG_BLCKSZ,
NULL, NULL, NULL
Add unchangeable GUC "variables" segment_size, wal_block_size, and wal_segment_size to make those configuration parameters available to clients, in the same way that block_size was previously exposed. Bernd Helmle, with comments from Abhijit Menon-Sen and some further tweaking by me.
2008-07-11 00:08:17 +02:00
},
{
{"wal_segment_size", PGC_INTERNAL, PRESET_OPTIONS,
gettext_noop("Shows the number of pages per write ahead log segment."),
NULL,
GUC_UNIT_XBLOCKS | GUC_NOT_IN_SAMPLE | GUC_DISALLOW_IN_FILE
},
&wal_segment_size,
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
(XLOG_SEG_SIZE / XLOG_BLCKSZ),
(XLOG_SEG_SIZE / XLOG_BLCKSZ),
Add unchangeable GUC "variables" segment_size, wal_block_size, and wal_segment_size to make those configuration parameters available to clients, in the same way that block_size was previously exposed. Bernd Helmle, with comments from Abhijit Menon-Sen and some further tweaking by me.
2008-07-11 00:08:17 +02:00
(XLOG_SEG_SIZE / XLOG_BLCKSZ),
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
NULL, NULL, NULL
Add unchangeable GUC "variables" segment_size, wal_block_size, and wal_segment_size to make those configuration parameters available to clients, in the same way that block_size was previously exposed. Bernd Helmle, with comments from Abhijit Menon-Sen and some further tweaking by me.
2008-07-11 00:08:17 +02:00
},
Integrate autovacuum functionality into the backend. There's still a few loose ends to be dealt with, but it seems to work. Alvaro Herrera, based on the contrib code by Matthew O'Connor.
2005-07-14 07:13:45 +02:00
{
{"autovacuum_naptime", PGC_SIGHUP, AUTOVACUUM,
Revert addition of units to GUC descriptions; doesn't affect postgresql.conf.
2009-04-08 01:27:34 +02:00
gettext_noop("Time to sleep between autovacuum runs."),
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
NULL,
GUC_UNIT_S
Integrate autovacuum functionality into the backend. There's still a few loose ends to be dealt with, but it seems to work. Alvaro Herrera, based on the contrib code by Matthew O'Connor.
2005-07-14 07:13:45 +02:00
},
&autovacuum_naptime,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
60, 1, INT_MAX / 1000,
NULL, NULL, NULL
Integrate autovacuum functionality into the backend. There's still a few loose ends to be dealt with, but it seems to work. Alvaro Herrera, based on the contrib code by Matthew O'Connor.
2005-07-14 07:13:45 +02:00
},
{
{"autovacuum_vacuum_threshold", PGC_SIGHUP, AUTOVACUUM,
gettext_noop("Minimum number of tuple updates or deletes prior to vacuum."),
NULL
},
&autovacuum_vac_thresh,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
50, 0, INT_MAX,
NULL, NULL, NULL
Integrate autovacuum functionality into the backend. There's still a few loose ends to be dealt with, but it seems to work. Alvaro Herrera, based on the contrib code by Matthew O'Connor.
2005-07-14 07:13:45 +02:00
},
{
{"autovacuum_analyze_threshold", PGC_SIGHUP, AUTOVACUUM,
gettext_noop("Minimum number of tuple inserts, updates or deletes prior to analyze."),
NULL
},
&autovacuum_anl_thresh,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
50, 0, INT_MAX,
NULL, NULL, NULL
Integrate autovacuum functionality into the backend. There's still a few loose ends to be dealt with, but it seems to work. Alvaro Herrera, based on the contrib code by Matthew O'Connor.
2005-07-14 07:13:45 +02:00
},
Fix recently-understood problems with handling of XID freezing, particularly in PITR scenarios. We now WAL-log the replacement of old XIDs with FrozenTransactionId, so that such replacement is guaranteed to propagate to PITR slave databases. Also, rather than relying on hint-bit updates to be preserved, pg_clog is not truncated until all instances of an XID are known to have been replaced by FrozenTransactionId. Add new GUC variables and pg_autovacuum columns to allow management of the freezing policy, so that users can trade off the size of pg_clog against the amount of freezing work done. Revise the already-existing code that forces autovacuum of tables approaching the wraparound point to make it more bulletproof; also, revise the autovacuum logic so that anti-wraparound vacuuming is done per-table rather than per-database. initdb forced because of changes in pg_class, pg_database, and pg_autovacuum catalogs. Heikki Linnakangas, Simon Riggs, and Tom Lane.
2006-11-05 23:42:10 +01:00
{
/* see varsup.c for why this is PGC_POSTMASTER not PGC_SIGHUP */
{"autovacuum_freeze_max_age", PGC_POSTMASTER, AUTOVACUUM,
Spelling fix
2006-11-29 15:50:07 +01:00
gettext_noop("Age at which to autovacuum a table to prevent transaction ID wraparound."),
Fix recently-understood problems with handling of XID freezing, particularly in PITR scenarios. We now WAL-log the replacement of old XIDs with FrozenTransactionId, so that such replacement is guaranteed to propagate to PITR slave databases. Also, rather than relying on hint-bit updates to be preserved, pg_clog is not truncated until all instances of an XID are known to have been replaced by FrozenTransactionId. Add new GUC variables and pg_autovacuum columns to allow management of the freezing policy, so that users can trade off the size of pg_clog against the amount of freezing work done. Revise the already-existing code that forces autovacuum of tables approaching the wraparound point to make it more bulletproof; also, revise the autovacuum logic so that anti-wraparound vacuuming is done per-table rather than per-database. initdb forced because of changes in pg_class, pg_database, and pg_autovacuum catalogs. Heikki Linnakangas, Simon Riggs, and Tom Lane.
2006-11-05 23:42:10 +01:00
NULL
},
&autovacuum_freeze_max_age,
Track the current XID wrap limit (or more accurately, the oldest unfrozen XID) in checkpoint records. This eliminates the need to recompute the value from scratch during database startup, which is one of the two remaining reasons for the flatfile code to exist. It should also simplify life for hot-standby operation. To avoid bloating the checkpoint records unreasonably, I switched from tracking the oldest database by name to tracking it by OID. This turns out to save cycles in general (everywhere but the warning-generating paths, which we hardly care about) and also helps us deal with the case that the oldest database got dropped instead of being vacuumed. The prior coding might go for a long time without updating the wrap limit in that case, which is bad because it might result in a lot of useless autovacuum activity.
2009-08-31 04:23:23 +02:00
/* see pg_resetxlog if you change the upper-limit value */
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
200000000, 100000000, 2000000000,
NULL, NULL, NULL
Fix recently-understood problems with handling of XID freezing, particularly in PITR scenarios. We now WAL-log the replacement of old XIDs with FrozenTransactionId, so that such replacement is guaranteed to propagate to PITR slave databases. Also, rather than relying on hint-bit updates to be preserved, pg_clog is not truncated until all instances of an XID are known to have been replaced by FrozenTransactionId. Add new GUC variables and pg_autovacuum columns to allow management of the freezing policy, so that users can trade off the size of pg_clog against the amount of freezing work done. Revise the already-existing code that forces autovacuum of tables approaching the wraparound point to make it more bulletproof; also, revise the autovacuum logic so that anti-wraparound vacuuming is done per-table rather than per-database. initdb forced because of changes in pg_class, pg_database, and pg_autovacuum catalogs. Heikki Linnakangas, Simon Riggs, and Tom Lane.
2006-11-05 23:42:10 +01:00
},
Add a multi-worker capability to autovacuum. This allows multiple worker processes to be running simultaneously. Also, now autovacuum processes do not count towards the max_connections limit; they are counted separately from regular processes, and are limited by the new GUC variable autovacuum_max_workers. The launcher now has intelligence to launch workers on each database every autovacuum_naptime seconds, limited only on the max amount of worker slots available. Also, the global worker I/O utilization is limited by the vacuum cost-based delay feature. Workers are "balanced" so that the total I/O consumption does not exceed the established limit. This part of the patch was contributed by ITAGAKI Takahiro. Per discussion.
2007-04-16 20:30:04 +02:00
{
/* see max_connections */
{"autovacuum_max_workers", PGC_POSTMASTER, AUTOVACUUM,
gettext_noop("Sets the maximum number of simultaneously running autovacuum worker processes."),
NULL
},
&autovacuum_max_workers,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
3, 1, MAX_BACKENDS,
check_autovacuum_max_workers, assign_autovacuum_max_workers, NULL
Add a multi-worker capability to autovacuum. This allows multiple worker processes to be running simultaneously. Also, now autovacuum processes do not count towards the max_connections limit; they are counted separately from regular processes, and are limited by the new GUC variable autovacuum_max_workers. The launcher now has intelligence to launch workers on each database every autovacuum_naptime seconds, limited only on the max amount of worker slots available. Also, the global worker I/O utilization is limited by the vacuum cost-based delay feature. Workers are "balanced" so that the total I/O consumption does not exceed the established limit. This part of the patch was contributed by ITAGAKI Takahiro. Per discussion.
2007-04-16 20:30:04 +02:00
},
Integrate autovacuum functionality into the backend. There's still a few loose ends to be dealt with, but it seems to work. Alvaro Herrera, based on the contrib code by Matthew O'Connor.
2005-07-14 07:13:45 +02:00
Add GUC variables to control keep-alive times for idle, interval, and count. Oliver Jowett
2005-07-30 17:17:26 +02:00
{
{"tcp_keepalives_idle", PGC_USERSET, CLIENT_CONN_OTHER,
Revert addition of units to GUC descriptions; doesn't affect postgresql.conf.
2009-04-08 01:27:34 +02:00
gettext_noop("Time between issuing TCP keepalives."),
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
gettext_noop("A value of 0 uses the system default."),
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
GUC_UNIT_S
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
},
Add GUC variables to control keep-alive times for idle, interval, and count. Oliver Jowett
2005-07-30 17:17:26 +02:00
&tcp_keepalives_idle,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
0, 0, INT_MAX,
NULL, assign_tcp_keepalives_idle, show_tcp_keepalives_idle
Add GUC variables to control keep-alive times for idle, interval, and count. Oliver Jowett
2005-07-30 17:17:26 +02:00
},
{
{"tcp_keepalives_interval", PGC_USERSET, CLIENT_CONN_OTHER,
Revert addition of units to GUC descriptions; doesn't affect postgresql.conf.
2009-04-08 01:27:34 +02:00
gettext_noop("Time between TCP keepalive retransmits."),
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
gettext_noop("A value of 0 uses the system default."),
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
GUC_UNIT_S
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
},
Add GUC variables to control keep-alive times for idle, interval, and count. Oliver Jowett
2005-07-30 17:17:26 +02:00
&tcp_keepalives_interval,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
0, 0, INT_MAX,
NULL, assign_tcp_keepalives_interval, show_tcp_keepalives_interval
Add GUC variables to control keep-alive times for idle, interval, and count. Oliver Jowett
2005-07-30 17:17:26 +02:00
},
Add configuration parameter ssl_renegotiation_limit to control how often we do SSL session key renegotiation. Can be set to 0 to disable renegotiation completely, which is required if a broken SSL library is used (broken patches to CVE-2009-3555 a known cause) or when using a client library that can't do renegotiation.
2010-02-25 14:26:16 +01:00
{
{"ssl_renegotiation_limit", PGC_USERSET, CONN_AUTH_SECURITY,
gettext_noop("Set the amount of traffic to send and receive before renegotiating the encryption keys."),
NULL,
GUC_UNIT_KB,
},
&ssl_renegotiation_limit,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
512 * 1024, 0, MAX_KILOBYTES,
NULL, NULL, NULL
Add configuration parameter ssl_renegotiation_limit to control how often we do SSL session key renegotiation. Can be set to 0 to disable renegotiation completely, which is required if a broken SSL library is used (broken patches to CVE-2009-3555 a known cause) or when using a client library that can't do renegotiation.
2010-02-25 14:26:16 +01:00
},
Add GUC variables to control keep-alive times for idle, interval, and count. Oliver Jowett
2005-07-30 17:17:26 +02:00
{
{"tcp_keepalives_count", PGC_USERSET, CLIENT_CONN_OTHER,
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
gettext_noop("Maximum number of TCP keepalive retransmits."),
gettext_noop("This controls the number of consecutive keepalive retransmits that can be "
"lost before a connection is considered dead. A value of 0 uses the "
"system default."),
},
Add GUC variables to control keep-alive times for idle, interval, and count. Oliver Jowett
2005-07-30 17:17:26 +02:00
&tcp_keepalives_count,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
0, 0, INT_MAX,
NULL, assign_tcp_keepalives_count, show_tcp_keepalives_count
Add GUC variables to control keep-alive times for idle, interval, and count. Oliver Jowett
2005-07-30 17:17:26 +02:00
},
GIN: Generalized Inverted iNdex. text[], int4[], Tsearch2 support for GIN.
2006-05-02 13:28:56 +02:00
{
On platforms that have getrlimit(RLIMIT_STACK), use it to ensure that max_stack_depth is not set to an unsafe value. This commit also provides configure-time checking for <sys/resource.h>, and cleans up some perhaps-unportable code associated with use of that include file and getrlimit().
2006-10-07 21:25:29 +02:00
{"gin_fuzzy_search_limit", PGC_USERSET, CLIENT_CONN_OTHER,
Message style improvements
2006-10-06 19:14:01 +02:00
gettext_noop("Sets the maximum allowed result for exact search by GIN."),
GIN: Generalized Inverted iNdex. text[], int4[], Tsearch2 support for GIN.
2006-05-02 13:28:56 +02:00
NULL,
0
},
&GinFuzzySearchLimit,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
0, 0, INT_MAX,
NULL, NULL, NULL
GIN: Generalized Inverted iNdex. text[], int4[], Tsearch2 support for GIN.
2006-05-02 13:28:56 +02:00
},
Convert effective_cache_size to an integer, for better integration with upcoming units feature.
2006-07-26 13:35:56 +02:00
{
{"effective_cache_size", PGC_USERSET, QUERY_TUNING_COST,
Various improvements to the GUC description strings. Punctuate and capitalize the strings like sentences. Remove unnecessarily specific descriptions of the units used by GUC variables, since we now allow any reasonable unit to be specified.
2006-12-06 18:35:49 +01:00
gettext_noop("Sets the planner's assumption about the size of the disk cache."),
Convert effective_cache_size to an integer, for better integration with upcoming units feature.
2006-07-26 13:35:56 +02:00
gettext_noop("That is, the portion of the kernel's disk cache that "
"will be used for PostgreSQL data files. This is measured in disk "
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
"pages, which are normally 8 kB each."),
GUC_UNIT_BLOCKS,
Convert effective_cache_size to an integer, for better integration with upcoming units feature.
2006-07-26 13:35:56 +02:00
},
&effective_cache_size,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
DEFAULT_EFFECTIVE_CACHE_SIZE, 1, INT_MAX,
NULL, NULL, NULL
Convert effective_cache_size to an integer, for better integration with upcoming units feature.
2006-07-26 13:35:56 +02:00
},
Add new variable "server_version_num", which is almost the same as "server_version" but uses the handy PG_VERSION_NUM which allows apps to do things like if ($version >= 80200) without having to parse apart the value of server_version themselves. Greg Sabino Mullane greg@turnstep.com
2006-09-02 15:12:50 +02:00
{
/* Can't be set in postgresql.conf */
{"server_version_num", PGC_INTERNAL, PRESET_OPTIONS,
gettext_noop("Shows the server version as an integer."),
NULL,
Remove GUC_REPORT for new "server_version_num" GUC variable. Added overhead for every connection, per Tom.
2006-09-02 19:08:10 +02:00
GUC_NOT_IN_SAMPLE | GUC_DISALLOW_IN_FILE
Add new variable "server_version_num", which is almost the same as "server_version" but uses the handy PG_VERSION_NUM which allows apps to do things like if ($version >= 80200) without having to parse apart the value of server_version themselves. Greg Sabino Mullane greg@turnstep.com
2006-09-02 15:12:50 +02:00
},
&server_version_num,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
PG_VERSION_NUM, PG_VERSION_NUM, PG_VERSION_NUM,
NULL, NULL, NULL
Add new variable "server_version_num", which is almost the same as "server_version" but uses the handy PG_VERSION_NUM which allows apps to do things like if ($version >= 80200) without having to parse apart the value of server_version themselves. Greg Sabino Mullane greg@turnstep.com
2006-09-02 15:12:50 +02:00
},
Teach tuplesort.c about "top N" sorting, in which only the first N tuples need be returned. We keep a heap of the current best N tuples and sift-up new tuples into it as we scan the input. For M input tuples this means only about M*log(N) comparisons instead of M*log(M), not to mention a lot less workspace when N is small --- avoiding spill-to-disk for large M is actually the most attractive thing about it. Patch includes planner and executor support for invoking this facility in ORDER BY ... LIMIT queries. Greg Stark, with some editorialization by moi.
2007-05-04 03:13:45 +02:00
Add GUC log_temp_files to log the use of temporary files. Bill Moran
2007-01-09 22:31:17 +01:00
{
Make "log_temp_files" super-user set only, like other logging options. Simon Riggs
2008-08-22 20:47:07 +02:00
{"log_temp_files", PGC_SUSET, LOGGING_WHAT,
Make log_temp_files based on kB, and revert docs & comments to match. Per extensive discussion on pgsql-hackers. We are deliberately not back-patching this even though the behavior of 8.3 and 8.4 is unquestionably broken, for fear of breaking existing users of this parameter. This incompatibility should be release-noted.
2010-07-07 00:55:26 +02:00
gettext_noop("Log the use of temporary files larger than this number of kilobytes."),
Add GUC log_temp_files to log the use of temporary files. Bill Moran
2007-01-09 22:31:17 +01:00
gettext_noop("Zero logs all files. The default is -1 (turning this feature off)."),
Have log_temp_files be in kilobytes, remove trace call.
2007-01-09 23:16:46 +01:00
GUC_UNIT_KB
Add GUC log_temp_files to log the use of temporary files. Bill Moran
2007-01-09 22:31:17 +01:00
},
&log_temp_files,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
-1, -1, INT_MAX,
NULL, NULL, NULL
Add GUC log_temp_files to log the use of temporary files. Bill Moran
2007-01-09 22:31:17 +01:00
},
Add new variable "server_version_num", which is almost the same as "server_version" but uses the handy PG_VERSION_NUM which allows apps to do things like if ($version >= 80200) without having to parse apart the value of server_version themselves. Greg Sabino Mullane greg@turnstep.com
2006-09-02 15:12:50 +02:00
Turn PGBE_ACTIVITY_SIZE into a GUC variable, track_activity_query_size. As the buffer could now be a lot larger than before, and copying it could thus be a lot more expensive than before, use strcpy instead of memcpy to copy the query string, as was already suggested in comments. Also, only copy the PgBackendStatus struct and string if the slot is in use. Patch by Thomas Lee, with some changes by me.
2008-06-30 12:58:47 +02:00
{
{"track_activity_query_size", PGC_POSTMASTER, RESOURCES_MEM,
gettext_noop("Sets the size reserved for pg_stat_activity.current_query, in bytes."),
NULL,
},
&pgstat_track_activity_query_size,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
1024, 100, 102400,
NULL, NULL, NULL
Turn PGBE_ACTIVITY_SIZE into a GUC variable, track_activity_query_size. As the buffer could now be a lot larger than before, and copying it could thus be a lot more expensive than before, use strcpy instead of memcpy to copy the query string, as was already suggested in comments. Also, only copy the PgBackendStatus struct and string if the slot is in use. Patch by Thomas Lee, with some changes by me.
2008-06-30 12:58:47 +02:00
},
Add add_missing_from GUC variable. Nigel J. Andrews
2003-06-12 00:13:22 +02:00
/* End-of-list marker */
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
{NULL, 0, 0, NULL, NULL}, NULL, 0, 0, 0, NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
}
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
};
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
static struct config_real ConfigureNamesReal[] =
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
{
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Add a GUC parameter seq_page_cost, and use that everywhere we formerly assumed that a sequential page fetch has cost 1.0. This patch doesn't in itself change the system's behavior at all, but it opens the door to people adopting other units of measurement for EXPLAIN costs. Also, if we ever decide it's worth inventing per-tablespace access cost settings, this change provides a workable intellectual framework for that.
2006-06-05 04:49:58 +02:00
{"seq_page_cost", PGC_USERSET, QUERY_TUNING_COST,
gettext_noop("Sets the planner's estimate of the cost of a "
"sequentially fetched disk page."),
NULL
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
Add a GUC parameter seq_page_cost, and use that everywhere we formerly assumed that a sequential page fetch has cost 1.0. This patch doesn't in itself change the system's behavior at all, but it opens the door to people adopting other units of measurement for EXPLAIN costs. Also, if we ever decide it's worth inventing per-tablespace access cost settings, this change provides a workable intellectual framework for that.
2006-06-05 04:49:58 +02:00
&seq_page_cost,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
DEFAULT_SEQ_PAGE_COST, 0, DBL_MAX,
NULL, NULL, NULL
Another pgindent run. Fixes enum indenting, and improves #endif spacing. Also adds space for one-line comments.
2001-10-28 07:26:15 +01:00
},
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"random_page_cost", PGC_USERSET, QUERY_TUNING_COST,
Add a GUC parameter seq_page_cost, and use that everywhere we formerly assumed that a sequential page fetch has cost 1.0. This patch doesn't in itself change the system's behavior at all, but it opens the door to people adopting other units of measurement for EXPLAIN costs. Also, if we ever decide it's worth inventing per-tablespace access cost settings, this change provides a workable intellectual framework for that.
2006-06-05 04:49:58 +02:00
gettext_noop("Sets the planner's estimate of the cost of a "
"nonsequentially fetched disk page."),
NULL
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
&random_page_cost,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
DEFAULT_RANDOM_PAGE_COST, 0, DBL_MAX,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"cpu_tuple_cost", PGC_USERSET, QUERY_TUNING_COST,
Add a GUC parameter seq_page_cost, and use that everywhere we formerly assumed that a sequential page fetch has cost 1.0. This patch doesn't in itself change the system's behavior at all, but it opens the door to people adopting other units of measurement for EXPLAIN costs. Also, if we ever decide it's worth inventing per-tablespace access cost settings, this change provides a workable intellectual framework for that.
2006-06-05 04:49:58 +02:00
gettext_noop("Sets the planner's estimate of the cost of "
"processing each tuple (row)."),
NULL
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
&cpu_tuple_cost,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
DEFAULT_CPU_TUPLE_COST, 0, DBL_MAX,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"cpu_index_tuple_cost", PGC_USERSET, QUERY_TUNING_COST,
Add a GUC parameter seq_page_cost, and use that everywhere we formerly assumed that a sequential page fetch has cost 1.0. This patch doesn't in itself change the system's behavior at all, but it opens the door to people adopting other units of measurement for EXPLAIN costs. Also, if we ever decide it's worth inventing per-tablespace access cost settings, this change provides a workable intellectual framework for that.
2006-06-05 04:49:58 +02:00
gettext_noop("Sets the planner's estimate of the cost of "
"processing each index entry during an index scan."),
NULL
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
&cpu_index_tuple_cost,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
DEFAULT_CPU_INDEX_TUPLE_COST, 0, DBL_MAX,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"cpu_operator_cost", PGC_USERSET, QUERY_TUNING_COST,
Add a GUC parameter seq_page_cost, and use that everywhere we formerly assumed that a sequential page fetch has cost 1.0. This patch doesn't in itself change the system's behavior at all, but it opens the door to people adopting other units of measurement for EXPLAIN costs. Also, if we ever decide it's worth inventing per-tablespace access cost settings, this change provides a workable intellectual framework for that.
2006-06-05 04:49:58 +02:00
gettext_noop("Sets the planner's estimate of the cost of "
"processing each operator or function call."),
NULL
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
&cpu_operator_cost,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
DEFAULT_CPU_OPERATOR_COST, 0, DBL_MAX,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
Allow the planner's estimate of the fraction of a cursor's rows that will be retrieved to be controlled through a GUC variable. Robert Hell
2008-05-02 23:26:10 +02:00
{
{"cursor_tuple_fraction", PGC_USERSET, QUERY_TUNING_OTHER,
gettext_noop("Sets the planner's estimate of the fraction of "
"a cursor's rows that will be retrieved."),
NULL
},
&cursor_tuple_fraction,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
DEFAULT_CURSOR_TUPLE_FRACTION, 0.0, 1.0,
NULL, NULL, NULL
Allow the planner's estimate of the fraction of a cursor's rows that will be retrieved to be controlled through a GUC variable. Robert Hell
2008-05-02 23:26:10 +02:00
},
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"geqo_selection_bias", PGC_USERSET, QUERY_TUNING_GEQO,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("GEQO: selective pressure within the population."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL
},
&Geqo_selection_bias,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
DEFAULT_GEQO_SELECTION_BIAS,
MIN_GEQO_SELECTION_BIAS, MAX_GEQO_SELECTION_BIAS,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
Make GEQO's planning deterministic by having it start from a predictable random number seed each time. This is how it used to work years ago, but we got rid of the seed reset because it was resetting the main random() sequence and thus having undesirable effects on the rest of the system. To fix, establish a private random number state for each execution of geqo(), and initialize the state using the new GUC variable geqo_seed. People who want to experiment with different random searches can do so by changing geqo_seed, but you'll always get the same plan for the same value of geqo_seed (if holding all other planner inputs constant, of course). The new state is kept in PlannerInfo by adding a "void *" field reserved for use by join_search hooks. Most of the rather bulky code changes in this commit are just arranging to pass PlannerInfo around to all the GEQO functions (many of which formerly didn't receive it). Andres Freund, with some editorialization by Tom
2009-07-16 22:55:44 +02:00
{
{"geqo_seed", PGC_USERSET, QUERY_TUNING_GEQO,
gettext_noop("GEQO: seed for random path selection."),
NULL
},
&Geqo_seed,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
0.0, 0.0, 1.0,
NULL, NULL, NULL
Make GEQO's planning deterministic by having it start from a predictable random number seed each time. This is how it used to work years ago, but we got rid of the seed reset because it was resetting the main random() sequence and thus having undesirable effects on the rest of the system. To fix, establish a private random number state for each execution of geqo(), and initialize the state using the new GUC variable geqo_seed. People who want to experiment with different random searches can do so by changing geqo_seed, but you'll always get the same plan for the same value of geqo_seed (if holding all other planner inputs constant, of course). The new state is kept in PlannerInfo by adding a "void *" field reserved for use by join_search hooks. Most of the rather bulky code changes in this commit are just arranging to pass PlannerInfo around to all the GEQO functions (many of which formerly didn't receive it). Andres Freund, with some editorialization by Tom
2009-07-16 22:55:44 +02:00
},
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
Replace the BufMgrLock with separate locks on the lookup hashtable and the freelist, plus per-buffer spinlocks that protect access to individual shared buffer headers. This requires abandoning a global freelist (since the freelist is a global contention point), which shoots down ARC and 2Q as well as plain LRU management. Adopt a clock sweep algorithm instead. Preliminary results show substantial improvement in multi-backend situations.
2005-03-04 21:21:07 +01:00
{
Add new GUC categories corresponding to sections in docs, and move description for vacuum_defer_cleanup_age to the correct category. Sections in postgresql.conf are also sorted in the same order with docs. Per gripe by Fujii Masao, suggestion by Heikki Linnakangas, and patch by me.
2010-06-15 09:52:11 +02:00
{"bgwriter_lru_multiplier", PGC_SIGHUP, RESOURCES_BGWRITER,
Improve "bgwriter_lru_multiplier" GUC description.
2008-03-06 17:31:42 +01:00
gettext_noop("Multiple of the average buffer usage to free per round."),
Replace the BufMgrLock with separate locks on the lookup hashtable and the freelist, plus per-buffer spinlocks that protect access to individual shared buffer headers. This requires abandoning a global freelist (since the freelist is a global contention point), which shoots down ARC and 2Q as well as plain LRU management. Adopt a clock sweep algorithm instead. Preliminary results show substantial improvement in multi-backend situations.
2005-03-04 21:21:07 +01:00
NULL
},
Just-in-time background writing strategy. This code avoids re-scanning buffers that cannot possibly need to be cleaned, and estimates how many buffers it should try to clean based on moving averages of recent allocation requests and density of reusable buffers. The patch also adds a couple more columns to pg_stat_bgwriter to help measure the effectiveness of the bgwriter. Greg Smith, building on his own work and ideas from several other people, in particular a much older patch from Itagaki Takahiro.
2007-09-25 22:03:38 +02:00
&bgwriter_lru_multiplier,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
2.0, 0.0, 10.0,
NULL, NULL, NULL
Replace the BufMgrLock with separate locks on the lookup hashtable and the freelist, plus per-buffer spinlocks that protect access to individual shared buffer headers. This requires abandoning a global freelist (since the freelist is a global contention point), which shoots down ARC and 2Q as well as plain LRU management. Adopt a clock sweep algorithm instead. Preliminary results show substantial improvement in multi-backend situations.
2005-03-04 21:21:07 +01:00
},
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"seed", PGC_USERSET, UNGROUPED,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Sets the seed for random-number generation."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL,
GUC_NO_SHOW_ALL | GUC_NO_RESET_ALL | GUC_NOT_IN_SAMPLE | GUC_DISALLOW_IN_FILE
},
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
&phony_random_seed,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
0.0, -1.0, 1.0,
check_random_seed, assign_random_seed, show_random_seed
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
},
Integrate autovacuum functionality into the backend. There's still a few loose ends to be dealt with, but it seems to work. Alvaro Herrera, based on the contrib code by Matthew O'Connor.
2005-07-14 07:13:45 +02:00
{
{"autovacuum_vacuum_scale_factor", PGC_SIGHUP, AUTOVACUUM,
gettext_noop("Number of tuple updates or deletes prior to vacuum as a fraction of reltuples."),
NULL
},
&autovacuum_vac_scale,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
0.2, 0.0, 100.0,
NULL, NULL, NULL
Integrate autovacuum functionality into the backend. There's still a few loose ends to be dealt with, but it seems to work. Alvaro Herrera, based on the contrib code by Matthew O'Connor.
2005-07-14 07:13:45 +02:00
},
{
{"autovacuum_analyze_scale_factor", PGC_SIGHUP, AUTOVACUUM,
gettext_noop("Number of tuple inserts, updates or deletes prior to analyze as a fraction of reltuples."),
NULL
},
&autovacuum_anl_scale,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
0.1, 0.0, 100.0,
NULL, NULL, NULL
Integrate autovacuum functionality into the backend. There's still a few loose ends to be dealt with, but it seems to work. Alvaro Herrera, based on the contrib code by Matthew O'Connor.
2005-07-14 07:13:45 +02:00
},
Implement "distributed" checkpoints in which the checkpoint I/O is spread over a fairly long period of time, rather than being spat out in a burst. This happens only for background checkpoints carried out by the bgwriter; other cases, such as a shutdown checkpoint, are still done at full speed. Remove the "all buffers" scan in the bgwriter, and associated stats infrastructure, since this seems no longer very useful when the checkpoint itself is properly throttled. Original patch by Itagaki Takahiro, reworked by Heikki Linnakangas, and some minor API editorialization by me.
2007-06-28 02:02:40 +02:00
{
{"checkpoint_completion_target", PGC_SIGHUP, WAL_CHECKPOINTS,
gettext_noop("Time spent flushing dirty buffers during checkpoint, as fraction of checkpoint interval."),
NULL
},
&CheckPointCompletionTarget,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
0.5, 0.0, 1.0,
NULL, NULL, NULL
Implement "distributed" checkpoints in which the checkpoint I/O is spread over a fairly long period of time, rather than being spat out in a burst. This happens only for background checkpoints carried out by the bgwriter; other cases, such as a shutdown checkpoint, are still done at full speed. Remove the "all buffers" scan in the bgwriter, and associated stats infrastructure, since this seems no longer very useful when the checkpoint itself is properly throttled. Original patch by Itagaki Takahiro, reworked by Heikki Linnakangas, and some minor API editorialization by me.
2007-06-28 02:02:40 +02:00
},
Add add_missing_from GUC variable. Nigel J. Andrews
2003-06-12 00:13:22 +02:00
/* End-of-list marker */
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
{NULL, 0, 0, NULL, NULL}, NULL, 0.0, 0.0, 0.0, NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
}
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
};
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
static struct config_string ConfigureNamesString[] =
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
{
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
{
Add new GUC categories corresponding to sections in docs, and move description for vacuum_defer_cleanup_age to the correct category. Sections in postgresql.conf are also sorted in the same order with docs. Per gripe by Fujii Masao, suggestion by Heikki Linnakangas, and patch by me.
2010-06-15 09:52:11 +02:00
{"archive_command", PGC_SIGHUP, WAL_ARCHIVING,
Various improvements to the GUC description strings. Punctuate and capitalize the strings like sentences. Remove unnecessarily specific descriptions of the units used by GUC variables, since we now allow any reasonable unit to be specified.
2006-12-06 18:35:49 +01:00
gettext_noop("Sets the shell command that will be called to archive a WAL file."),
NULL
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
},
&XLogArchiveCommand,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
"",
NULL, NULL, show_archive_command
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
},
XLOG file archiving and point-in-time recovery. There are still some loose ends and a glaring lack of documentation, but it basically works. Simon Riggs with some editorialization by Tom Lane.
2004-07-19 04:47:16 +02:00
Commit to match discussed elog() changes. Only update is that LOG is now just below FATAL in server_min_messages. Added more text to highlight ordering difference between it and client_min_messages. --------------------------------------------------------------------------- REALLYFATAL => PANIC STOP => PANIC New INFO level the prints to client by default New LOG level the prints to server log by default Cause VACUUM information to print only to the client NOTICE => INFO where purely information messages are sent DEBUG => LOG for purely server status messages DEBUG removed, kept as backward compatible DEBUG5, DEBUG4, DEBUG3, DEBUG2, DEBUG1 added DebugLvl removed in favor of new DEBUG[1-5] symbols New server_min_messages GUC parameter with values: DEBUG[5-1], INFO, NOTICE, ERROR, LOG, FATAL, PANIC New client_min_messages GUC parameter with values: DEBUG[5-1], LOG, INFO, NOTICE, ERROR, FATAL, PANIC Server startup now logged with LOG instead of DEBUG Remove debug_level GUC parameter elog() numbers now start at 10 Add test to print error message if older elog() values are passed to elog() Bootstrap mode now has a -d that requires an argument, like postmaster
2002-03-02 22:39:36 +01:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"client_encoding", PGC_USERSET, CLIENT_CONN_LOCALE,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Sets the client's character set encoding."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL,
Fix bug that allowed any logged-in user to SET ROLE to any other database user id (CVE-2006-0553). Also fix related bug in SET SESSION AUTHORIZATION that allows unprivileged users to crash the server, if it has been compiled with Asserts enabled. The escalation-of-privilege risk exists only in 8.1.0-8.1.2. However, the Assert-crash risk exists in all releases back to 7.3. Thanks to Akio Ishida for reporting this problem.
2006-02-12 23:32:43 +01:00
GUC_IS_NAME | GUC_REPORT
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
In the continuing saga of FE/BE protocol revisions, add reporting of initial values and runtime changes in selected parameters. This gets rid of the need for an initial 'select pg_client_encoding()' query in libpq, bringing us back to one message transmitted in each direction for a standard connection startup. To allow server version to be sent using the same GUC mechanism that handles other parameters, invent the concept of a never-settable GUC parameter: you can 'show server_version' but it's not settable by any GUC input source. Create 'lc_collate' and 'lc_ctype' never-settable parameters so that people can find out these settings without need for pg_controldata. (These side ideas were all discussed some time ago in pgsql-hackers, but not yet implemented.)
2003-04-25 21:45:10 +02:00
&client_encoding_string,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
"SQL_ASCII",
check_client_encoding, assign_client_encoding, NULL
Commit to match discussed elog() changes. Only update is that LOG is now just below FATAL in server_min_messages. Added more text to highlight ordering difference between it and client_min_messages. --------------------------------------------------------------------------- REALLYFATAL => PANIC STOP => PANIC New INFO level the prints to client by default New LOG level the prints to server log by default Cause VACUUM information to print only to the client NOTICE => INFO where purely information messages are sent DEBUG => LOG for purely server status messages DEBUG removed, kept as backward compatible DEBUG5, DEBUG4, DEBUG3, DEBUG2, DEBUG1 added DebugLvl removed in favor of new DEBUG[1-5] symbols New server_min_messages GUC parameter with values: DEBUG[5-1], INFO, NOTICE, ERROR, LOG, FATAL, PANIC New client_min_messages GUC parameter with values: DEBUG[5-1], LOG, INFO, NOTICE, ERROR, FATAL, PANIC Server startup now logged with LOG instead of DEBUG Remove debug_level GUC parameter elog() numbers now start at 10 Add test to print error message if older elog() values are passed to elog() Bootstrap mode now has a -d that requires an argument, like postmaster
2002-03-02 22:39:36 +01:00
},
Add: #log_line_prefix = '' # e.g. '<%u%%%d> ' # %u=user name %d=database name # %r=remote host and port # %p=PID %t=timestamp %i=command tag # %c=session id %l=session line number # %s=session start timestamp # %x=stop here in non-session processes # %%='%' Andrew Dunstan
2004-03-09 05:43:07 +01:00
{
{"log_line_prefix", PGC_SIGHUP, LOGGING_WHAT,
Various improvements to the GUC description strings. Punctuate and capitalize the strings like sentences. Remove unnecessarily specific descriptions of the units used by GUC variables, since we now allow any reasonable unit to be specified.
2006-12-06 18:35:49 +01:00
gettext_noop("Controls information prefixed to each log line."),
gettext_noop("If blank, no prefix is used.")
Add: #log_line_prefix = '' # e.g. '<%u%%%d> ' # %u=user name %d=database name # %r=remote host and port # %p=PID %t=timestamp %i=command tag # %c=session id %l=session line number # %s=session start timestamp # %x=stop here in non-session processes # %%='%' Andrew Dunstan
2004-03-09 05:43:07 +01:00
},
&Log_line_prefix,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
"",
NULL, NULL, NULL
Add: #log_line_prefix = '' # e.g. '<%u%%%d> ' # %u=user name %d=database name # %r=remote host and port # %p=PID %t=timestamp %i=command tag # %c=session id %l=session line number # %s=session start timestamp # %x=stop here in non-session processes # %%='%' Andrew Dunstan
2004-03-09 05:43:07 +01:00
},
Switch over to using the src/timezone functions for formatting timestamps displayed in the postmaster log. This avoids Windows-specific problems with localized time zone names that are in the wrong encoding, and generally seems like a good idea to forestall other potential platform-dependent issues. To preserve the existing behavior that all backends will log in the same time zone, create a new GUC variable log_timezone that can only be changed on a system-wide basis, and reference log-related calculations to that zone instead of the TimeZone variable. This fixes the issue reported by Hiroshi Saito that timestamps printed by xlog.c startup could be improperly localized on Windows. We still need a simpler patch for that problem in the back branches, however.
2007-08-04 03:26:54 +02:00
{
{"log_timezone", PGC_SIGHUP, LOGGING_WHAT,
gettext_noop("Sets the time zone to use in log messages."),
NULL
},
&log_timezone_string,
Simplify handling of the timezone GUC by making initdb choose the default. We were doing some amazingly complicated things in order to avoid running the very expensive identify_system_timezone() procedure during GUC initialization. But there is an obvious fix for that, which is to do it once during initdb and have initdb install the system-specific default into postgresql.conf, as it already does for most other GUC variables that need system-environment-dependent defaults. This means that the timezone (and log_timezone) settings no longer have any magic behavior in the server. Per discussion.
2011-09-09 23:59:11 +02:00
"GMT",
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
check_log_timezone, assign_log_timezone, show_log_timezone
Switch over to using the src/timezone functions for formatting timestamps displayed in the postmaster log. This avoids Windows-specific problems with localized time zone names that are in the wrong encoding, and generally seems like a good idea to forestall other potential platform-dependent issues. To preserve the existing behavior that all backends will log in the same time zone, create a new GUC variable log_timezone that can only be changed on a system-wide basis, and reference log-related calculations to that zone instead of the TimeZone variable. This fixes the issue reported by Hiroshi Saito that timestamps printed by xlog.c startup could be improperly localized on Windows. We still need a simpler patch for that problem in the back branches, however.
2007-08-04 03:26:54 +02:00
},
Add: #log_line_prefix = '' # e.g. '<%u%%%d> ' # %u=user name %d=database name # %r=remote host and port # %p=PID %t=timestamp %i=command tag # %c=session id %l=session line number # %s=session start timestamp # %x=stop here in non-session processes # %%='%' Andrew Dunstan
2004-03-09 05:43:07 +01:00
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"DateStyle", PGC_USERSET, CLIENT_CONN_LOCALE,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Sets the display format for date and time values."),
Fix my own mistake in GUC variable annotation.
2003-08-01 03:23:11 +02:00
gettext_noop("Also controls interpretation of ambiguous "
Message editing: remove gratuitous variations in message wording, standardize terms, add some clarifications, fix some untranslatable attempts at dynamic message building.
2003-09-25 08:58:07 +02:00
"date inputs."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
GUC_LIST_INPUT | GUC_REPORT
},
In the continuing saga of FE/BE protocol revisions, add reporting of initial values and runtime changes in selected parameters. This gets rid of the need for an initial 'select pg_client_encoding()' query in libpq, bringing us back to one message transmitted in each direction for a standard connection startup. To allow server version to be sent using the same GUC mechanism that handles other parameters, invent the concept of a never-settable GUC parameter: you can 'show server_version' but it's not settable by any GUC input source. Create 'lc_collate' and 'lc_ctype' never-settable parameters so that people can find out these settings without need for pg_controldata. (These side ideas were all discussed some time ago in pgsql-hackers, but not yet implemented.)
2003-04-25 21:45:10 +02:00
&datestyle_string,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
"ISO, MDY",
check_datestyle, assign_datestyle, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
Add dynamic_library_path parameter and automatic appending of shared library extension.
2001-05-17 19:44:18 +02:00
Create 'default_tablespace' GUC variable that supplies a TABLESPACE clause implicitly whenever one is not given explicitly. Remove concept of a schema having an associated tablespace, and simplify the rules for selecting a default tablespace for a table or index. It's now just (a) explicit TABLESPACE clause; (b) default_tablespace if that's not an empty string; (c) database's default. This will allow pg_dump to use SET commands instead of tablespace clauses to determine object locations (but I didn't actually make it do so). All per recent discussions.
2004-11-05 20:17:13 +01:00
{
{"default_tablespace", PGC_USERSET, CLIENT_CONN_STATEMENT,
gettext_noop("Sets the default tablespace to create tables and indexes in."),
Fix bug that allowed any logged-in user to SET ROLE to any other database user id (CVE-2006-0553). Also fix related bug in SET SESSION AUTHORIZATION that allows unprivileged users to crash the server, if it has been compiled with Asserts enabled. The escalation-of-privilege risk exists only in 8.1.0-8.1.2. However, the Assert-crash risk exists in all releases back to 7.3. Thanks to Akio Ishida for reporting this problem.
2006-02-12 23:32:43 +01:00
gettext_noop("An empty string selects the database's default tablespace."),
pgindent run for 8.2.
2006-10-04 02:30:14 +02:00
GUC_IS_NAME
Create 'default_tablespace' GUC variable that supplies a TABLESPACE clause implicitly whenever one is not given explicitly. Remove concept of a schema having an associated tablespace, and simplify the rules for selecting a default tablespace for a table or index. It's now just (a) explicit TABLESPACE clause; (b) default_tablespace if that's not an empty string; (c) database's default. This will allow pg_dump to use SET commands instead of tablespace clauses to determine object locations (but I didn't actually make it do so). All per recent discussions.
2004-11-05 20:17:13 +01:00
},
&default_tablespace,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
"",
check_default_tablespace, NULL, NULL
Create 'default_tablespace' GUC variable that supplies a TABLESPACE clause implicitly whenever one is not given explicitly. Remove concept of a schema having an associated tablespace, and simplify the rules for selecting a default tablespace for a table or index. It's now just (a) explicit TABLESPACE clause; (b) default_tablespace if that's not an empty string; (c) database's default. This will allow pg_dump to use SET commands instead of tablespace clauses to determine object locations (but I didn't actually make it do so). All per recent discussions.
2004-11-05 20:17:13 +01:00
},
Create a GUC parameter temp_tablespaces that allows selection of the tablespace(s) in which to store temp tables and temporary files. This is a list to allow spreading the load across multiple tablespaces (a random list element is chosen each time a temp object is to be created). Temp files are not stored in per-database pgsql_tmp/ directories anymore, but per-tablespace directories. Jaime Casanova and Albert Cervera, with review by Bernd Helmle and Tom Lane.
2007-06-03 19:08:34 +02:00
{
{"temp_tablespaces", PGC_USERSET, CLIENT_CONN_STATEMENT,
gettext_noop("Sets the tablespace(s) to use for temporary tables and sort files."),
NULL,
GUC_LIST_INPUT | GUC_LIST_QUOTE
},
&temp_tablespaces,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
"",
check_temp_tablespaces, assign_temp_tablespaces, NULL
Create a GUC parameter temp_tablespaces that allows selection of the tablespace(s) in which to store temp tables and temporary files. This is a list to allow spreading the load across multiple tablespaces (a random list element is chosen each time a temp object is to be created). Temp files are not stored in per-database pgsql_tmp/ directories anymore, but per-tablespace directories. Jaime Casanova and Albert Cervera, with review by Bernd Helmle and Tom Lane.
2007-06-03 19:08:34 +02:00
},
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"dynamic_library_path", PGC_SUSET, CLIENT_CONN_OTHER,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Sets the path for dynamically loadable modules."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
gettext_noop("If a dynamically loadable module needs to be opened and "
More message editing, some suggested by Alvaro Herrera
2003-09-29 02:05:25 +02:00
"the specified name does not have a directory component (i.e., the "
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
"name does not contain a slash), the system will search this path for "
Add a GUC_SUPERUSER_ONLY flag to mark GUC variables that should not be examinable by non-superusers, and use it to protect the recently-added GUC variables for data directory and config files. For now I have only flagged those variables that could be used to deduce something about the server's filesystem layout, but possibly we should also mark vars related to logging settings and other admin-only information?
2004-10-22 21:48:19 +02:00
"the specified file."),
GUC_SUPERUSER_ONLY
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
&Dynamic_library_path,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
"$libdir",
NULL, NULL, NULL
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
},
{
Make it possible to change Kerberos/GSSAPI parameters without restarting the postmaster. They are only used in backend processes, so it's just a matter of re-labeling the GUCs.
2009-01-02 11:33:20 +01:00
{"krb_server_keyfile", PGC_SIGHUP, CONN_AUTH_SECURITY,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Sets the location of the Kerberos server key file."),
Add a GUC_SUPERUSER_ONLY flag to mark GUC variables that should not be examinable by non-superusers, and use it to protect the recently-added GUC variables for data directory and config files. For now I have only flagged those variables that could be used to deduce something about the server's filesystem layout, but possibly we should also mark vars related to logging settings and other admin-only information?
2004-10-22 21:48:19 +02:00
NULL,
GUC_SUPERUSER_ONLY
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
&pg_krb_server_keyfile,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
PG_KRB_SRVTAB,
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
Small polishing of syslog facility and ident settings. Only allow setting at postmaster start, rename syslog_progid to syslog_ident, since syslog itself uses that term, fix doc markup.
2000-11-14 20:13:27 +01:00
Allow kerberos name and username case sensitivity to be specified from postgresql.conf. --------------------------------------------------------------------------- Here's an updated version of the patch, with the following changes: 1) No longer uses "service name" as "application version". It's instead hardcoded as "postgres". It could be argued that this part should be backpatched to 8.0, but it doesn't make a big difference until you can start changing it with GUC / connection parameters. This change only affects kerberos 5, not 4. 2) Now downcases kerberos usernames when the client is running on win32. 3) Adds guc option for "krb_caseins_users" to make the server ignore case mismatch which is required by some KDCs such as Active Directory. Off by default, per discussion with Tom. This change only affects kerberos 5, not 4. 4) Updated so it doesn't conflict with the rendevouz/bonjour patch already in ;-) Magnus Hagander
2005-06-04 22:42:43 +02:00
{
Make it possible to change Kerberos/GSSAPI parameters without restarting the postmaster. They are only used in backend processes, so it's just a matter of re-labeling the GUCs.
2009-01-02 11:33:20 +01:00
{"krb_srvname", PGC_SIGHUP, CONN_AUTH_SECURITY,
Allow kerberos name and username case sensitivity to be specified from postgresql.conf. --------------------------------------------------------------------------- Here's an updated version of the patch, with the following changes: 1) No longer uses "service name" as "application version". It's instead hardcoded as "postgres". It could be argued that this part should be backpatched to 8.0, but it doesn't make a big difference until you can start changing it with GUC / connection parameters. This change only affects kerberos 5, not 4. 2) Now downcases kerberos usernames when the client is running on win32. 3) Adds guc option for "krb_caseins_users" to make the server ignore case mismatch which is required by some KDCs such as Active Directory. Off by default, per discussion with Tom. This change only affects kerberos 5, not 4. 4) Updated so it doesn't conflict with the rendevouz/bonjour patch already in ;-) Magnus Hagander
2005-06-04 22:42:43 +02:00
gettext_noop("Sets the name of the Kerberos service."),
NULL
},
&pg_krb_srvnam,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
PG_KRB_SRVNAM,
NULL, NULL, NULL
Allow kerberos name and username case sensitivity to be specified from postgresql.conf. --------------------------------------------------------------------------- Here's an updated version of the patch, with the following changes: 1) No longer uses "service name" as "application version". It's instead hardcoded as "postgres". It could be argued that this part should be backpatched to 8.0, but it doesn't make a big difference until you can start changing it with GUC / connection parameters. This change only affects kerberos 5, not 4. 2) Now downcases kerberos usernames when the client is running on win32. 3) Adds guc option for "krb_caseins_users" to make the server ignore case mismatch which is required by some KDCs such as Active Directory. Off by default, per discussion with Tom. This change only affects kerberos 5, not 4. 4) Updated so it doesn't conflict with the rendevouz/bonjour patch already in ;-) Magnus Hagander
2005-06-04 22:42:43 +02:00
},
Add GUC parameter to control rendezvous name.
2003-07-22 22:29:13 +02:00
{
Rename Rendezvous to Bonjour to match OS/X renaming.
2005-05-15 02:26:19 +02:00
{"bonjour_name", PGC_POSTMASTER, CONN_AUTH_SETTINGS,
Add a boolean GUC parameter "bonjour" to control whether a Bonjour-enabled build actually attempts to advertise itself via Bonjour. Formerly it always did so, which meant that packagers had to decide for their users whether this behavior was wanted or not. The default is "off" to be on the safe side, though this represents a change in the default behavior of a Bonjour-enabled build. Per discussion.
2009-09-08 19:08:36 +02:00
gettext_noop("Sets the Bonjour service name."),
Add GUC parameter to control rendezvous name.
2003-07-22 22:29:13 +02:00
NULL
},
Rename Rendezvous to Bonjour to match OS/X renaming.
2005-05-15 02:26:19 +02:00
&bonjour_name,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
"",
NULL, NULL, NULL
Add GUC parameter to control rendezvous name.
2003-07-22 22:29:13 +02:00
},
Powerup defaults for LC_foo GUC variables should match what main.c does.
2002-09-10 18:09:02 +02:00
/* See main.c about why defaults for LC_foo are not all alike */
In the continuing saga of FE/BE protocol revisions, add reporting of initial values and runtime changes in selected parameters. This gets rid of the need for an initial 'select pg_client_encoding()' query in libpq, bringing us back to one message transmitted in each direction for a standard connection startup. To allow server version to be sent using the same GUC mechanism that handles other parameters, invent the concept of a never-settable GUC parameter: you can 'show server_version' but it's not settable by any GUC input source. Create 'lc_collate' and 'lc_ctype' never-settable parameters so that people can find out these settings without need for pg_controldata. (These side ideas were all discussed some time ago in pgsql-hackers, but not yet implemented.)
2003-04-25 21:45:10 +02:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"lc_collate", PGC_INTERNAL, CLIENT_CONN_LOCALE,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Shows the collation order locale."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL,
GUC_NOT_IN_SAMPLE | GUC_DISALLOW_IN_FILE
},
&locale_collate,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
"C",
NULL, NULL, NULL
In the continuing saga of FE/BE protocol revisions, add reporting of initial values and runtime changes in selected parameters. This gets rid of the need for an initial 'select pg_client_encoding()' query in libpq, bringing us back to one message transmitted in each direction for a standard connection startup. To allow server version to be sent using the same GUC mechanism that handles other parameters, invent the concept of a never-settable GUC parameter: you can 'show server_version' but it's not settable by any GUC input source. Create 'lc_collate' and 'lc_ctype' never-settable parameters so that people can find out these settings without need for pg_controldata. (These side ideas were all discussed some time ago in pgsql-hackers, but not yet implemented.)
2003-04-25 21:45:10 +02:00
},
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"lc_ctype", PGC_INTERNAL, CLIENT_CONN_LOCALE,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Shows the character classification and case conversion locale."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL,
GUC_NOT_IN_SAMPLE | GUC_DISALLOW_IN_FILE
},
&locale_ctype,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
"C",
NULL, NULL, NULL
In the continuing saga of FE/BE protocol revisions, add reporting of initial values and runtime changes in selected parameters. This gets rid of the need for an initial 'select pg_client_encoding()' query in libpq, bringing us back to one message transmitted in each direction for a standard connection startup. To allow server version to be sent using the same GUC mechanism that handles other parameters, invent the concept of a never-settable GUC parameter: you can 'show server_version' but it's not settable by any GUC input source. Create 'lc_collate' and 'lc_ctype' never-settable parameters so that people can find out these settings without need for pg_controldata. (These side ideas were all discussed some time ago in pgsql-hackers, but not yet implemented.)
2003-04-25 21:45:10 +02:00
},
Locale support is on by default. The choice of locale is done in initdb and/or with GUC variables.
2002-04-03 07:39:33 +02:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"lc_messages", PGC_SUSET, CLIENT_CONN_LOCALE,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Sets the language in which messages are displayed."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL
},
&locale_messages,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
"",
check_locale_messages, assign_locale_messages, NULL
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
},
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"lc_monetary", PGC_USERSET, CLIENT_CONN_LOCALE,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Sets the locale for formatting monetary amounts."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL
},
&locale_monetary,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
"C",
check_locale_monetary, assign_locale_monetary, NULL
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
},
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"lc_numeric", PGC_USERSET, CLIENT_CONN_LOCALE,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Sets the locale for formatting numbers."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL
},
&locale_numeric,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
"C",
check_locale_numeric, assign_locale_numeric, NULL
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
},
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"lc_time", PGC_USERSET, CLIENT_CONN_LOCALE,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Sets the locale for formatting date and time values."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL
},
&locale_time,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
"C",
check_locale_time, assign_locale_time, NULL
Locale support is on by default. The choice of locale is done in initdb and/or with GUC variables.
2002-04-03 07:39:33 +02:00
},
> I can see a couple possible downsides: (a) the library might have some > weird behavior across fork boundaries; (b) the additional memory space > that has to be duplicated into child processes will cost something per > child launch, even if the child never uses it. But these are only > arguments that it might not *always* be a prudent thing to do, not that > we shouldn't give the DBA the tool to do it if he wants. So fire away. Here is a patch for the above, including a documentation update. It creates a new GUC variable "preload_libraries", that accepts a list in the form: preload_libraries = '$libdir/mylib1:initfunc,$libdir/mylib2' If ":initfunc" is omitted or not found, no initialization function is executed, but the library is still preloaded. If "$libdir/mylib" isn't found, the postmaster refuses to start. In my testing with PL/R, it reduces the first call to a PL/R function (after connecting) from almost 2 seconds, down to about 8 ms. Joe Conway
2003-03-20 05:51:44 +01:00
{
Add server support for "plugin" libraries that can be used for add-on tasks such as debugging and performance measurement. This consists of two features: a table of "rendezvous variables" that allows separately-loaded shared libraries to communicate, and a new GUC setting "local_preload_libraries" that allows libraries to be loaded into specific sessions without explicit cooperation from the client application. To make local_preload_libraries as flexible as possible, we do not restrict its use to superusers; instead, it is restricted to load only libraries stored in $libdir/plugins/. The existing LOAD command has also been modified to allow non-superusers to LOAD libraries stored in this directory. This patch also renames the existing GUC variable preload_libraries to shared_preload_libraries (after a suggestion by Simon Riggs) and does some code refactoring in dfmgr.c to improve clarity. Korry Douglas, with a little help from Tom Lane.
2006-08-15 20:26:59 +02:00
{"shared_preload_libraries", PGC_POSTMASTER, RESOURCES_KERNEL,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Lists shared libraries to preload into server."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL,
Add a GUC_SUPERUSER_ONLY flag to mark GUC variables that should not be examinable by non-superusers, and use it to protect the recently-added GUC variables for data directory and config files. For now I have only flagged those variables that could be used to deduce something about the server's filesystem layout, but possibly we should also mark vars related to logging settings and other admin-only information?
2004-10-22 21:48:19 +02:00
GUC_LIST_INPUT | GUC_LIST_QUOTE | GUC_SUPERUSER_ONLY
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
Add server support for "plugin" libraries that can be used for add-on tasks such as debugging and performance measurement. This consists of two features: a table of "rendezvous variables" that allows separately-loaded shared libraries to communicate, and a new GUC setting "local_preload_libraries" that allows libraries to be loaded into specific sessions without explicit cooperation from the client application. To make local_preload_libraries as flexible as possible, we do not restrict its use to superusers; instead, it is restricted to load only libraries stored in $libdir/plugins/. The existing LOAD command has also been modified to allow non-superusers to LOAD libraries stored in this directory. This patch also renames the existing GUC variable preload_libraries to shared_preload_libraries (after a suggestion by Simon Riggs) and does some code refactoring in dfmgr.c to improve clarity. Korry Douglas, with a little help from Tom Lane.
2006-08-15 20:26:59 +02:00
&shared_preload_libraries_string,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
"",
NULL, NULL, NULL
Add server support for "plugin" libraries that can be used for add-on tasks such as debugging and performance measurement. This consists of two features: a table of "rendezvous variables" that allows separately-loaded shared libraries to communicate, and a new GUC setting "local_preload_libraries" that allows libraries to be loaded into specific sessions without explicit cooperation from the client application. To make local_preload_libraries as flexible as possible, we do not restrict its use to superusers; instead, it is restricted to load only libraries stored in $libdir/plugins/. The existing LOAD command has also been modified to allow non-superusers to LOAD libraries stored in this directory. This patch also renames the existing GUC variable preload_libraries to shared_preload_libraries (after a suggestion by Simon Riggs) and does some code refactoring in dfmgr.c to improve clarity. Korry Douglas, with a little help from Tom Lane.
2006-08-15 20:26:59 +02:00
},
{
{"local_preload_libraries", PGC_BACKEND, CLIENT_CONN_OTHER,
gettext_noop("Lists shared libraries to preload into each backend."),
NULL,
GUC_LIST_INPUT | GUC_LIST_QUOTE
},
&local_preload_libraries_string,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
"",
NULL, NULL, NULL
> I can see a couple possible downsides: (a) the library might have some > weird behavior across fork boundaries; (b) the additional memory space > that has to be duplicated into child processes will cost something per > child launch, even if the child never uses it. But these are only > arguments that it might not *always* be a prudent thing to do, not that > we shouldn't give the DBA the tool to do it if he wants. So fire away. Here is a patch for the above, including a documentation update. It creates a new GUC variable "preload_libraries", that accepts a list in the form: preload_libraries = '$libdir/mylib1:initfunc,$libdir/mylib2' If ":initfunc" is omitted or not found, no initialization function is executed, but the library is still preloaded. If "$libdir/mylib" isn't found, the postmaster refuses to start. In my testing with PL/R, it reduces the first call to a PL/R function (after connecting) from almost 2 seconds, down to about 8 ms. Joe Conway
2003-03-20 05:51:44 +01:00
},
Locale support is on by default. The choice of locale is done in initdb and/or with GUC variables.
2002-04-03 07:39:33 +02:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"search_path", PGC_USERSET, CLIENT_CONN_STATEMENT,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Sets the schema search order for names that are not schema-qualified."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL,
GUC_LIST_INPUT | GUC_LIST_QUOTE
},
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
&namespace_search_path,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
"\"$user\",public",
check_search_path, assign_search_path, NULL
Locale support is on by default. The choice of locale is done in initdb and/or with GUC variables.
2002-04-03 07:39:33 +02:00
},
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
/* Can't be set in postgresql.conf */
{"server_encoding", PGC_INTERNAL, CLIENT_CONN_LOCALE,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Sets the server (database) character set encoding."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL,
Fix bug that allowed any logged-in user to SET ROLE to any other database user id (CVE-2006-0553). Also fix related bug in SET SESSION AUTHORIZATION that allows unprivileged users to crash the server, if it has been compiled with Asserts enabled. The escalation-of-privilege risk exists only in 8.1.0-8.1.2. However, the Assert-crash risk exists in all releases back to 7.3. Thanks to Akio Ishida for reporting this problem.
2006-02-12 23:32:43 +01:00
GUC_IS_NAME | GUC_REPORT | GUC_NOT_IN_SAMPLE | GUC_DISALLOW_IN_FILE
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
In the continuing saga of FE/BE protocol revisions, add reporting of initial values and runtime changes in selected parameters. This gets rid of the need for an initial 'select pg_client_encoding()' query in libpq, bringing us back to one message transmitted in each direction for a standard connection startup. To allow server version to be sent using the same GUC mechanism that handles other parameters, invent the concept of a never-settable GUC parameter: you can 'show server_version' but it's not settable by any GUC input source. Create 'lc_collate' and 'lc_ctype' never-settable parameters so that people can find out these settings without need for pg_controldata. (These side ideas were all discussed some time ago in pgsql-hackers, but not yet implemented.)
2003-04-25 21:45:10 +02:00
&server_encoding_string,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
"SQL_ASCII",
NULL, NULL, NULL
In the continuing saga of FE/BE protocol revisions, add reporting of initial values and runtime changes in selected parameters. This gets rid of the need for an initial 'select pg_client_encoding()' query in libpq, bringing us back to one message transmitted in each direction for a standard connection startup. To allow server version to be sent using the same GUC mechanism that handles other parameters, invent the concept of a never-settable GUC parameter: you can 'show server_version' but it's not settable by any GUC input source. Create 'lc_collate' and 'lc_ctype' never-settable parameters so that people can find out these settings without need for pg_controldata. (These side ideas were all discussed some time ago in pgsql-hackers, but not yet implemented.)
2003-04-25 21:45:10 +02:00
},
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
/* Can't be set in postgresql.conf */
Whack some sense into the configuration-file-location patch. Refactor code into something reasonably understandable, cause use of the feature to not fail in standalone backends or in EXEC_BACKEND case, fix sloppy guc.c table entries, make the documentation minimally usable.
2004-10-08 03:36:36 +02:00
{"server_version", PGC_INTERNAL, PRESET_OPTIONS,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Shows the server version."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL,
GUC_REPORT | GUC_NOT_IN_SAMPLE | GUC_DISALLOW_IN_FILE
},
In the continuing saga of FE/BE protocol revisions, add reporting of initial values and runtime changes in selected parameters. This gets rid of the need for an initial 'select pg_client_encoding()' query in libpq, bringing us back to one message transmitted in each direction for a standard connection startup. To allow server version to be sent using the same GUC mechanism that handles other parameters, invent the concept of a never-settable GUC parameter: you can 'show server_version' but it's not settable by any GUC input source. Create 'lc_collate' and 'lc_ctype' never-settable parameters so that people can find out these settings without need for pg_controldata. (These side ideas were all discussed some time ago in pgsql-hackers, but not yet implemented.)
2003-04-25 21:45:10 +02:00
&server_version_string,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
PG_VERSION,
NULL, NULL, NULL
Locale support is on by default. The choice of locale is done in initdb and/or with GUC variables.
2002-04-03 07:39:33 +02:00
},
Add SET ROLE. This is a partial commit of Stephen Frost's recent patch; I'm still working on the has_role function and information_schema changes.
2005-07-26 00:12:34 +02:00
{
/* Not for general use --- used by SET ROLE */
{"role", PGC_USERSET, UNGROUPED,
gettext_noop("Sets the current role."),
NULL,
Prevent indirect security attacks via changing session-local state within an allegedly immutable index function. It was previously recognized that we had to prevent such a function from executing SET/RESET ROLE/SESSION AUTHORIZATION, or it could trivially obtain the privileges of the session user. However, since there is in general no privilege checking for changes of session-local state, it is also possible for such a function to change settings in a way that might subvert later operations in the same session. Examples include changing search_path to cause an unexpected function to be called, or replacing an existing prepared statement with another one that will execute a function of the attacker's choosing. The present patch secures VACUUM, ANALYZE, and CREATE INDEX/REINDEX against these threats, which are the same places previously deemed to need protection against the SET ROLE issue. GUC changes are still allowed, since there are many useful cases for that, but we prevent security problems by forcing a rollback of any GUC change after completing the operation. Other cases are handled by throwing an error if any change is attempted; these include temp table creation, closing a cursor, and creating or deleting a prepared statement. (In 7.4, the infrastructure to roll back GUC changes doesn't exist, so we settle for rejecting changes of "search_path" in these contexts.) Original report and patch by Gurjeet Singh, additional analysis by Tom Lane. Security: CVE-2009-4136
2009-12-09 22:57:51 +01:00
GUC_IS_NAME | GUC_NO_SHOW_ALL | GUC_NO_RESET_ALL | GUC_NOT_IN_SAMPLE | GUC_DISALLOW_IN_FILE | GUC_NOT_WHILE_SEC_REST
Add SET ROLE. This is a partial commit of Stephen Frost's recent patch; I'm still working on the has_role function and information_schema changes.
2005-07-26 00:12:34 +02:00
},
&role_string,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
"none",
check_role, assign_role, show_role
Add SET ROLE. This is a partial commit of Stephen Frost's recent patch; I'm still working on the has_role function and information_schema changes.
2005-07-26 00:12:34 +02:00
},
Commit to match discussed elog() changes. Only update is that LOG is now just below FATAL in server_min_messages. Added more text to highlight ordering difference between it and client_min_messages. --------------------------------------------------------------------------- REALLYFATAL => PANIC STOP => PANIC New INFO level the prints to client by default New LOG level the prints to server log by default Cause VACUUM information to print only to the client NOTICE => INFO where purely information messages are sent DEBUG => LOG for purely server status messages DEBUG removed, kept as backward compatible DEBUG5, DEBUG4, DEBUG3, DEBUG2, DEBUG1 added DebugLvl removed in favor of new DEBUG[1-5] symbols New server_min_messages GUC parameter with values: DEBUG[5-1], INFO, NOTICE, ERROR, LOG, FATAL, PANIC New client_min_messages GUC parameter with values: DEBUG[5-1], LOG, INFO, NOTICE, ERROR, FATAL, PANIC Server startup now logged with LOG instead of DEBUG Remove debug_level GUC parameter elog() numbers now start at 10 Add test to print error message if older elog() values are passed to elog() Bootstrap mode now has a -d that requires an argument, like postmaster
2002-03-02 22:39:36 +01:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
/* Not for general use --- used by SET SESSION AUTHORIZATION */
{"session_authorization", PGC_USERSET, UNGROUPED,
Seems there are three GUC variables that are defined as "Shows ..." while you can actually set them with SET. This applied patch changes the wording from "Show" to "Set".
2003-12-01 04:55:21 +01:00
gettext_noop("Sets the session user name."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL,
Prevent indirect security attacks via changing session-local state within an allegedly immutable index function. It was previously recognized that we had to prevent such a function from executing SET/RESET ROLE/SESSION AUTHORIZATION, or it could trivially obtain the privileges of the session user. However, since there is in general no privilege checking for changes of session-local state, it is also possible for such a function to change settings in a way that might subvert later operations in the same session. Examples include changing search_path to cause an unexpected function to be called, or replacing an existing prepared statement with another one that will execute a function of the attacker's choosing. The present patch secures VACUUM, ANALYZE, and CREATE INDEX/REINDEX against these threats, which are the same places previously deemed to need protection against the SET ROLE issue. GUC changes are still allowed, since there are many useful cases for that, but we prevent security problems by forcing a rollback of any GUC change after completing the operation. Other cases are handled by throwing an error if any change is attempted; these include temp table creation, closing a cursor, and creating or deleting a prepared statement. (In 7.4, the infrastructure to roll back GUC changes doesn't exist, so we settle for rejecting changes of "search_path" in these contexts.) Original report and patch by Gurjeet Singh, additional analysis by Tom Lane. Security: CVE-2009-4136
2009-12-09 22:57:51 +01:00
GUC_IS_NAME | GUC_REPORT | GUC_NO_SHOW_ALL | GUC_NO_RESET_ALL | GUC_NOT_IN_SAMPLE | GUC_DISALLOW_IN_FILE | GUC_NOT_WHILE_SEC_REST
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
&session_authorization_string,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
NULL,
check_session_authorization, assign_session_authorization, NULL
Commit to match discussed elog() changes. Only update is that LOG is now just below FATAL in server_min_messages. Added more text to highlight ordering difference between it and client_min_messages. --------------------------------------------------------------------------- REALLYFATAL => PANIC STOP => PANIC New INFO level the prints to client by default New LOG level the prints to server log by default Cause VACUUM information to print only to the client NOTICE => INFO where purely information messages are sent DEBUG => LOG for purely server status messages DEBUG removed, kept as backward compatible DEBUG5, DEBUG4, DEBUG3, DEBUG2, DEBUG1 added DebugLvl removed in favor of new DEBUG[1-5] symbols New server_min_messages GUC parameter with values: DEBUG[5-1], INFO, NOTICE, ERROR, LOG, FATAL, PANIC New client_min_messages GUC parameter with values: DEBUG[5-1], LOG, INFO, NOTICE, ERROR, FATAL, PANIC Server startup now logged with LOG instead of DEBUG Remove debug_level GUC parameter elog() numbers now start at 10 Add test to print error message if older elog() values are passed to elog() Bootstrap mode now has a -d that requires an argument, like postmaster
2002-03-02 22:39:36 +01:00
},
Remove 'syslog' GUC variable, and add more logical 'log_destination' variable to control logoutput location on Unix and Win32. Magnus Hagander
2004-04-05 05:02:11 +02:00
{
log_destination should only be PGC_SIGHUP, not PGC_POSTMASTER, as per discussion yesterday. Also a few improvements in the associated documentation.
2004-07-24 21:51:23 +02:00
{"log_destination", PGC_SIGHUP, LOGGING_WHERE,
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
gettext_noop("Sets the destination for server log output."),
Tsearch2 functionality migrates to core. The bulk of this work is by Oleg Bartunov and Teodor Sigaev, but I did a lot of editorializing, so anything that's broken is probably my fault. Documentation is nonexistent as yet, but let's land the patch so we can get some portability testing done.
2007-08-21 03:11:32 +02:00
gettext_noop("Valid values are combinations of \"stderr\", "
"\"syslog\", \"csvlog\", and \"eventlog\", "
"depending on the platform."),
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
GUC_LIST_INPUT
Remove 'syslog' GUC variable, and add more logical 'log_destination' variable to control logoutput location on Unix and Win32. Magnus Hagander
2004-04-05 05:02:11 +02:00
},
&log_destination_string,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
"stderr",
check_log_destination, assign_log_destination, NULL
Remove 'syslog' GUC variable, and add more logical 'log_destination' variable to control logoutput location on Unix and Win32. Magnus Hagander
2004-04-05 05:02:11 +02:00
},
Create a built-in log rotation program, so that we no longer have to recommend that people go get Apache's rotatelogs program. Additional benefits are that configuration is done through GUC, rather than externally, and that the postmaster can monitor the log rotator and restart it after failure (though we certainly hope that won't happen often). Andreas Pflug, some rework by Tom Lane.
2004-08-06 01:32:13 +02:00
{
{"log_directory", PGC_SIGHUP, LOGGING_WHERE,
Replace log_filename_prefix with more general log_filename parameter, to allow DBA to choose the form in which log filenames reflect the current time. Also allow for truncating instead of appending to pre-existing files --- this is convenient when the log filename pattern rewrites the same names cyclically. Per Ed L.
2004-08-31 06:53:44 +02:00
gettext_noop("Sets the destination directory for log files."),
Wording cleanup for error messages. Also change can't -> cannot. Standard English uses "may", "can", and "might" in different ways: may - permission, "You may borrow my rake." can - ability, "I can lift that log." might - possibility, "It might rain today." Unfortunately, in conversational English, their use is often mixed, as in, "You may use this variable to do X", when in fact, "can" is a better choice. Similarly, "It may crash" is better stated, "It might crash".
2007-02-01 20:10:30 +01:00
gettext_noop("Can be specified as relative to the data directory "
Add a GUC_SUPERUSER_ONLY flag to mark GUC variables that should not be examinable by non-superusers, and use it to protect the recently-added GUC variables for data directory and config files. For now I have only flagged those variables that could be used to deduce something about the server's filesystem layout, but possibly we should also mark vars related to logging settings and other admin-only information?
2004-10-22 21:48:19 +02:00
"or as absolute path."),
GUC_SUPERUSER_ONLY
Create a built-in log rotation program, so that we no longer have to recommend that people go get Apache's rotatelogs program. Additional benefits are that configuration is done through GUC, rather than externally, and that the postmaster can monitor the log rotator and restart it after failure (though we certainly hope that won't happen often). Andreas Pflug, some rework by Tom Lane.
2004-08-06 01:32:13 +02:00
},
&Log_directory,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
"pg_log",
check_canonical_path, NULL, NULL
Create a built-in log rotation program, so that we no longer have to recommend that people go get Apache's rotatelogs program. Additional benefits are that configuration is done through GUC, rather than externally, and that the postmaster can monitor the log rotator and restart it after failure (though we certainly hope that won't happen often). Andreas Pflug, some rework by Tom Lane.
2004-08-06 01:32:13 +02:00
},
{
Replace log_filename_prefix with more general log_filename parameter, to allow DBA to choose the form in which log filenames reflect the current time. Also allow for truncating instead of appending to pre-existing files --- this is convenient when the log filename pattern rewrites the same names cyclically. Per Ed L.
2004-08-31 06:53:44 +02:00
{"log_filename", PGC_SIGHUP, LOGGING_WHERE,
gettext_noop("Sets the file name pattern for log files."),
Add a GUC_SUPERUSER_ONLY flag to mark GUC variables that should not be examinable by non-superusers, and use it to protect the recently-added GUC variables for data directory and config files. For now I have only flagged those variables that could be used to deduce something about the server's filesystem layout, but possibly we should also mark vars related to logging settings and other admin-only information?
2004-10-22 21:48:19 +02:00
NULL,
GUC_SUPERUSER_ONLY
Create a built-in log rotation program, so that we no longer have to recommend that people go get Apache's rotatelogs program. Additional benefits are that configuration is done through GUC, rather than externally, and that the postmaster can monitor the log rotator and restart it after failure (though we certainly hope that won't happen often). Andreas Pflug, some rework by Tom Lane.
2004-08-06 01:32:13 +02:00
},
Replace log_filename_prefix with more general log_filename parameter, to allow DBA to choose the form in which log filenames reflect the current time. Also allow for truncating instead of appending to pre-existing files --- this is convenient when the log filename pattern rewrites the same names cyclically. Per Ed L.
2004-08-31 06:53:44 +02:00
&Log_filename,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
"postgresql-%Y-%m-%d_%H%M%S.log",
NULL, NULL, NULL
Create a built-in log rotation program, so that we no longer have to recommend that people go get Apache's rotatelogs program. Additional benefits are that configuration is done through GUC, rather than externally, and that the postmaster can monitor the log rotator and restart it after failure (though we certainly hope that won't happen often). Andreas Pflug, some rework by Tom Lane.
2004-08-06 01:32:13 +02:00
},
Remove 'syslog' GUC variable, and add more logical 'log_destination' variable to control logoutput location on Unix and Win32. Magnus Hagander
2004-04-05 05:02:11 +02:00
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Fix syslog bug: if any messages are emitted to write_syslog before the facility has been set, the facility gets set to LOCAL0 and cannot be changed later. This seems reasonably plausible to happen, particularly at higher debug log levels, though I am not certain it explains Han Holl's recent report. Easiest fix is to teach the code how to change the value on-the-fly, which is nicer anyway. I made the settings PGC_SIGHUP to conform with log_destination.
2005-10-14 22:53:56 +02:00
{"syslog_ident", PGC_SIGHUP, LOGGING_WHERE,
gettext_noop("Sets the program name used to identify PostgreSQL "
"messages in syslog."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL
},
Fix syslog bug: if any messages are emitted to write_syslog before the facility has been set, the facility gets set to LOCAL0 and cannot be changed later. This seems reasonably plausible to happen, particularly at higher debug log levels, though I am not certain it explains Han Holl's recent report. Easiest fix is to teach the code how to change the value on-the-fly, which is nicer anyway. I made the settings PGC_SIGHUP to conform with log_destination.
2005-10-14 22:53:56 +02:00
&syslog_ident_str,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
"postgres",
NULL, assign_syslog_ident, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
Add runtime configuration options to control permission bits and group owner of unix socket.
2000-11-01 22:14:03 +01:00
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"TimeZone", PGC_USERSET, CLIENT_CONN_LOCALE,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Sets the time zone for displaying and interpreting time stamps."),
Mark the TimeZone parameter as GUC_REPORT, so that JDBC can find out when it changes. Per request from Kris Jurka.
2004-12-20 19:15:07 +01:00
NULL,
GUC_REPORT
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
&timezone_string,
Simplify handling of the timezone GUC by making initdb choose the default. We were doing some amazingly complicated things in order to avoid running the very expensive identify_system_timezone() procedure during GUC initialization. But there is an obvious fix for that, which is to do it once during initdb and have initdb install the system-specific default into postgresql.conf, as it already does for most other GUC variables that need system-environment-dependent defaults. This means that the timezone (and log_timezone) settings no longer have any magic behavior in the server. Per discussion.
2011-09-09 23:59:11 +02:00
"GMT",
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
check_timezone, assign_timezone, show_timezone
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
},
Remove hard-wired lists of timezone abbreviations in favor of providing configuration files that can be altered by a DBA. The australian_timezones GUC setting disappears, replaced by a timezone_abbreviations setting (set this to 'Australia' to get the effect of australian_timezones). The list of zone names defined by default has undergone a bit of cleanup, too. Documentation still needs some work --- in particular, should we fix Table B-4, or just get rid of it? Joachim Wieland, with some editorializing by moi.
2006-07-25 05:51:23 +02:00
{
{"timezone_abbreviations", PGC_USERSET, CLIENT_CONN_LOCALE,
Various improvements to the GUC description strings. Punctuate and capitalize the strings like sentences. Remove unnecessarily specific descriptions of the units used by GUC variables, since we now allow any reasonable unit to be specified.
2006-12-06 18:35:49 +01:00
gettext_noop("Selects a file of time zone abbreviations."),
Remove GUC extra_desc strings that are redundant with the enum value lists.
2008-07-01 23:07:33 +02:00
NULL
Remove hard-wired lists of timezone abbreviations in favor of providing configuration files that can be altered by a DBA. The australian_timezones GUC setting disappears, replaced by a timezone_abbreviations setting (set this to 'Australia' to get the effect of australian_timezones). The list of zone names defined by default has undergone a bit of cleanup, too. Documentation still needs some work --- in particular, should we fix Table B-4, or just get rid of it? Joachim Wieland, with some editorializing by moi.
2006-07-25 05:51:23 +02:00
},
Adjust initialization sequence for timezone_abbreviations so that it's handled just about like timezone; in particular, don't try to read anything during InitializeGUCOptions. Should solve current startup failure on Windows, and avoid wasted cycles if a nondefault setting is specified in postgresql.conf too. Possibly we need to think about a more general solution for handling 'expensive to set' GUC options.
2006-07-29 05:02:56 +02:00
&timezone_abbreviations_string,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
NULL,
check_timezone_abbreviations, assign_timezone_abbreviations, NULL
Remove hard-wired lists of timezone abbreviations in favor of providing configuration files that can be altered by a DBA. The australian_timezones GUC setting disappears, replaced by a timezone_abbreviations setting (set this to 'Australia' to get the effect of australian_timezones). The list of zone names defined by default has undergone a bit of cleanup, too. Documentation still needs some work --- in particular, should we fix Table B-4, or just get rid of it? Joachim Wieland, with some editorializing by moi.
2006-07-25 05:51:23 +02:00
},
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"transaction_isolation", PGC_USERSET, CLIENT_CONN_STATEMENT,
Seems there are three GUC variables that are defined as "Shows ..." while you can actually set them with SET. This applied patch changes the wording from "Show" to "Set".
2003-12-01 04:55:21 +01:00
gettext_noop("Sets the current transaction's isolation level."),
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
NULL,
GUC_NO_RESET_ALL | GUC_NOT_IN_SAMPLE | GUC_DISALLOW_IN_FILE
},
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
&XactIsoLevel_string,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
"default",
check_XactIsoLevel, assign_XactIsoLevel, show_XactIsoLevel
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
},
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"unix_socket_group", PGC_POSTMASTER, CONN_AUTH_SETTINGS,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Sets the owning group of the Unix-domain socket."),
Various improvements to the GUC description strings. Punctuate and capitalize the strings like sentences. Remove unnecessarily specific descriptions of the units used by GUC variables, since we now allow any reasonable unit to be specified.
2006-12-06 18:35:49 +01:00
gettext_noop("The owning user of the socket is always the user "
"that starts the server.")
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
&Unix_socket_group,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
"",
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
Support syncing WAL log to disk using either fsync(), fdatasync(), O_SYNC, or O_DSYNC (as available on a given platform). Add GUC parameter to control sync method. Also, add defense to XLogWrite to prevent it from going nuts if passed a target write position that's past the end of the buffers so far filled by XLogInsert.
2001-03-16 06:44:33 +01:00
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
{"unix_socket_directory", PGC_POSTMASTER, CONN_AUTH_SETTINGS,
Cleanup on --help-config: Now called --describe-config, no further options, machine readable, without headers, not sorted. Parameter descriptions adjusted to fit first sentence + rest convention.
2003-10-19 00:59:09 +02:00
gettext_noop("Sets the directory where the Unix-domain socket will be created."),
Add a GUC_SUPERUSER_ONLY flag to mark GUC variables that should not be examinable by non-superusers, and use it to protect the recently-added GUC variables for data directory and config files. For now I have only flagged those variables that could be used to deduce something about the server's filesystem layout, but possibly we should also mark vars related to logging settings and other admin-only information?
2004-10-22 21:48:19 +02:00
NULL,
GUC_SUPERUSER_ONLY
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
&UnixSocketDir,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
"",
check_canonical_path, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
UUNET is looking into offering PostgreSQL as a part of a managed web hosting product, on both shared and dedicated machines. We currently offer Oracle and MySQL, and it would be a nice middle-ground. However, as shipped, PostgreSQL lacks the following features we need that MySQL has: 1. The ability to listen only on a particular IP address. Each hosting customer has their own IP address, on which all of their servers (http, ftp, real media, etc.) run. 2. The ability to place the Unix-domain socket in a mode 700 directory. This allows us to automatically create an empty database, with an empty DBA password, for new or upgrading customers without having to interactively set a DBA password and communicate it to (or from) the customer. This in turn cuts down our install and upgrade times. 3. The ability to connect to the Unix-domain socket from within a change-rooted environment. We run CGI programs chrooted to the user's home directory, which is another reason why we need to be able to specify where the Unix-domain socket is, instead of /tmp. 4. The ability to, if run as root, open a pid file in /var/run as root, and then setuid to the desired user. (mysqld -u can almost do this; I had to patch it, too). The patch below fixes problem 1-3. I plan to address #4, also, but haven't done so yet. These diffs are big enough that they should give the PG development team something to think about in the meantime :-) Also, I'm about to leave for 2 weeks' vacation, so I thought I'd get out what I have, which works (for the problems it tackles), now. With these changes, we can set up and run PostgreSQL with scripts the same way we can with apache or proftpd or mysql. In summary, this patch makes the following enhancements: 1. Adds an environment variable PGUNIXSOCKET, analogous to MYSQL_UNIX_PORT, and command line options -k --unix-socket to the relevant programs. 2. Adds a -h option to postmaster to set the hostname or IP address to listen on instead of the default INADDR_ANY. 3. Extends some library interfaces to support the above. 4. Fixes a few memory leaks in PQconnectdb(). The default behavior is unchanged from stock 7.0.2; if you don't use any of these new features, they don't change the operation. David J. MacKenzie
2000-11-13 16:18:15 +01:00
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Replace the virtual_host and tcpip_socket parameters with a unified listen_addresses parameter, as per recent discussion. The default behavior is now to listen on localhost, which eliminates the need for the -i postmaster switch in many scenarios. Andrew Dunstan
2004-03-23 02:23:48 +01:00
{"listen_addresses", PGC_POSTMASTER, CONN_AUTH_SETTINGS,
Make listen_addresses be a comma-separated list instead of a space-separated list. More consistent with our other list-containing GUC variables.
2004-08-08 22:17:36 +02:00
gettext_noop("Sets the host name or IP address(es) to listen to."),
NULL,
GUC_LIST_INPUT
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
},
Replace the virtual_host and tcpip_socket parameters with a unified listen_addresses parameter, as per recent discussion. The default behavior is now to listen on localhost, which eliminates the need for the -i postmaster switch in many scenarios. Andrew Dunstan
2004-03-23 02:23:48 +01:00
&ListenAddresses,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
"localhost",
NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
},
UUNET is looking into offering PostgreSQL as a part of a managed web hosting product, on both shared and dedicated machines. We currently offer Oracle and MySQL, and it would be a nice middle-ground. However, as shipped, PostgreSQL lacks the following features we need that MySQL has: 1. The ability to listen only on a particular IP address. Each hosting customer has their own IP address, on which all of their servers (http, ftp, real media, etc.) run. 2. The ability to place the Unix-domain socket in a mode 700 directory. This allows us to automatically create an empty database, with an empty DBA password, for new or upgrading customers without having to interactively set a DBA password and communicate it to (or from) the customer. This in turn cuts down our install and upgrade times. 3. The ability to connect to the Unix-domain socket from within a change-rooted environment. We run CGI programs chrooted to the user's home directory, which is another reason why we need to be able to specify where the Unix-domain socket is, instead of /tmp. 4. The ability to, if run as root, open a pid file in /var/run as root, and then setuid to the desired user. (mysqld -u can almost do this; I had to patch it, too). The patch below fixes problem 1-3. I plan to address #4, also, but haven't done so yet. These diffs are big enough that they should give the PG development team something to think about in the meantime :-) Also, I'm about to leave for 2 weeks' vacation, so I thought I'd get out what I have, which works (for the problems it tackles), now. With these changes, we can set up and run PostgreSQL with scripts the same way we can with apache or proftpd or mysql. In summary, this patch makes the following enhancements: 1. Adds an environment variable PGUNIXSOCKET, analogous to MYSQL_UNIX_PORT, and command line options -k --unix-socket to the relevant programs. 2. Adds a -h option to postmaster to set the hostname or IP address to listen on instead of the default INADDR_ANY. 3. Extends some library interfaces to support the above. 4. Fixes a few memory leaks in PQconnectdb(). The default behavior is unchanged from stock 7.0.2; if you don't use any of these new features, they don't change the operation. David J. MacKenzie
2000-11-13 16:18:15 +01:00
Allow configuration files to be placed outside the data directory. Add new postgresql.conf variables to point to data, pg_hba.conf, and pg_ident.conf files. Needs more documentation.
2004-07-11 02:18:45 +02:00
{
Adjust configuration-files GUC behavior as per my recent proposal. The vars are renamed to data_directory, config_file, hba_file, and ident_file, and are guaranteed to be set to accurate absolute paths during postmaster startup. This commit does not yet do anything about hiding path values from non-superusers.
2004-10-10 01:13:22 +02:00
{"data_directory", PGC_POSTMASTER, FILE_LOCATIONS,
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
gettext_noop("Sets the server's data directory."),
NULL,
GUC_SUPERUSER_ONLY
Whack some sense into the configuration-file-location patch. Refactor code into something reasonably understandable, cause use of the feature to not fail in standalone backends or in EXEC_BACKEND case, fix sloppy guc.c table entries, make the documentation minimally usable.
2004-10-08 03:36:36 +02:00
},
Adjust configuration-files GUC behavior as per my recent proposal. The vars are renamed to data_directory, config_file, hba_file, and ident_file, and are guaranteed to be set to accurate absolute paths during postmaster startup. This commit does not yet do anything about hiding path values from non-superusers.
2004-10-10 01:13:22 +02:00
&data_directory,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
NULL,
Adjust configuration-files GUC behavior as per my recent proposal. The vars are renamed to data_directory, config_file, hba_file, and ident_file, and are guaranteed to be set to accurate absolute paths during postmaster startup. This commit does not yet do anything about hiding path values from non-superusers.
2004-10-10 01:13:22 +02:00
NULL, NULL, NULL
Allow configuration files to be placed outside the data directory. Add new postgresql.conf variables to point to data, pg_hba.conf, and pg_ident.conf files. Needs more documentation.
2004-07-11 02:18:45 +02:00
},
{
Adjust configuration-files GUC behavior as per my recent proposal. The vars are renamed to data_directory, config_file, hba_file, and ident_file, and are guaranteed to be set to accurate absolute paths during postmaster startup. This commit does not yet do anything about hiding path values from non-superusers.
2004-10-10 01:13:22 +02:00
{"config_file", PGC_POSTMASTER, FILE_LOCATIONS,
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
gettext_noop("Sets the server's main configuration file."),
NULL,
GUC_DISALLOW_IN_FILE | GUC_SUPERUSER_ONLY
Adjust configuration-files GUC behavior as per my recent proposal. The vars are renamed to data_directory, config_file, hba_file, and ident_file, and are guaranteed to be set to accurate absolute paths during postmaster startup. This commit does not yet do anything about hiding path values from non-superusers.
2004-10-10 01:13:22 +02:00
},
&ConfigFileName,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
NULL,
Adjust configuration-files GUC behavior as per my recent proposal. The vars are renamed to data_directory, config_file, hba_file, and ident_file, and are guaranteed to be set to accurate absolute paths during postmaster startup. This commit does not yet do anything about hiding path values from non-superusers.
2004-10-10 01:13:22 +02:00
NULL, NULL, NULL
},
{
{"hba_file", PGC_POSTMASTER, FILE_LOCATIONS,
Various improvements to the GUC description strings. Punctuate and capitalize the strings like sentences. Remove unnecessarily specific descriptions of the units used by GUC variables, since we now allow any reasonable unit to be specified.
2006-12-06 18:35:49 +01:00
gettext_noop("Sets the server's \"hba\" configuration file."),
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
NULL,
GUC_SUPERUSER_ONLY
Whack some sense into the configuration-file-location patch. Refactor code into something reasonably understandable, cause use of the feature to not fail in standalone backends or in EXEC_BACKEND case, fix sloppy guc.c table entries, make the documentation minimally usable.
2004-10-08 03:36:36 +02:00
},
Adjust configuration-files GUC behavior as per my recent proposal. The vars are renamed to data_directory, config_file, hba_file, and ident_file, and are guaranteed to be set to accurate absolute paths during postmaster startup. This commit does not yet do anything about hiding path values from non-superusers.
2004-10-10 01:13:22 +02:00
&HbaFileName,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
NULL,
Adjust configuration-files GUC behavior as per my recent proposal. The vars are renamed to data_directory, config_file, hba_file, and ident_file, and are guaranteed to be set to accurate absolute paths during postmaster startup. This commit does not yet do anything about hiding path values from non-superusers.
2004-10-10 01:13:22 +02:00
NULL, NULL, NULL
Allow configuration files to be placed outside the data directory. Add new postgresql.conf variables to point to data, pg_hba.conf, and pg_ident.conf files. Needs more documentation.
2004-07-11 02:18:45 +02:00
},
{
Adjust configuration-files GUC behavior as per my recent proposal. The vars are renamed to data_directory, config_file, hba_file, and ident_file, and are guaranteed to be set to accurate absolute paths during postmaster startup. This commit does not yet do anything about hiding path values from non-superusers.
2004-10-10 01:13:22 +02:00
{"ident_file", PGC_POSTMASTER, FILE_LOCATIONS,
Various improvements to the GUC description strings. Punctuate and capitalize the strings like sentences. Remove unnecessarily specific descriptions of the units used by GUC variables, since we now allow any reasonable unit to be specified.
2006-12-06 18:35:49 +01:00
gettext_noop("Sets the server's \"ident\" configuration file."),
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
NULL,
GUC_SUPERUSER_ONLY
Whack some sense into the configuration-file-location patch. Refactor code into something reasonably understandable, cause use of the feature to not fail in standalone backends or in EXEC_BACKEND case, fix sloppy guc.c table entries, make the documentation minimally usable.
2004-10-08 03:36:36 +02:00
},
Adjust configuration-files GUC behavior as per my recent proposal. The vars are renamed to data_directory, config_file, hba_file, and ident_file, and are guaranteed to be set to accurate absolute paths during postmaster startup. This commit does not yet do anything about hiding path values from non-superusers.
2004-10-10 01:13:22 +02:00
&IdentFileName,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
NULL,
Adjust configuration-files GUC behavior as per my recent proposal. The vars are renamed to data_directory, config_file, hba_file, and ident_file, and are guaranteed to be set to accurate absolute paths during postmaster startup. This commit does not yet do anything about hiding path values from non-superusers.
2004-10-10 01:13:22 +02:00
NULL, NULL, NULL
Allow configuration files to be placed outside the data directory. Add new postgresql.conf variables to point to data, pg_hba.conf, and pg_ident.conf files. Needs more documentation.
2004-07-11 02:18:45 +02:00
},
{
Adjust configuration-files GUC behavior as per my recent proposal. The vars are renamed to data_directory, config_file, hba_file, and ident_file, and are guaranteed to be set to accurate absolute paths during postmaster startup. This commit does not yet do anything about hiding path values from non-superusers.
2004-10-10 01:13:22 +02:00
{"external_pid_file", PGC_POSTMASTER, FILE_LOCATIONS,
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
gettext_noop("Writes the postmaster PID to the specified file."),
NULL,
GUC_SUPERUSER_ONLY
Whack some sense into the configuration-file-location patch. Refactor code into something reasonably understandable, cause use of the feature to not fail in standalone backends or in EXEC_BACKEND case, fix sloppy guc.c table entries, make the documentation minimally usable.
2004-10-08 03:36:36 +02:00
},
Adjust configuration-files GUC behavior as per my recent proposal. The vars are renamed to data_directory, config_file, hba_file, and ident_file, and are guaranteed to be set to accurate absolute paths during postmaster startup. This commit does not yet do anything about hiding path values from non-superusers.
2004-10-10 01:13:22 +02:00
&external_pid_file,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
NULL,
check_canonical_path, NULL, NULL
Allow configuration files to be placed outside the data directory. Add new postgresql.conf variables to point to data, pg_hba.conf, and pg_ident.conf files. Needs more documentation.
2004-07-11 02:18:45 +02:00
},
Make the temporary directory for pgstat files configurable by the GUC variable stats_temp_directory, instead of requiring the admin to mount/symlink the pg_stat_tmp directory manually. For now the config variable is PGC_POSTMASTER. Room for further improvment that would allow it to be changed on-the-fly.
2008-08-15 10:37:41 +02:00
{
Make stats_temp_directory PGC_SIGHUP, and document how it may cause a temporary "outage" of the statistics views. This requires making the stats collector respond to SIGHUP, like the other utility processes already did.
2008-08-25 17:11:01 +02:00
{"stats_temp_directory", PGC_SIGHUP, STATS_COLLECTOR,
Make the temporary directory for pgstat files configurable by the GUC variable stats_temp_directory, instead of requiring the admin to mount/symlink the pg_stat_tmp directory manually. For now the config variable is PGC_POSTMASTER. Room for further improvment that would allow it to be changed on-the-fly.
2008-08-15 10:37:41 +02:00
gettext_noop("Writes temporary statistics files to the specified directory."),
NULL,
GUC_SUPERUSER_ONLY
},
&pgstat_temp_directory,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
"pg_stat_tmp",
check_canonical_path, assign_pgstat_temp_directory, NULL
Make the temporary directory for pgstat files configurable by the GUC variable stats_temp_directory, instead of requiring the admin to mount/symlink the pg_stat_tmp directory manually. For now the config variable is PGC_POSTMASTER. Room for further improvment that would allow it to be changed on-the-fly.
2008-08-15 10:37:41 +02:00
},
Efficient transaction-controlled synchronous replication. If a standby is broadcasting reply messages and we have named one or more standbys in synchronous_standby_names then allow users who set synchronous_replication to wait for commit, which then provides strict data integrity guarantees. Design avoids sending and receiving transaction state information so minimises bookkeeping overheads. We synchronize with the highest priority standby that is connected and ready to synchronize. Other standbys can be defined to takeover in case of standby failure. This version has very strict behaviour; more relaxed options may be added at a later date. Simon Riggs and Fujii Masao, with reviews by Yeb Havinga, Jaime Casanova, Heikki Linnakangas and Robert Haas, plus the assistance of many other design reviewers.
2011-03-06 23:49:16 +01:00
{
Reclassify replication-related GUC variables as "master" and "standby". Per discussion, this structure seems more understandable than what was there before. Make config.sgml and postgresql.conf.sample agree. In passing do a bit of editorial work on the variable descriptions.
2011-07-07 21:10:32 +02:00
{"synchronous_standby_names", PGC_SIGHUP, REPLICATION_MASTER,
Message style tweaks
2011-07-04 23:01:35 +02:00
gettext_noop("List of names of potential synchronous standbys."),
Efficient transaction-controlled synchronous replication. If a standby is broadcasting reply messages and we have named one or more standbys in synchronous_standby_names then allow users who set synchronous_replication to wait for commit, which then provides strict data integrity guarantees. Design avoids sending and receiving transaction state information so minimises bookkeeping overheads. We synchronize with the highest priority standby that is connected and ready to synchronize. Other standbys can be defined to takeover in case of standby failure. This version has very strict behaviour; more relaxed options may be added at a later date. Simon Riggs and Fujii Masao, with reviews by Yeb Havinga, Jaime Casanova, Heikki Linnakangas and Robert Haas, plus the assistance of many other design reviewers.
2011-03-06 23:49:16 +01:00
NULL,
GUC_LIST_INPUT
},
&SyncRepStandbyNames,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
"",
check_synchronous_standby_names, NULL, NULL
Efficient transaction-controlled synchronous replication. If a standby is broadcasting reply messages and we have named one or more standbys in synchronous_standby_names then allow users who set synchronous_replication to wait for commit, which then provides strict data integrity guarantees. Design avoids sending and receiving transaction state information so minimises bookkeeping overheads. We synchronize with the highest priority standby that is connected and ready to synchronize. Other standbys can be defined to takeover in case of standby failure. This version has very strict behaviour; more relaxed options may be added at a later date. Simon Riggs and Fujii Masao, with reviews by Yeb Havinga, Jaime Casanova, Heikki Linnakangas and Robert Haas, plus the assistance of many other design reviewers.
2011-03-06 23:49:16 +01:00
},
Tsearch2 functionality migrates to core. The bulk of this work is by Oleg Bartunov and Teodor Sigaev, but I did a lot of editorializing, so anything that's broken is probably my fault. Documentation is nonexistent as yet, but let's land the patch so we can get some portability testing done.
2007-08-21 03:11:32 +02:00
{
{"default_text_search_config", PGC_USERSET, CLIENT_CONN_LOCALE,
gettext_noop("Sets default text search configuration."),
NULL
},
&TSCurrentConfig,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
"pg_catalog.simple",
check_TSCurrentConfig, assign_TSCurrentConfig, NULL
Tsearch2 functionality migrates to core. The bulk of this work is by Oleg Bartunov and Teodor Sigaev, but I did a lot of editorializing, so anything that's broken is probably my fault. Documentation is nonexistent as yet, but let's land the patch so we can get some portability testing done.
2007-08-21 03:11:32 +02:00
},
SSL improvements: o read global SSL configuration file o add GUC "ssl_ciphers" to control allowed ciphers o add libpq environment variable PGSSLKEY to control SSL hardware keys Victor B. Wagner
2007-02-16 03:59:41 +01:00
{
{"ssl_ciphers", PGC_POSTMASTER, CONN_AUTH_SECURITY,
gettext_noop("Sets the list of allowed SSL ciphers."),
NULL,
GUC_SUPERUSER_ONLY
},
&SSLCipherSuites,
Make GUC variables for syslog and SSL always visible Make the variables visible (but not used) even when support is not compiled in.
2010-12-18 16:53:59 +01:00
#ifdef USE_SSL
"ALL:!ADH:!LOW:!EXP:!MD5:@STRENGTH",
#else
"none",
#endif
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
NULL, NULL, NULL
SSL improvements: o read global SSL configuration file o add GUC "ssl_ciphers" to control allowed ciphers o add libpq environment variable PGSSLKEY to control SSL hardware keys Victor B. Wagner
2007-02-16 03:59:41 +01:00
},
Code review for SSLKEY patch.
2007-02-16 18:07:00 +01:00
Add support for an application_name parameter, which is displayed in pg_stat_activity and recorded in log entries. Dave Page, reviewed by Andres Freund
2009-11-29 00:38:08 +01:00
{
Add new GUC categories corresponding to sections in docs, and move description for vacuum_defer_cleanup_age to the correct category. Sections in postgresql.conf are also sorted in the same order with docs. Per gripe by Fujii Masao, suggestion by Heikki Linnakangas, and patch by me.
2010-06-15 09:52:11 +02:00
{"application_name", PGC_USERSET, LOGGING_WHAT,
pgindent run for 9.0
2010-02-26 03:01:40 +01:00
gettext_noop("Sets the application name to be reported in statistics and logs."),
NULL,
GUC_IS_NAME | GUC_REPORT | GUC_NOT_IN_SAMPLE
Add support for an application_name parameter, which is displayed in pg_stat_activity and recorded in log entries. Dave Page, reviewed by Andres Freund
2009-11-29 00:38:08 +01:00
},
&application_name,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
"",
check_application_name, assign_application_name, NULL
Add support for an application_name parameter, which is displayed in pg_stat_activity and recorded in log entries. Dave Page, reviewed by Andres Freund
2009-11-29 00:38:08 +01:00
},
Add add_missing_from GUC variable. Nigel J. Andrews
2003-06-12 00:13:22 +02:00
/* End-of-list marker */
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
{NULL, 0, 0, NULL, NULL}, NULL, NULL, NULL, NULL, NULL
pgindent run on all C files. Java run to follow. initdb/regression tests pass.
2001-10-25 07:50:21 +02:00
}
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
};
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
static struct config_enum ConfigureNamesEnum[] =
{
Convert backslash_quote guc to use enum.
2008-04-04 13:47:19 +02:00
{
{"backslash_quote", PGC_USERSET, COMPAT_OPTIONS_PREVIOUS,
gettext_noop("Sets whether \"\\'\" is allowed in string literals."),
Remove GUC extra_desc strings that are redundant with the enum value lists.
2008-07-01 23:07:33 +02:00
NULL
Convert backslash_quote guc to use enum.
2008-04-04 13:47:19 +02:00
},
&backslash_quote,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
BACKSLASH_QUOTE_SAFE_ENCODING, backslash_quote_options,
NULL, NULL, NULL
Convert backslash_quote guc to use enum.
2008-04-04 13:47:19 +02:00
},
Support hex-string input and output for type BYTEA. Both hex format and the traditional "escape" format are automatically handled on input. The output format is selected by the new GUC variable bytea_output. As committed, bytea_output defaults to HEX, which is an *incompatible change*. We will keep it this way for awhile for testing purposes, but should consider whether to switch to the more backwards-compatible default of ESCAPE before 8.5 is released. Peter Eisentraut
2009-08-04 18:08:37 +02:00
{
{"bytea_output", PGC_USERSET, CLIENT_CONN_STATEMENT,
gettext_noop("Sets the output format for bytea."),
NULL
},
&bytea_output,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
BYTEA_OUTPUT_HEX, bytea_output_options,
NULL, NULL, NULL
Support hex-string input and output for type BYTEA. Both hex format and the traditional "escape" format are automatically handled on input. The output format is selected by the new GUC variable bytea_output. As committed, bytea_output defaults to HEX, which is an *incompatible change*. We will keep it this way for awhile for testing purposes, but should consider whether to switch to the more backwards-compatible default of ESCAPE before 8.5 is released. Peter Eisentraut
2009-08-04 18:08:37 +02:00
},
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
{
{"client_min_messages", PGC_USERSET, LOGGING_WHEN,
gettext_noop("Sets the message levels that are sent to the client."),
Split apart message_level_options into one set for server-side settings and one for client-side, restoring the previous behaviour with different sort order for the 'log' level. Also, remove redundant list of available options, since the enum code will output it automatically.
2008-07-01 08:08:31 +02:00
gettext_noop("Each level includes all the levels that follow it. The later"
" the level, the fewer messages are sent.")
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
},
&client_min_messages,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
NOTICE, client_message_level_options,
NULL, NULL, NULL
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
},
Create a third option named "partition" for constraint_exclusion, and make it the default. This setting enables constraint exclusion checks only for appendrel members (ie, inheritance children and UNION ALL arms), which are the cases in which constraint exclusion is most likely to be useful. Avoiding the overhead for simple queries that are unlikely to benefit should bring the cost down to the point where this is a reasonable default setting. Per today's discussion.
2009-01-07 23:40:49 +01:00
{
{"constraint_exclusion", PGC_USERSET, QUERY_TUNING_OTHER,
gettext_noop("Enables the planner to use constraints to optimize queries."),
gettext_noop("Table scans will be skipped if their constraints"
" guarantee that no rows match the query.")
},
&constraint_exclusion,
CONSTRAINT_EXCLUSION_PARTITION, constraint_exclusion_options,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
NULL, NULL, NULL
Create a third option named "partition" for constraint_exclusion, and make it the default. This setting enables constraint exclusion checks only for appendrel members (ie, inheritance children and UNION ALL arms), which are the cases in which constraint exclusion is most likely to be useful. Avoiding the overhead for simple queries that are unlikely to benefit should bring the cost down to the point where this is a reasonable default setting. Per today's discussion.
2009-01-07 23:40:49 +01:00
},
Convert three more guc settings to enum type: default_transaction_isolation, session_replication_role and regex_flavor.
2008-04-02 16:42:56 +02:00
{
{"default_transaction_isolation", PGC_USERSET, CLIENT_CONN_STATEMENT,
gettext_noop("Sets the transaction isolation level of each new transaction."),
Remove GUC extra_desc strings that are redundant with the enum value lists.
2008-07-01 23:07:33 +02:00
NULL
Convert three more guc settings to enum type: default_transaction_isolation, session_replication_role and regex_flavor.
2008-04-02 16:42:56 +02:00
},
&DefaultXactIsoLevel,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
XACT_READ_COMMITTED, isolation_level_options,
NULL, NULL, NULL
Convert three more guc settings to enum type: default_transaction_isolation, session_replication_role and regex_flavor.
2008-04-02 16:42:56 +02:00
},
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
{
Add a new GUC variable called "IntervalStyle" that decouples interval output from DateStyle, and create a new interval style that produces output matching the SQL standard (at least for interval values that fall within the standard's restrictions). IntervalStyle is also used to resolve the conflict between the standard and traditional Postgres rules for interpreting negative interval input. Ron Mayer
2008-11-09 01:28:35 +01:00
{"IntervalStyle", PGC_USERSET, CLIENT_CONN_LOCALE,
gettext_noop("Sets the display format for interval values."),
NULL,
GUC_REPORT
},
&IntervalStyle,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
INTSTYLE_POSTGRES, intervalstyle_options,
NULL, NULL, NULL
Add a new GUC variable called "IntervalStyle" that decouples interval output from DateStyle, and create a new interval style that produces output matching the SQL standard (at least for interval values that fall within the standard's restrictions). IntervalStyle is also used to resolve the conflict between the standard and traditional Postgres rules for interpreting negative interval input. Ron Mayer
2008-11-09 01:28:35 +01:00
},
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
{
log_error_verbosity is now LOGGING_WHAT category.
2010-02-17 02:39:35 +01:00
{"log_error_verbosity", PGC_SUSET, LOGGING_WHAT,
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
gettext_noop("Sets the verbosity of logged messages."),
Remove GUC extra_desc strings that are redundant with the enum value lists.
2008-07-01 23:07:33 +02:00
NULL
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
},
&Log_error_verbosity,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
PGERROR_DEFAULT, log_error_verbosity_options,
NULL, NULL, NULL
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
},
{
{"log_min_messages", PGC_SUSET, LOGGING_WHEN,
gettext_noop("Sets the message levels that are logged."),
Remove GUC extra_desc strings that are redundant with the enum value lists.
2008-07-01 23:07:33 +02:00
gettext_noop("Each level includes all the levels that follow it. The later"
" the level, the fewer messages are sent.")
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
},
&log_min_messages,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
WARNING, server_message_level_options,
NULL, NULL, NULL
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
},
{
{"log_min_error_statement", PGC_SUSET, LOGGING_WHEN,
gettext_noop("Causes all statements generating error at or above this level to be logged."),
Remove GUC extra_desc strings that are redundant with the enum value lists.
2008-07-01 23:07:33 +02:00
gettext_noop("Each level includes all the levels that follow it. The later"
" the level, the fewer messages are sent.")
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
},
&log_min_error_statement,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
ERROR, server_message_level_options,
NULL, NULL, NULL
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
},
{
{"log_statement", PGC_SUSET, LOGGING_WHAT,
gettext_noop("Sets the type of statements logged."),
Remove GUC extra_desc strings that are redundant with the enum value lists.
2008-07-01 23:07:33 +02:00
NULL
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
},
&log_statement,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
LOGSTMT_NONE, log_statement_options,
NULL, NULL, NULL
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
},
Convert syslog_facility guc to enum type.
2008-04-03 11:21:15 +02:00
{
{"syslog_facility", PGC_SIGHUP, LOGGING_WHERE,
gettext_noop("Sets the syslog \"facility\" to be used when syslog enabled."),
Remove GUC extra_desc strings that are redundant with the enum value lists.
2008-07-01 23:07:33 +02:00
NULL
Convert syslog_facility guc to enum type.
2008-04-03 11:21:15 +02:00
},
&syslog_facility,
Make GUC variables for syslog and SSL always visible Make the variables visible (but not used) even when support is not compiled in.
2010-12-18 16:53:59 +01:00
#ifdef HAVE_SYSLOG
LOG_LOCAL0,
#else
0,
Convert syslog_facility guc to enum type.
2008-04-03 11:21:15 +02:00
#endif
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
syslog_facility_options,
NULL, assign_syslog_facility, NULL
Make GUC variables for syslog and SSL always visible Make the variables visible (but not used) even when support is not compiled in.
2010-12-18 16:53:59 +01:00
},
Convert syslog_facility guc to enum type.
2008-04-03 11:21:15 +02:00
Convert three more guc settings to enum type: default_transaction_isolation, session_replication_role and regex_flavor.
2008-04-02 16:42:56 +02:00
{
{"session_replication_role", PGC_SUSET, CLIENT_CONN_STATEMENT,
gettext_noop("Sets the session's behavior for triggers and rewrite rules."),
Remove GUC extra_desc strings that are redundant with the enum value lists.
2008-07-01 23:07:33 +02:00
NULL
Convert three more guc settings to enum type: default_transaction_isolation, session_replication_role and regex_flavor.
2008-04-02 16:42:56 +02:00
},
&SessionReplicationRole,
SESSION_REPLICATION_ROLE_ORIGIN, session_replication_role_options,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
NULL, assign_session_replication_role, NULL
Convert three more guc settings to enum type: default_transaction_isolation, session_replication_role and regex_flavor.
2008-04-02 16:42:56 +02:00
},
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
Merge synchronous_replication setting into synchronous_commit. This means one less thing to configure when setting up synchronous replication, and also avoids some ambiguity around what the behavior should be when the settings of these variables conflict. Fujii Masao, with additional hacking by me.
2011-04-04 22:13:01 +02:00
{
{"synchronous_commit", PGC_USERSET, WAL_SETTINGS,
gettext_noop("Sets the current transaction's synchronization level."),
NULL
},
&synchronous_commit,
SYNCHRONOUS_COMMIT_ON, synchronous_commit_options,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
NULL, NULL, NULL
Merge synchronous_replication setting into synchronous_commit. This means one less thing to configure when setting up synchronous replication, and also avoids some ambiguity around what the behavior should be when the settings of these variables conflict. Fujii Masao, with additional hacking by me.
2011-04-04 22:13:01 +02:00
},
Allow read only connections during recovery, known as Hot Standby. Enabled by recovery_connections = on (default) and forcing archive recovery using a recovery.conf. Recovery processing now emulates the original transactions as they are replayed, providing full locking and MVCC behaviour for read only queries. Recovery must enter consistent state before connections are allowed, so there is a delay, typically short, before connections succeed. Replay of recovering transactions can conflict and in some cases deadlock with queries during recovery; these result in query cancellation after max_standby_delay seconds have expired. Infrastructure changes have minor effects on normal running, though introduce four new types of WAL record. New test mode "make standbycheck" allows regression tests of static command behaviour on a standby server while in recovery. Typical and extreme dynamic behaviours have been checked via code inspection and manual testing. Few port specific behaviours have been utilised, though primary testing has been on Linux only so far. This commit is the basic patch. Additional changes will follow in this release to enhance some aspects of behaviour, notably improved handling of conflicts, deadlock detection and query cancellation. Changes to VACUUM FULL are also required. Simon Riggs, with significant and lengthy review by Heikki Linnakangas, including streamlined redesign of snapshot creation and two-phase commit. Important contributions from Florian Pflug, Mark Kirkwood, Merlin Moncure, Greg Stark, Gianni Ciolli, Gabriele Bartolini, Hannu Krosing, Robert Haas, Tatsuo Ishii, Hiroyuki Yamada plus support and feedback from many other community members.
2009-12-19 02:32:45 +01:00
{
Bring some sanity to the trace_recovery_messages code and docs. Per gripe from Fujii Masao, though this is not exactly his proposed patch. Categorize as DEVELOPER_OPTIONS and set context PGC_SIGHUP, as per Fujii, but set the default to LOG because higher values aren't really sensible (see the code for trace_recovery()). Fix the documentation to agree with the code and to try to explain what the variable actually does. Get rid of no-op calls trace_recovery(LOG), which accomplish nothing except to demonstrate that this option confuses even its author.
2010-08-20 00:55:01 +02:00
{"trace_recovery_messages", PGC_SIGHUP, DEVELOPER_OPTIONS,
gettext_noop("Enables logging of recovery-related debugging information."),
Allow read only connections during recovery, known as Hot Standby. Enabled by recovery_connections = on (default) and forcing archive recovery using a recovery.conf. Recovery processing now emulates the original transactions as they are replayed, providing full locking and MVCC behaviour for read only queries. Recovery must enter consistent state before connections are allowed, so there is a delay, typically short, before connections succeed. Replay of recovering transactions can conflict and in some cases deadlock with queries during recovery; these result in query cancellation after max_standby_delay seconds have expired. Infrastructure changes have minor effects on normal running, though introduce four new types of WAL record. New test mode "make standbycheck" allows regression tests of static command behaviour on a standby server while in recovery. Typical and extreme dynamic behaviours have been checked via code inspection and manual testing. Few port specific behaviours have been utilised, though primary testing has been on Linux only so far. This commit is the basic patch. Additional changes will follow in this release to enhance some aspects of behaviour, notably improved handling of conflicts, deadlock detection and query cancellation. Changes to VACUUM FULL are also required. Simon Riggs, with significant and lengthy review by Heikki Linnakangas, including streamlined redesign of snapshot creation and two-phase commit. Important contributions from Florian Pflug, Mark Kirkwood, Merlin Moncure, Greg Stark, Gianni Ciolli, Gabriele Bartolini, Hannu Krosing, Robert Haas, Tatsuo Ishii, Hiroyuki Yamada plus support and feedback from many other community members.
2009-12-19 02:32:45 +01:00
gettext_noop("Each level includes all the levels that follow it. The later"
" the level, the fewer messages are sent.")
},
&trace_recovery_messages,
pgindent run before PG 9.1 beta 1.
2011-04-10 17:42:00 +02:00
Bring some sanity to the trace_recovery_messages code and docs. Per gripe from Fujii Masao, though this is not exactly his proposed patch. Categorize as DEVELOPER_OPTIONS and set context PGC_SIGHUP, as per Fujii, but set the default to LOG because higher values aren't really sensible (see the code for trace_recovery()). Fix the documentation to agree with the code and to try to explain what the variable actually does. Get rid of no-op calls trace_recovery(LOG), which accomplish nothing except to demonstrate that this option confuses even its author.
2010-08-20 00:55:01 +02:00
/*
pgindent run before PG 9.1 beta 1.
2011-04-10 17:42:00 +02:00
* client_message_level_options allows too many values, really, but
* it's not worth having a separate options array for this.
Bring some sanity to the trace_recovery_messages code and docs. Per gripe from Fujii Masao, though this is not exactly his proposed patch. Categorize as DEVELOPER_OPTIONS and set context PGC_SIGHUP, as per Fujii, but set the default to LOG because higher values aren't really sensible (see the code for trace_recovery()). Fix the documentation to agree with the code and to try to explain what the variable actually does. Get rid of no-op calls trace_recovery(LOG), which accomplish nothing except to demonstrate that this option confuses even its author.
2010-08-20 00:55:01 +02:00
*/
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
LOG, client_message_level_options,
NULL, NULL, NULL
Allow read only connections during recovery, known as Hot Standby. Enabled by recovery_connections = on (default) and forcing archive recovery using a recovery.conf. Recovery processing now emulates the original transactions as they are replayed, providing full locking and MVCC behaviour for read only queries. Recovery must enter consistent state before connections are allowed, so there is a delay, typically short, before connections succeed. Replay of recovering transactions can conflict and in some cases deadlock with queries during recovery; these result in query cancellation after max_standby_delay seconds have expired. Infrastructure changes have minor effects on normal running, though introduce four new types of WAL record. New test mode "make standbycheck" allows regression tests of static command behaviour on a standby server while in recovery. Typical and extreme dynamic behaviours have been checked via code inspection and manual testing. Few port specific behaviours have been utilised, though primary testing has been on Linux only so far. This commit is the basic patch. Additional changes will follow in this release to enhance some aspects of behaviour, notably improved handling of conflicts, deadlock detection and query cancellation. Changes to VACUUM FULL are also required. Simon Riggs, with significant and lengthy review by Heikki Linnakangas, including streamlined redesign of snapshot creation and two-phase commit. Important contributions from Florian Pflug, Mark Kirkwood, Merlin Moncure, Greg Stark, Gianni Ciolli, Gabriele Bartolini, Hannu Krosing, Robert Haas, Tatsuo Ishii, Hiroyuki Yamada plus support and feedback from many other community members.
2009-12-19 02:32:45 +01:00
},
Add support for tracking call counts and elapsed runtime for user-defined functions. Note that because this patch changes FmgrInfo, any external C functions you might be testing with 8.4 will need to be recompiled. Patch by Martin Pihlak, some editorialization by me (principally, removing tracking of getrusage() numbers)
2008-05-15 02:17:41 +02:00
{
{"track_functions", PGC_SUSET, STATS_COLLECTOR,
gettext_noop("Collects function-level statistics on database activity."),
Remove GUC extra_desc strings that are redundant with the enum value lists.
2008-07-01 23:07:33 +02:00
NULL
Add support for tracking call counts and elapsed runtime for user-defined functions. Note that because this patch changes FmgrInfo, any external C functions you might be testing with 8.4 will need to be recompiled. Patch by Martin Pihlak, some editorialization by me (principally, removing tracking of getrusage() numbers)
2008-05-15 02:17:41 +02:00
},
&pgstat_track_functions,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
TRACK_FUNC_OFF, track_function_options,
NULL, NULL, NULL
Add support for tracking call counts and elapsed runtime for user-defined functions. Note that because this patch changes FmgrInfo, any external C functions you might be testing with 8.4 will need to be recompiled. Patch by Martin Pihlak, some editorialization by me (principally, removing tracking of getrusage() numbers)
2008-05-15 02:17:41 +02:00
},
Introduce wal_level GUC to explicitly control if information needed for archival or hot standby should be WAL-logged, instead of deducing that from other options like archive_mode. This replaces recovery_connections GUC in the primary, where it now has no effect, but it's still used in the standby to enable/disable hot standby. Remove the WAL-logging of "unlogged operations", like creating an index without WAL-logging and fsyncing it at the end. Instead, we keep a copy of the wal_mode setting and the settings that affect how much shared memory a hot standby server needs to track master transactions (max_connections, max_prepared_xacts, max_locks_per_xact) in pg_control. Whenever the settings change, at server restart, write a WAL record noting the new settings and update pg_control. This allows us to notice the change in those settings in the standby at the right moment, they used to be included in checkpoint records, but that meant that a changed value was not reflected in the standby until the first checkpoint after the change. Bump PG_CONTROL_VERSION and XLOG_PAGE_MAGIC. Whack XLOG_PAGE_MAGIC back to the sequence it used to follow, before hot standby and subsequent patches changed it to 0x9003.
2010-04-28 18:10:43 +02:00
{
{"wal_level", PGC_POSTMASTER, WAL_SETTINGS,
gettext_noop("Set the level of information written to the WAL."),
NULL
},
&wal_level,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
WAL_LEVEL_MINIMAL, wal_level_options,
NULL, NULL, NULL
Introduce wal_level GUC to explicitly control if information needed for archival or hot standby should be WAL-logged, instead of deducing that from other options like archive_mode. This replaces recovery_connections GUC in the primary, where it now has no effect, but it's still used in the standby to enable/disable hot standby. Remove the WAL-logging of "unlogged operations", like creating an index without WAL-logging and fsyncing it at the end. Instead, we keep a copy of the wal_mode setting and the settings that affect how much shared memory a hot standby server needs to track master transactions (max_connections, max_prepared_xacts, max_locks_per_xact) in pg_control. Whenever the settings change, at server restart, write a WAL record noting the new settings and update pg_control. This allows us to notice the change in those settings in the standby at the right moment, they used to be included in checkpoint records, but that meant that a changed value was not reflected in the standby until the first checkpoint after the change. Bump PG_CONTROL_VERSION and XLOG_PAGE_MAGIC. Whack XLOG_PAGE_MAGIC back to the sequence it used to follow, before hot standby and subsequent patches changed it to 0x9003.
2010-04-28 18:10:43 +02:00
},
Convert wal_sync_method to guc enum.
2008-05-12 10:35:05 +02:00
{
{"wal_sync_method", PGC_SIGHUP, WAL_SETTINGS,
gettext_noop("Selects the method used for forcing WAL updates to disk."),
NULL
},
&sync_method,
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
DEFAULT_SYNC_METHOD, sync_method_options,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
NULL, assign_xlog_sync_method, NULL
Convert wal_sync_method to guc enum.
2008-05-12 10:35:05 +02:00
},
Turn xmlbinary and xmloption GUC variables into enumsTurn xmlbinary and xmloption GUC variables into enums..
2008-04-04 10:33:15 +02:00
{
{"xmlbinary", PGC_USERSET, CLIENT_CONN_STATEMENT,
gettext_noop("Sets how binary values are to be encoded in XML."),
Remove GUC extra_desc strings that are redundant with the enum value lists.
2008-07-01 23:07:33 +02:00
NULL
Turn xmlbinary and xmloption GUC variables into enumsTurn xmlbinary and xmloption GUC variables into enums..
2008-04-04 10:33:15 +02:00
},
&xmlbinary,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
XMLBINARY_BASE64, xmlbinary_options,
NULL, NULL, NULL
Turn xmlbinary and xmloption GUC variables into enumsTurn xmlbinary and xmloption GUC variables into enums..
2008-04-04 10:33:15 +02:00
},
{
{"xmloption", PGC_USERSET, CLIENT_CONN_STATEMENT,
gettext_noop("Sets whether XML data in implicit parsing and serialization "
"operations is to be considered as documents or content fragments."),
Remove GUC extra_desc strings that are redundant with the enum value lists.
2008-07-01 23:07:33 +02:00
NULL
Turn xmlbinary and xmloption GUC variables into enumsTurn xmlbinary and xmloption GUC variables into enums..
2008-04-04 10:33:15 +02:00
},
&xmloption,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
XMLOPTION_CONTENT, xmloption_options,
NULL, NULL, NULL
Turn xmlbinary and xmloption GUC variables into enumsTurn xmlbinary and xmloption GUC variables into enums..
2008-04-04 10:33:15 +02:00
},
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
/* End-of-list marker */
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
{NULL, 0, 0, NULL, NULL}, NULL, 0, NULL, NULL, NULL, NULL
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
}
};
Allow configuration files to be placed outside the data directory. Add new postgresql.conf variables to point to data, pg_hba.conf, and pg_ident.conf files. Needs more documentation.
2004-07-11 02:18:45 +02:00
/******** end of options list ********/
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
Rename SortMem and VacuumMem to work_mem and maintenance_work_mem. Make btree index creation and initial validation of foreign-key constraints use maintenance_work_mem rather than work_mem as their memory limit. Add some code to guc.c to allow these variables to be referenced by their old names in SHOW and SET commands, for backwards compatibility.
2004-02-03 18:34:04 +01:00
/*
* To allow continued support of obsolete names for GUC variables, we apply
* the following mappings to any unrecognized name. Note that an old name
* should be mapped to a new one only if the new variable has very similar
* semantics to the old.
*/
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
static const char *const map_old_guc_names[] = {
Rename SortMem and VacuumMem to work_mem and maintenance_work_mem. Make btree index creation and initial validation of foreign-key constraints use maintenance_work_mem rather than work_mem as their memory limit. Add some code to guc.c to allow these variables to be referenced by their old names in SHOW and SET commands, for backwards compatibility.
2004-02-03 18:34:04 +01:00
"sort_mem", "work_mem",
"vacuum_mem", "maintenance_work_mem",
NULL
};
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
/*
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
* Actual lookup of variables is done through this single, sorted array.
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
*/
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
static struct config_generic **guc_variables;
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
/* Current number of variables contained in the vector */
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
static int num_guc_variables;
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
/* Vector capacity */
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
static int size_guc_variables;
Add all possible config file options.
2001-01-24 19:37:31 +01:00
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
static bool guc_dirty; /* TRUE if need to do commit/abort work */
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
In the continuing saga of FE/BE protocol revisions, add reporting of initial values and runtime changes in selected parameters. This gets rid of the need for an initial 'select pg_client_encoding()' query in libpq, bringing us back to one message transmitted in each direction for a standard connection startup. To allow server version to be sent using the same GUC mechanism that handles other parameters, invent the concept of a never-settable GUC parameter: you can 'show server_version' but it's not settable by any GUC input source. Create 'lc_collate' and 'lc_ctype' never-settable parameters so that people can find out these settings without need for pg_controldata. (These side ideas were all discussed some time ago in pgsql-hackers, but not yet implemented.)
2003-04-25 21:45:10 +02:00
static bool reporting_enabled; /* TRUE to enable GUC_REPORT */
Implement function-local GUC parameter settings, as per recent discussion. There are still some loose ends: I didn't do anything about the SET FROM CURRENT idea yet, and it's not real clear whether we are happy with the interaction of SET LOCAL with function-local settings. The documentation is a bit spartan, too.
2007-09-03 02:39:26 +02:00
static int GUCNestLevel = 0; /* 1 when in main transaction */
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
pgindent run.
2002-09-04 22:31:48 +02:00
static int guc_var_compare(const void *a, const void *b);
Rename SortMem and VacuumMem to work_mem and maintenance_work_mem. Make btree index creation and initial validation of foreign-key constraints use maintenance_work_mem rather than work_mem as their memory limit. Add some code to guc.c to allow these variables to be referenced by their old names in SHOW and SET commands, for backwards compatibility.
2004-02-03 18:34:04 +01:00
static int guc_name_compare(const char *namea, const char *nameb);
Split PGC_S_DEFAULT into two values, for true boot_val vs computed default. Failure to distinguish these cases is the real cause behind the recent reports of Windows builds crashing on 'infinity'::timestamp, which was directly due to failure to establish a value of timezone_abbreviations in postmaster child processes. The postmaster had the desired value, but write_one_nondefault_variable() didn't transmit it to backends. To fix that, invent a new value PGC_S_DYNAMIC_DEFAULT, and be sure to use that or PGC_S_ENV_VAR (as appropriate) for "default" settings that are computed during initialization. (We need both because there's at least one variable that could receive a value from either source.) This commit also fixes ProcessConfigFile's failure to restore the correct default value for certain GUC variables if they are set in postgresql.conf and then removed/commented out of the file. We have to recompute and reinstall the value for any GUC variable that could have received a value from PGC_S_DYNAMIC_DEFAULT or PGC_S_ENV_VAR sources, and there were a number of oversights. (That whole thing is a crock that needs to be redesigned, but not today.) However, I intentionally didn't make it work "exactly right" for the cases of timezone and log_timezone. The exactly right behavior would involve running select_default_timezone, which we'd have to do independently in each postgres process, causing the whole database to become entirely unresponsive for as much as several seconds. That didn't seem like a good idea, especially since the variable's removal from postgresql.conf might be just an accidental edit. Instead the behavior is to adopt the previously active setting as if it were default. Note that this patch creates an ABI break for extensions that use any of the PGC_S_XXX constants; they'll need to be recompiled.
2011-05-12 01:57:38 +02:00
static void InitializeGUCOptionsFromEnvironment(void);
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
static void InitializeOneGUCOption(struct config_generic * gconf);
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
static void push_old_value(struct config_generic * gconf, GucAction action);
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
static void ReportGUCOption(struct config_generic * record);
Add a GUC_SUPERUSER_ONLY flag to mark GUC variables that should not be examinable by non-superusers, and use it to protect the recently-added GUC variables for data directory and config files. For now I have only flagged those variables that could be used to deduce something about the server's filesystem layout, but possibly we should also mark vars related to logging settings and other admin-only information?
2004-10-22 21:48:19 +02:00
static void ShowGUCConfigOption(const char *name, DestReceiver *dest);
static void ShowAllGUCConfig(DestReceiver *dest);
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
static char *_ShowOption(struct config_generic * record, bool use_units);
Enforce superuser permissions checks during ALTER ROLE/DATABASE SET, rather than during define_custom_variable(). This entails rejecting an ALTER command if the target variable doesn't have a known (non-placeholder) definition, unless the calling user is superuser. When the variable *is* known, we can correctly apply the rule that only superusers can issue ALTER for SUSET parameters. This allows define_custom_variable to apply ALTER's values for SUSET parameters at module load time, secure in the knowledge that only a superuser could have set the ALTER value. This change fixes a longstanding gotcha in the usage of SUSET-level custom parameters; which is a good thing to fix now that plpgsql defines such a parameter.
2010-04-21 22:54:19 +02:00
static bool validate_option_array_item(const char *name, const char *value,
pgindent run for 9.0, second run
2010-07-06 21:19:02 +02:00
bool skipIfNoPermissions);
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
Fix unchecked mallocs/strdups added by recent placeholder-config-vars patch. Thomas Hallgren
2004-07-06 01:14:14 +02:00
/*
* Some infrastructure for checking malloc/strdup/realloc calls
*/
static void *
guc_malloc(int elevel, size_t size)
{
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
void *data;
Fix unchecked mallocs/strdups added by recent placeholder-config-vars patch. Thomas Hallgren
2004-07-06 01:14:14 +02:00
data = malloc(size);
if (data == NULL)
ereport(elevel,
(errcode(ERRCODE_OUT_OF_MEMORY),
errmsg("out of memory")));
return data;
}
static void *
guc_realloc(int elevel, void *old, size_t size)
{
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
void *data;
Fix unchecked mallocs/strdups added by recent placeholder-config-vars patch. Thomas Hallgren
2004-07-06 01:14:14 +02:00
data = realloc(old, size);
if (data == NULL)
ereport(elevel,
(errcode(ERRCODE_OUT_OF_MEMORY),
errmsg("out of memory")));
return data;
}
static char *
guc_strdup(int elevel, const char *src)
{
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
char *data;
Fix unchecked mallocs/strdups added by recent placeholder-config-vars patch. Thomas Hallgren
2004-07-06 01:14:14 +02:00
data = strdup(src);
if (data == NULL)
ereport(elevel,
(errcode(ERRCODE_OUT_OF_MEMORY),
errmsg("out of memory")));
return data;
}
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
/*
* Detect whether strval is referenced anywhere in a GUC string item
*/
static bool
string_field_used(struct config_string * conf, char *strval)
{
GucStack *stack;
if (strval == *(conf->variable) ||
strval == conf->reset_val ||
strval == conf->boot_val)
return true;
for (stack = conf->gen.stack; stack; stack = stack->prev)
{
if (strval == stack->prior.val.stringval ||
strval == stack->masked.val.stringval)
return true;
}
return false;
}
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
/*
* Support for assigning to a field of a string GUC item. Free the prior
* value if it's not referenced anywhere else in the item (including stacked
* states).
*/
static void
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
set_string_field(struct config_string * conf, char **field, char *newval)
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
{
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
char *oldval = *field;
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
/* Do the assignment */
*field = newval;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
/* Free old value if it's not NULL and isn't referenced anymore */
if (oldval && !string_field_used(conf, oldval))
free(oldval);
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
}
/*
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
* Detect whether an "extra" struct is referenced anywhere in a GUC item
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
*/
static bool
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
extra_field_used(struct config_generic * gconf, void *extra)
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
{
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
GucStack *stack;
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
if (extra == gconf->extra)
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
return true;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
switch (gconf->vartype)
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
case PGC_BOOL:
if (extra == ((struct config_bool *) gconf)->reset_extra)
return true;
break;
case PGC_INT:
if (extra == ((struct config_int *) gconf)->reset_extra)
return true;
break;
case PGC_REAL:
if (extra == ((struct config_real *) gconf)->reset_extra)
return true;
break;
case PGC_STRING:
if (extra == ((struct config_string *) gconf)->reset_extra)
return true;
break;
case PGC_ENUM:
if (extra == ((struct config_enum *) gconf)->reset_extra)
return true;
break;
}
for (stack = gconf->stack; stack; stack = stack->prev)
{
if (extra == stack->prior.extra ||
extra == stack->masked.extra)
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
return true;
}
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
return false;
}
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
/*
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
* Support for assigning to an "extra" field of a GUC item. Free the prior
* value if it's not referenced anywhere else in the item (including stacked
* states).
*/
static void
set_extra_field(struct config_generic * gconf, void **field, void *newval)
{
void *oldval = *field;
/* Do the assignment */
*field = newval;
/* Free old value if it's not NULL and isn't referenced anymore */
if (oldval && !extra_field_used(gconf, oldval))
free(oldval);
}
/*
* Support for copying a variable's active value into a stack entry.
* The "extra" field associated with the active value is copied, too.
*
* NB: be sure stringval and extra fields of a new stack entry are
* initialized to NULL before this is used, else we'll try to free() them.
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
*/
static void
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
set_stack_value(struct config_generic * gconf, config_var_value *val)
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
{
switch (gconf->vartype)
{
case PGC_BOOL:
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
val->val.boolval =
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
*((struct config_bool *) gconf)->variable;
break;
case PGC_INT:
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
val->val.intval =
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
*((struct config_int *) gconf)->variable;
break;
case PGC_REAL:
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
val->val.realval =
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
*((struct config_real *) gconf)->variable;
break;
case PGC_STRING:
set_string_field((struct config_string *) gconf,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
&(val->val.stringval),
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
*((struct config_string *) gconf)->variable);
break;
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
case PGC_ENUM:
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
val->val.enumval =
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
*((struct config_enum *) gconf)->variable;
break;
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
}
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
set_extra_field(gconf, &(val->extra), gconf->extra);
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
}
/*
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
* Support for discarding a no-longer-needed value in a stack entry.
* The "extra" field associated with the stack entry is cleared, too.
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
*/
static void
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
discard_stack_value(struct config_generic * gconf, config_var_value *val)
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
{
switch (gconf->vartype)
{
case PGC_BOOL:
case PGC_INT:
case PGC_REAL:
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
case PGC_ENUM:
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
/* no need to do anything */
break;
case PGC_STRING:
set_string_field((struct config_string *) gconf,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
&(val->val.stringval),
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
NULL);
break;
}
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
set_extra_field(gconf, &(val->extra), NULL);
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
}
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
/*
* Fetch the sorted array pointer (exported for help_config.c's use ONLY)
*/
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
struct config_generic **
get_guc_variables(void)
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
{
return guc_variables;
}
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
/*
pgindent run.
2002-09-04 22:31:48 +02:00
* Build the sorted array. This is split out so that it could be
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
* re-executed after startup (eg, we could allow loadable modules to
* add vars, and then we'd need to re-sort).
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
*/
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
void
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
build_guc_variables(void)
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
{
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
int size_vars;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
int num_vars = 0;
struct config_generic **guc_vars;
pgindent run. Make it all clean.
2001-03-22 05:01:46 +01:00
int i;
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
for (i = 0; ConfigureNamesBool[i].gen.name; i++)
Extend GUC concepts of parse_hook and assign_hook to all four supported datatypes, not only strings. parse_hook is useless for bool, I suppose, but it seems possibly useful for int and double to apply variable-specific constraints that are more complex than simple range limits. assign_hook is definitely useful for all datatypes --- we need it right now for bool to support date cache reset when changing Australian timezone rule setting. Also, clean up some residual problems with the reset all/show all patch, including memory leaks and mistaken reset of PostPortNumber. It seems best that RESET ALL not touch variables that don't have SUSET or USERSET context.
2001-06-13 00:54:06 +02:00
{
struct config_bool *conf = &ConfigureNamesBool[i];
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
/* Rather than requiring vartype to be filled in by hand, do this: */
conf->gen.vartype = PGC_BOOL;
num_vars++;
Extend GUC concepts of parse_hook and assign_hook to all four supported datatypes, not only strings. parse_hook is useless for bool, I suppose, but it seems possibly useful for int and double to apply variable-specific constraints that are more complex than simple range limits. assign_hook is definitely useful for all datatypes --- we need it right now for bool to support date cache reset when changing Australian timezone rule setting. Also, clean up some residual problems with the reset all/show all patch, including memory leaks and mistaken reset of PostPortNumber. It seems best that RESET ALL not touch variables that don't have SUSET or USERSET context.
2001-06-13 00:54:06 +02:00
}
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
for (i = 0; ConfigureNamesInt[i].gen.name; i++)
Extend GUC concepts of parse_hook and assign_hook to all four supported datatypes, not only strings. parse_hook is useless for bool, I suppose, but it seems possibly useful for int and double to apply variable-specific constraints that are more complex than simple range limits. assign_hook is definitely useful for all datatypes --- we need it right now for bool to support date cache reset when changing Australian timezone rule setting. Also, clean up some residual problems with the reset all/show all patch, including memory leaks and mistaken reset of PostPortNumber. It seems best that RESET ALL not touch variables that don't have SUSET or USERSET context.
2001-06-13 00:54:06 +02:00
{
struct config_int *conf = &ConfigureNamesInt[i];
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
conf->gen.vartype = PGC_INT;
num_vars++;
Extend GUC concepts of parse_hook and assign_hook to all four supported datatypes, not only strings. parse_hook is useless for bool, I suppose, but it seems possibly useful for int and double to apply variable-specific constraints that are more complex than simple range limits. assign_hook is definitely useful for all datatypes --- we need it right now for bool to support date cache reset when changing Australian timezone rule setting. Also, clean up some residual problems with the reset all/show all patch, including memory leaks and mistaken reset of PostPortNumber. It seems best that RESET ALL not touch variables that don't have SUSET or USERSET context.
2001-06-13 00:54:06 +02:00
}
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
for (i = 0; ConfigureNamesReal[i].gen.name; i++)
Extend GUC concepts of parse_hook and assign_hook to all four supported datatypes, not only strings. parse_hook is useless for bool, I suppose, but it seems possibly useful for int and double to apply variable-specific constraints that are more complex than simple range limits. assign_hook is definitely useful for all datatypes --- we need it right now for bool to support date cache reset when changing Australian timezone rule setting. Also, clean up some residual problems with the reset all/show all patch, including memory leaks and mistaken reset of PostPortNumber. It seems best that RESET ALL not touch variables that don't have SUSET or USERSET context.
2001-06-13 00:54:06 +02:00
{
struct config_real *conf = &ConfigureNamesReal[i];
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
conf->gen.vartype = PGC_REAL;
num_vars++;
Extend GUC concepts of parse_hook and assign_hook to all four supported datatypes, not only strings. parse_hook is useless for bool, I suppose, but it seems possibly useful for int and double to apply variable-specific constraints that are more complex than simple range limits. assign_hook is definitely useful for all datatypes --- we need it right now for bool to support date cache reset when changing Australian timezone rule setting. Also, clean up some residual problems with the reset all/show all patch, including memory leaks and mistaken reset of PostPortNumber. It seems best that RESET ALL not touch variables that don't have SUSET or USERSET context.
2001-06-13 00:54:06 +02:00
}
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
for (i = 0; ConfigureNamesString[i].gen.name; i++)
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
{
Extend GUC concepts of parse_hook and assign_hook to all four supported datatypes, not only strings. parse_hook is useless for bool, I suppose, but it seems possibly useful for int and double to apply variable-specific constraints that are more complex than simple range limits. assign_hook is definitely useful for all datatypes --- we need it right now for bool to support date cache reset when changing Australian timezone rule setting. Also, clean up some residual problems with the reset all/show all patch, including memory leaks and mistaken reset of PostPortNumber. It seems best that RESET ALL not touch variables that don't have SUSET or USERSET context.
2001-06-13 00:54:06 +02:00
struct config_string *conf = &ConfigureNamesString[i];
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
conf->gen.vartype = PGC_STRING;
num_vars++;
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
}
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
for (i = 0; ConfigureNamesEnum[i].gen.name; i++)
{
struct config_enum *conf = &ConfigureNamesEnum[i];
conf->gen.vartype = PGC_ENUM;
num_vars++;
}
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
/*
* Create table with 20% slack
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
*/
size_vars = num_vars + num_vars / 4;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
guc_vars = (struct config_generic **)
Fix unchecked mallocs/strdups added by recent placeholder-config-vars patch. Thomas Hallgren
2004-07-06 01:14:14 +02:00
guc_malloc(FATAL, size_vars * sizeof(struct config_generic *));
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
num_vars = 0;
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
for (i = 0; ConfigureNamesBool[i].gen.name; i++)
pgindent run.
2002-09-04 22:31:48 +02:00
guc_vars[num_vars++] = &ConfigureNamesBool[i].gen;
Add all possible config file options.
2001-01-24 19:37:31 +01:00
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
for (i = 0; ConfigureNamesInt[i].gen.name; i++)
pgindent run.
2002-09-04 22:31:48 +02:00
guc_vars[num_vars++] = &ConfigureNamesInt[i].gen;
Add all possible config file options.
2001-01-24 19:37:31 +01:00
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
for (i = 0; ConfigureNamesReal[i].gen.name; i++)
pgindent run.
2002-09-04 22:31:48 +02:00
guc_vars[num_vars++] = &ConfigureNamesReal[i].gen;
Add all possible config file options.
2001-01-24 19:37:31 +01:00
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
for (i = 0; ConfigureNamesString[i].gen.name; i++)
pgindent run.
2002-09-04 22:31:48 +02:00
guc_vars[num_vars++] = &ConfigureNamesString[i].gen;
Add all possible config file options.
2001-01-24 19:37:31 +01:00
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
for (i = 0; ConfigureNamesEnum[i].gen.name; i++)
guc_vars[num_vars++] = &ConfigureNamesEnum[i].gen;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
if (guc_variables)
free(guc_variables);
guc_variables = guc_vars;
num_guc_variables = num_vars;
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
size_guc_variables = size_vars;
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
qsort((void *) guc_variables, num_guc_variables,
sizeof(struct config_generic *), guc_var_compare);
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
}
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
/*
* Add a new GUC variable to the list of known variables. The
* list is expanded if needed.
*/
Fix unchecked mallocs/strdups added by recent placeholder-config-vars patch. Thomas Hallgren
2004-07-06 01:14:14 +02:00
static bool
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
add_guc_variable(struct config_generic * var, int elevel)
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
{
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
if (num_guc_variables + 1 >= size_guc_variables)
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
{
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
/*
* Increase the vector by 25%
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
*/
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
int size_vars = size_guc_variables + size_guc_variables / 4;
struct config_generic **guc_vars;
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
if (size_vars == 0)
Fix unchecked mallocs/strdups added by recent placeholder-config-vars patch. Thomas Hallgren
2004-07-06 01:14:14 +02:00
{
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
size_vars = 100;
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
guc_vars = (struct config_generic **)
guc_malloc(elevel, size_vars * sizeof(struct config_generic *));
Fix unchecked mallocs/strdups added by recent placeholder-config-vars patch. Thomas Hallgren
2004-07-06 01:14:14 +02:00
}
else
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
{
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
guc_vars = (struct config_generic **)
guc_realloc(elevel, guc_variables, size_vars * sizeof(struct config_generic *));
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
}
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
if (guc_vars == NULL)
Fix unchecked mallocs/strdups added by recent placeholder-config-vars patch. Thomas Hallgren
2004-07-06 01:14:14 +02:00
return false; /* out of memory */
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
guc_variables = guc_vars;
size_guc_variables = size_vars;
}
guc_variables[num_guc_variables++] = var;
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
qsort((void *) guc_variables, num_guc_variables,
sizeof(struct config_generic *), guc_var_compare);
Fix unchecked mallocs/strdups added by recent placeholder-config-vars patch. Thomas Hallgren
2004-07-06 01:14:14 +02:00
return true;
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
}
/*
Remove the custom_variable_classes parameter. This variable provides only marginal error-prevention capability (since it can only check the prefix of a qualified GUC name), and the consensus is that that isn't worth the amount of hassle that maintaining the setting creates for DBAs. So, let's just remove it. With this commit, the system will silently accept a value for any qualified GUC name at all, whether it has anything to do with any known extension or not. (Unqualified names still have to match known built-in settings, though; and you will get a WARNING at extension load time if there's an unrecognized setting with that extension's prefix.) There's still some discussion ongoing about whether to tighten that up and if so how; but if we do come up with a solution, it's not likely to look anything like custom_variable_classes.
2011-10-04 18:36:18 +02:00
* Create and add a placeholder variable for a custom variable name.
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
*/
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
static struct config_generic *
Fix unchecked mallocs/strdups added by recent placeholder-config-vars patch. Thomas Hallgren
2004-07-06 01:14:14 +02:00
add_placeholder_variable(const char *name, int elevel)
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
{
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
size_t sz = sizeof(struct config_string) + sizeof(char *);
struct config_string *var;
struct config_generic *gen;
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
var = (struct config_string *) guc_malloc(elevel, sz);
if (var == NULL)
Fix unchecked mallocs/strdups added by recent placeholder-config-vars patch. Thomas Hallgren
2004-07-06 01:14:14 +02:00
return NULL;
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
memset(var, 0, sz);
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
gen = &var->gen;
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
gen->name = guc_strdup(elevel, name);
if (gen->name == NULL)
Fix unchecked mallocs/strdups added by recent placeholder-config-vars patch. Thomas Hallgren
2004-07-06 01:14:14 +02:00
{
free(var);
return NULL;
}
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
gen->context = PGC_USERSET;
gen->group = CUSTOM_OPTIONS;
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
gen->short_desc = "GUC placeholder variable";
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
gen->flags = GUC_NO_SHOW_ALL | GUC_NOT_IN_SAMPLE | GUC_CUSTOM_PLACEHOLDER;
gen->vartype = PGC_STRING;
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
/*
* The char* is allocated at the end of the struct since we have no
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
* 'static' place to point to. Note that the current value, as well as
* the boot and reset values, start out NULL.
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
*/
var->variable = (char **) (var + 1);
Fix unchecked mallocs/strdups added by recent placeholder-config-vars patch. Thomas Hallgren
2004-07-06 01:14:14 +02:00
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
if (!add_guc_variable((struct config_generic *) var, elevel))
Fix unchecked mallocs/strdups added by recent placeholder-config-vars patch. Thomas Hallgren
2004-07-06 01:14:14 +02:00
{
free((void *) gen->name);
free(var);
return NULL;
}
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
return gen;
}
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
/*
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
* Look up option NAME. If it exists, return a pointer to its record,
* else return NULL. If create_placeholders is TRUE, we'll create a
* placeholder record for a valid-looking custom variable name.
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
*/
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
static struct config_generic *
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
find_option(const char *name, bool create_placeholders, int elevel)
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
{
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
const char **key = &name;
struct config_generic **res;
Rename SortMem and VacuumMem to work_mem and maintenance_work_mem. Make btree index creation and initial validation of foreign-key constraints use maintenance_work_mem rather than work_mem as their memory limit. Add some code to guc.c to allow these variables to be referenced by their old names in SHOW and SET commands, for backwards compatibility.
2004-02-03 18:34:04 +01:00
int i;
Add all possible config file options.
2001-01-24 19:37:31 +01:00
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
Assert(name);
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
/*
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
* By equating const char ** with struct config_generic *, we are assuming
* the name field is first in config_generic.
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
*/
pgindent run.
2002-09-04 22:31:48 +02:00
res = (struct config_generic **) bsearch((void *) &key,
(void *) guc_variables,
num_guc_variables,
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
sizeof(struct config_generic *),
pgindent run.
2002-09-04 22:31:48 +02:00
guc_var_compare);
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
if (res)
return *res;
Rename SortMem and VacuumMem to work_mem and maintenance_work_mem. Make btree index creation and initial validation of foreign-key constraints use maintenance_work_mem rather than work_mem as their memory limit. Add some code to guc.c to allow these variables to be referenced by their old names in SHOW and SET commands, for backwards compatibility.
2004-02-03 18:34:04 +01:00
/*
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
* See if the name is an obsolete name for a variable. We assume that the
* set of supported old names is short enough that a brute-force search is
* the best way.
Rename SortMem and VacuumMem to work_mem and maintenance_work_mem. Make btree index creation and initial validation of foreign-key constraints use maintenance_work_mem rather than work_mem as their memory limit. Add some code to guc.c to allow these variables to be referenced by their old names in SHOW and SET commands, for backwards compatibility.
2004-02-03 18:34:04 +01:00
*/
for (i = 0; map_old_guc_names[i] != NULL; i += 2)
{
if (guc_name_compare(name, map_old_guc_names[i]) == 0)
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
return find_option(map_old_guc_names[i + 1], false, elevel);
Rename SortMem and VacuumMem to work_mem and maintenance_work_mem. Make btree index creation and initial validation of foreign-key constraints use maintenance_work_mem rather than work_mem as their memory limit. Add some code to guc.c to allow these variables to be referenced by their old names in SHOW and SET commands, for backwards compatibility.
2004-02-03 18:34:04 +01:00
}
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
if (create_placeholders)
{
/*
Remove the custom_variable_classes parameter. This variable provides only marginal error-prevention capability (since it can only check the prefix of a qualified GUC name), and the consensus is that that isn't worth the amount of hassle that maintaining the setting creates for DBAs. So, let's just remove it. With this commit, the system will silently accept a value for any qualified GUC name at all, whether it has anything to do with any known extension or not. (Unqualified names still have to match known built-in settings, though; and you will get a WARNING at extension load time if there's an unrecognized setting with that extension's prefix.) There's still some discussion ongoing about whether to tighten that up and if so how; but if we do come up with a solution, it's not likely to look anything like custom_variable_classes.
2011-10-04 18:36:18 +02:00
* Check if the name is qualified, and if so, add a placeholder.
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
*/
Remove the custom_variable_classes parameter. This variable provides only marginal error-prevention capability (since it can only check the prefix of a qualified GUC name), and the consensus is that that isn't worth the amount of hassle that maintaining the setting creates for DBAs. So, let's just remove it. With this commit, the system will silently accept a value for any qualified GUC name at all, whether it has anything to do with any known extension or not. (Unqualified names still have to match known built-in settings, though; and you will get a WARNING at extension load time if there's an unrecognized setting with that extension's prefix.) There's still some discussion ongoing about whether to tighten that up and if so how; but if we do come up with a solution, it's not likely to look anything like custom_variable_classes.
2011-10-04 18:36:18 +02:00
if (strchr(name, GUC_QUALIFIER_SEPARATOR) != NULL)
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
return add_placeholder_variable(name, elevel);
}
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
Rename SortMem and VacuumMem to work_mem and maintenance_work_mem. Make btree index creation and initial validation of foreign-key constraints use maintenance_work_mem rather than work_mem as their memory limit. Add some code to guc.c to allow these variables to be referenced by their old names in SHOW and SET commands, for backwards compatibility.
2004-02-03 18:34:04 +01:00
/* Unknown name */
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
return NULL;
}
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
/*
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
* comparator for qsorting and bsearching guc_variables array
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
*/
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
static int
guc_var_compare(const void *a, const void *b)
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
{
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
struct config_generic *confa = *(struct config_generic **) a;
struct config_generic *confb = *(struct config_generic **) b;
Add all possible config file options.
2001-01-24 19:37:31 +01:00
Rename SortMem and VacuumMem to work_mem and maintenance_work_mem. Make btree index creation and initial validation of foreign-key constraints use maintenance_work_mem rather than work_mem as their memory limit. Add some code to guc.c to allow these variables to be referenced by their old names in SHOW and SET commands, for backwards compatibility.
2004-02-03 18:34:04 +01:00
return guc_name_compare(confa->name, confb->name);
}
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
/*
* the bare comparison function for GUC names
*/
Rename SortMem and VacuumMem to work_mem and maintenance_work_mem. Make btree index creation and initial validation of foreign-key constraints use maintenance_work_mem rather than work_mem as their memory limit. Add some code to guc.c to allow these variables to be referenced by their old names in SHOW and SET commands, for backwards compatibility.
2004-02-03 18:34:04 +01:00
static int
guc_name_compare(const char *namea, const char *nameb)
{
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
/*
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
* The temptation to use strcasecmp() here must be resisted, because the
* array ordering has to remain stable across setlocale() calls. So, build
* our own with a simple ASCII-only downcasing.
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
*/
Dept. of second thoughts: interaction between DoIt and makeDepend in set_config_option wasn't quite right. Also clean up a couple other things that could have been done better.
2002-05-17 22:32:29 +02:00
while (*namea && *nameb)
{
char cha = *namea++;
char chb = *nameb++;
if (cha >= 'A' && cha <= 'Z')
cha += 'a' - 'A';
if (chb >= 'A' && chb <= 'Z')
chb += 'a' - 'A';
if (cha != chb)
return cha - chb;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
}
Dept. of second thoughts: interaction between DoIt and makeDepend in set_config_option wasn't quite right. Also clean up a couple other things that could have been done better.
2002-05-17 22:32:29 +02:00
if (*namea)
return 1; /* a is longer */
if (*nameb)
return -1; /* b is longer */
return 0;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
}
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
/*
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
* Initialize GUC options during program startup.
Whack some sense into the configuration-file-location patch. Refactor code into something reasonably understandable, cause use of the feature to not fail in standalone backends or in EXEC_BACKEND case, fix sloppy guc.c table entries, make the documentation minimally usable.
2004-10-08 03:36:36 +02:00
*
* Note that we cannot read the config file yet, since we have not yet
* processed command-line switches.
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
*/
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
void
InitializeGUCOptions(void)
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
{
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
int i;
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
Fix crash caused by log_timezone patch if we attempt to emit any elog messages between the setting of log_line_prefix and the setting of log_timezone. We can't realistically set log_timezone any earlier than we do now, so the best behavior seems to be to use GMT zone if any timestamps are to be logged during early startup. Create a dummy zone variable with a minimal definition of GMT (in particular it will never know about leap seconds), so that we can set it up without reference to any external files.
2007-08-04 21:29:25 +02:00
/*
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
* Before log_line_prefix could possibly receive a nonempty setting, make
* sure that timezone processing is minimally alive (see elog.c).
Fix crash caused by log_timezone patch if we attempt to emit any elog messages between the setting of log_line_prefix and the setting of log_timezone. We can't realistically set log_timezone any earlier than we do now, so the best behavior seems to be to use GMT zone if any timestamps are to be logged during early startup. Create a dummy zone variable with a minimal definition of GMT (in particular it will never know about leap seconds), so that we can set it up without reference to any external files.
2007-08-04 21:29:25 +02:00
*/
Simplify handling of the timezone GUC by making initdb choose the default. We were doing some amazingly complicated things in order to avoid running the very expensive identify_system_timezone() procedure during GUC initialization. But there is an obvious fix for that, which is to do it once during initdb and have initdb install the system-specific default into postgresql.conf, as it already does for most other GUC variables that need system-environment-dependent defaults. This means that the timezone (and log_timezone) settings no longer have any magic behavior in the server. Per discussion.
2011-09-09 23:59:11 +02:00
pg_timezone_initialize();
Fix crash caused by log_timezone patch if we attempt to emit any elog messages between the setting of log_line_prefix and the setting of log_timezone. We can't realistically set log_timezone any earlier than we do now, so the best behavior seems to be to use GMT zone if any timestamps are to be logged during early startup. Create a dummy zone variable with a minimal definition of GMT (in particular it will never know about leap seconds), so that we can set it up without reference to any external files.
2007-08-04 21:29:25 +02:00
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
/*
* Build sorted array of all GUC variables.
*/
build_guc_variables();
Save source of GUC settings, allowing different sources to be processed in any order without affecting results.
2002-02-23 02:31:37 +01:00
Second pass over run-time configuration system. Adjust priorities on some option settings. Sort out SIGHUP vs BACKEND -- there is no total ordering here, so make explicit checks. Add comments explaining all of this. Removed permissions check on SHOW command. Add examine_subclass to the game, rename to SQL_inheritance to fit the official data model better. Adjust documentation. Standalone backend needs to reset all options before it starts. To facilitate that, have IsUnderPostmaster be set by the postmaster itself, don't wait for the magic -p switch. Also make sure that all environment variables and argv's survive init_ps_display(). Use strdup where necessary. Have initdb make configuration files (postgresql.conf, pg_hba.conf) mode 0600 -- having configuration files is no fun if you can't edit them.
2000-06-23 00:31:24 +02:00
/*
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
* Load all variables with their compiled-in defaults, and initialize
* status fields as needed.
Second pass over run-time configuration system. Adjust priorities on some option settings. Sort out SIGHUP vs BACKEND -- there is no total ordering here, so make explicit checks. Add comments explaining all of this. Removed permissions check on SHOW command. Add examine_subclass to the game, rename to SQL_inheritance to fit the official data model better. Adjust documentation. Standalone backend needs to reset all options before it starts. To facilitate that, have IsUnderPostmaster be set by the postmaster itself, don't wait for the magic -p switch. Also make sure that all environment variables and argv's survive init_ps_display(). Use strdup where necessary. Have initdb make configuration files (postgresql.conf, pg_hba.conf) mode 0600 -- having configuration files is no fun if you can't edit them.
2000-06-23 00:31:24 +02:00
*/
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
for (i = 0; i < num_guc_variables; i++)
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
{
Some infrastructure changes for the upcoming auto-explain contrib module: * Refactor explain.c slightly to export a convenient-to-use subroutine for printing EXPLAIN results. * Provide hooks for plugins to get control at ExecutorStart and ExecutorEnd as well as ExecutorRun. * Add some minimal support for tracking the total runtime of ExecutorRun. This code won't actually do anything unless a plugin prods it to. * Change the API of the DefineCustomXXXVariable functions to allow nonzero "flags" to be specified for a custom GUC variable. While at it, also make the "bootstrap" default value for custom GUCs be explicitly specified as a parameter to these functions. This is to eliminate confusion over where the default comes from, as has been expressed in the past by some users of the custom-variable facility. * Refactor GUC code a bit to ensure that a custom variable gets initialized to something valid (like its default value) even if the placeholder value was invalid.
2008-11-19 02:10:24 +01:00
InitializeOneGUCOption(guc_variables[i]);
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
}
guc_dirty = false;
In the continuing saga of FE/BE protocol revisions, add reporting of initial values and runtime changes in selected parameters. This gets rid of the need for an initial 'select pg_client_encoding()' query in libpq, bringing us back to one message transmitted in each direction for a standard connection startup. To allow server version to be sent using the same GUC mechanism that handles other parameters, invent the concept of a never-settable GUC parameter: you can 'show server_version' but it's not settable by any GUC input source. Create 'lc_collate' and 'lc_ctype' never-settable parameters so that people can find out these settings without need for pg_controldata. (These side ideas were all discussed some time ago in pgsql-hackers, but not yet implemented.)
2003-04-25 21:45:10 +02:00
reporting_enabled = false;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
/*
Read-only transactions, as defined in SQL.
2003-01-10 23:03:30 +01:00
* Prevent any attempt to override the transaction modes from
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
* non-interactive sources.
*/
Read-only transactions, as defined in SQL.
2003-01-10 23:03:30 +01:00
SetConfigOption("transaction_isolation", "default",
PGC_POSTMASTER, PGC_S_OVERRIDE);
SetConfigOption("transaction_read_only", "no",
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
PGC_POSTMASTER, PGC_S_OVERRIDE);
Implement genuine serializable isolation level. Until now, our Serializable mode has in fact been what's called Snapshot Isolation, which allows some anomalies that could not occur in any serialized ordering of the transactions. This patch fixes that using a method called Serializable Snapshot Isolation, based on research papers by Michael J. Cahill (see README-SSI for full references). In Serializable Snapshot Isolation, transactions run like they do in Snapshot Isolation, but a predicate lock manager observes the reads and writes performed and aborts transactions if it detects that an anomaly might occur. This method produces some false positives, ie. it sometimes aborts transactions even though there is no anomaly. To track reads we implement predicate locking, see storage/lmgr/predicate.c. Whenever a tuple is read, a predicate lock is acquired on the tuple. Shared memory is finite, so when a transaction takes many tuple-level locks on a page, the locks are promoted to a single page-level lock, and further to a single relation level lock if necessary. To lock key values with no matching tuple, a sequential scan always takes a relation-level lock, and an index scan acquires a page-level lock that covers the search key, whether or not there are any matching keys at the moment. A predicate lock doesn't conflict with any regular locks or with another predicate locks in the normal sense. They're only used by the predicate lock manager to detect the danger of anomalies. Only serializable transactions participate in predicate locking, so there should be no extra overhead for for other transactions. Predicate locks can't be released at commit, but must be remembered until all the transactions that overlapped with it have completed. That means that we need to remember an unbounded amount of predicate locks, so we apply a lossy but conservative method of tracking locks for committed transactions. If we run short of shared memory, we overflow to a new "pg_serial" SLRU pool. We don't currently allow Serializable transactions in Hot Standby mode. That would be hard, because even read-only transactions can cause anomalies that wouldn't otherwise occur. Serializable isolation mode now means the new fully serializable level. Repeatable Read gives you the old Snapshot Isolation level that we have always had. Kevin Grittner and Dan Ports, reviewed by Jeff Davis, Heikki Linnakangas and Anssi Kääriäinen
2011-02-07 22:46:51 +01:00
SetConfigOption("transaction_deferrable", "no",
PGC_POSTMASTER, PGC_S_OVERRIDE);
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
/*
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
* For historical reasons, some GUC parameters can receive defaults from
Split PGC_S_DEFAULT into two values, for true boot_val vs computed default. Failure to distinguish these cases is the real cause behind the recent reports of Windows builds crashing on 'infinity'::timestamp, which was directly due to failure to establish a value of timezone_abbreviations in postmaster child processes. The postmaster had the desired value, but write_one_nondefault_variable() didn't transmit it to backends. To fix that, invent a new value PGC_S_DYNAMIC_DEFAULT, and be sure to use that or PGC_S_ENV_VAR (as appropriate) for "default" settings that are computed during initialization. (We need both because there's at least one variable that could receive a value from either source.) This commit also fixes ProcessConfigFile's failure to restore the correct default value for certain GUC variables if they are set in postgresql.conf and then removed/commented out of the file. We have to recompute and reinstall the value for any GUC variable that could have received a value from PGC_S_DYNAMIC_DEFAULT or PGC_S_ENV_VAR sources, and there were a number of oversights. (That whole thing is a crock that needs to be redesigned, but not today.) However, I intentionally didn't make it work "exactly right" for the cases of timezone and log_timezone. The exactly right behavior would involve running select_default_timezone, which we'd have to do independently in each postgres process, causing the whole database to become entirely unresponsive for as much as several seconds. That didn't seem like a good idea, especially since the variable's removal from postgresql.conf might be just an accidental edit. Instead the behavior is to adopt the previously active setting as if it were default. Note that this patch creates an ABI break for extensions that use any of the PGC_S_XXX constants; they'll need to be recompiled.
2011-05-12 01:57:38 +02:00
* environment variables. Process those settings.
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
*/
Split PGC_S_DEFAULT into two values, for true boot_val vs computed default. Failure to distinguish these cases is the real cause behind the recent reports of Windows builds crashing on 'infinity'::timestamp, which was directly due to failure to establish a value of timezone_abbreviations in postmaster child processes. The postmaster had the desired value, but write_one_nondefault_variable() didn't transmit it to backends. To fix that, invent a new value PGC_S_DYNAMIC_DEFAULT, and be sure to use that or PGC_S_ENV_VAR (as appropriate) for "default" settings that are computed during initialization. (We need both because there's at least one variable that could receive a value from either source.) This commit also fixes ProcessConfigFile's failure to restore the correct default value for certain GUC variables if they are set in postgresql.conf and then removed/commented out of the file. We have to recompute and reinstall the value for any GUC variable that could have received a value from PGC_S_DYNAMIC_DEFAULT or PGC_S_ENV_VAR sources, and there were a number of oversights. (That whole thing is a crock that needs to be redesigned, but not today.) However, I intentionally didn't make it work "exactly right" for the cases of timezone and log_timezone. The exactly right behavior would involve running select_default_timezone, which we'd have to do independently in each postgres process, causing the whole database to become entirely unresponsive for as much as several seconds. That didn't seem like a good idea, especially since the variable's removal from postgresql.conf might be just an accidental edit. Instead the behavior is to adopt the previously active setting as if it were default. Note that this patch creates an ABI break for extensions that use any of the PGC_S_XXX constants; they'll need to be recompiled.
2011-05-12 01:57:38 +02:00
InitializeGUCOptionsFromEnvironment();
}
/*
* Assign any GUC values that can come from the server's environment.
*
* This is called from InitializeGUCOptions, and also from ProcessConfigFile
* to deal with the possibility that a setting has been removed from
* postgresql.conf and should now get a value from the environment.
* (The latter is a kludge that should probably go away someday; if so,
* fold this back into InitializeGUCOptions.)
*/
static void
InitializeGUCOptionsFromEnvironment(void)
{
char *env;
long stack_rlimit;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
env = getenv("PGPORT");
if (env != NULL)
SetConfigOption("port", env, PGC_POSTMASTER, PGC_S_ENV_VAR);
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
env = getenv("PGDATESTYLE");
if (env != NULL)
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
SetConfigOption("datestyle", env, PGC_POSTMASTER, PGC_S_ENV_VAR);
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
env = getenv("PGCLIENTENCODING");
if (env != NULL)
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
SetConfigOption("client_encoding", env, PGC_POSTMASTER, PGC_S_ENV_VAR);
On platforms that have getrlimit(RLIMIT_STACK), use it to ensure that max_stack_depth is not set to an unsafe value. This commit also provides configure-time checking for <sys/resource.h>, and cleans up some perhaps-unportable code associated with use of that include file and getrlimit().
2006-10-07 21:25:29 +02:00
/*
* rlimit isn't exactly an "environment variable", but it behaves about
* the same. If we can identify the platform stack depth rlimit, increase
* default stack depth setting up to whatever is safe (but at most 2MB).
*/
stack_rlimit = get_stack_depth_rlimit();
if (stack_rlimit > 0)
{
Make get_stack_depth_rlimit() handle RLIM_INFINITY more sanely. Rather than considering this result as meaning "unknown", report LONG_MAX. This won't change what superusers can set max_stack_depth to, but it will cause InitializeGUCOptions() to set the built-in default to 2MB not 100kB. The latter seems like a fairly unreasonable interpretation of "infinity". Per my investigation of odd buildfarm results as well as an old complaint from Heikki. Since this should persuade all the buildfarm animals to use a reasonable stack depth setting during "make check", revert previous patch that dumbed down a recursive regression test to only 5 levels.
2010-11-06 21:50:18 +01:00
long new_limit = (stack_rlimit - STACK_DEPTH_SLOP) / 1024L;
On platforms that have getrlimit(RLIMIT_STACK), use it to ensure that max_stack_depth is not set to an unsafe value. This commit also provides configure-time checking for <sys/resource.h>, and cleans up some perhaps-unportable code associated with use of that include file and getrlimit().
2006-10-07 21:25:29 +02:00
if (new_limit > 100)
{
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
char limbuf[16];
On platforms that have getrlimit(RLIMIT_STACK), use it to ensure that max_stack_depth is not set to an unsafe value. This commit also provides configure-time checking for <sys/resource.h>, and cleans up some perhaps-unportable code associated with use of that include file and getrlimit().
2006-10-07 21:25:29 +02:00
new_limit = Min(new_limit, 2048);
Make get_stack_depth_rlimit() handle RLIM_INFINITY more sanely. Rather than considering this result as meaning "unknown", report LONG_MAX. This won't change what superusers can set max_stack_depth to, but it will cause InitializeGUCOptions() to set the built-in default to 2MB not 100kB. The latter seems like a fairly unreasonable interpretation of "infinity". Per my investigation of odd buildfarm results as well as an old complaint from Heikki. Since this should persuade all the buildfarm animals to use a reasonable stack depth setting during "make check", revert previous patch that dumbed down a recursive regression test to only 5 levels.
2010-11-06 21:50:18 +01:00
sprintf(limbuf, "%ld", new_limit);
On platforms that have getrlimit(RLIMIT_STACK), use it to ensure that max_stack_depth is not set to an unsafe value. This commit also provides configure-time checking for <sys/resource.h>, and cleans up some perhaps-unportable code associated with use of that include file and getrlimit().
2006-10-07 21:25:29 +02:00
SetConfigOption("max_stack_depth", limbuf,
PGC_POSTMASTER, PGC_S_ENV_VAR);
}
}
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
}
Some infrastructure changes for the upcoming auto-explain contrib module: * Refactor explain.c slightly to export a convenient-to-use subroutine for printing EXPLAIN results. * Provide hooks for plugins to get control at ExecutorStart and ExecutorEnd as well as ExecutorRun. * Add some minimal support for tracking the total runtime of ExecutorRun. This code won't actually do anything unless a plugin prods it to. * Change the API of the DefineCustomXXXVariable functions to allow nonzero "flags" to be specified for a custom GUC variable. While at it, also make the "bootstrap" default value for custom GUCs be explicitly specified as a parameter to these functions. This is to eliminate confusion over where the default comes from, as has been expressed in the past by some users of the custom-variable facility. * Refactor GUC code a bit to ensure that a custom variable gets initialized to something valid (like its default value) even if the placeholder value was invalid.
2008-11-19 02:10:24 +01:00
/*
* Initialize one GUC option variable to its compiled-in default.
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
*
* Note: the reason for calling check_hooks is not that we think the boot_val
* might fail, but that the hooks might wish to compute an "extra" struct.
Some infrastructure changes for the upcoming auto-explain contrib module: * Refactor explain.c slightly to export a convenient-to-use subroutine for printing EXPLAIN results. * Provide hooks for plugins to get control at ExecutorStart and ExecutorEnd as well as ExecutorRun. * Add some minimal support for tracking the total runtime of ExecutorRun. This code won't actually do anything unless a plugin prods it to. * Change the API of the DefineCustomXXXVariable functions to allow nonzero "flags" to be specified for a custom GUC variable. While at it, also make the "bootstrap" default value for custom GUCs be explicitly specified as a parameter to these functions. This is to eliminate confusion over where the default comes from, as has been expressed in the past by some users of the custom-variable facility. * Refactor GUC code a bit to ensure that a custom variable gets initialized to something valid (like its default value) even if the placeholder value was invalid.
2008-11-19 02:10:24 +01:00
*/
static void
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
InitializeOneGUCOption(struct config_generic * gconf)
Some infrastructure changes for the upcoming auto-explain contrib module: * Refactor explain.c slightly to export a convenient-to-use subroutine for printing EXPLAIN results. * Provide hooks for plugins to get control at ExecutorStart and ExecutorEnd as well as ExecutorRun. * Add some minimal support for tracking the total runtime of ExecutorRun. This code won't actually do anything unless a plugin prods it to. * Change the API of the DefineCustomXXXVariable functions to allow nonzero "flags" to be specified for a custom GUC variable. While at it, also make the "bootstrap" default value for custom GUCs be explicitly specified as a parameter to these functions. This is to eliminate confusion over where the default comes from, as has been expressed in the past by some users of the custom-variable facility. * Refactor GUC code a bit to ensure that a custom variable gets initialized to something valid (like its default value) even if the placeholder value was invalid.
2008-11-19 02:10:24 +01:00
{
gconf->status = 0;
gconf->source = PGC_S_DEFAULT;
Remember the source GucContext for each GUC parameter. We used to just remember the GucSource, but saving GucContext too provides a little more information --- notably, whether a SET was done by a superuser or regular user. This allows us to rip out the fairly dodgy code that define_custom_variable used to use to try to infer the context to re-install a pre-existing setting with. In particular, it now works for a superuser to SET a extension's SUSET custom variable before loading the associated extension, because GUC can remember whether the SET was done as a superuser or not. The plperl regression tests contain an example where this is useful.
2011-10-04 22:13:16 +02:00
gconf->reset_source = PGC_S_DEFAULT;
gconf->scontext = PGC_INTERNAL;
gconf->reset_scontext = PGC_INTERNAL;
Some infrastructure changes for the upcoming auto-explain contrib module: * Refactor explain.c slightly to export a convenient-to-use subroutine for printing EXPLAIN results. * Provide hooks for plugins to get control at ExecutorStart and ExecutorEnd as well as ExecutorRun. * Add some minimal support for tracking the total runtime of ExecutorRun. This code won't actually do anything unless a plugin prods it to. * Change the API of the DefineCustomXXXVariable functions to allow nonzero "flags" to be specified for a custom GUC variable. While at it, also make the "bootstrap" default value for custom GUCs be explicitly specified as a parameter to these functions. This is to eliminate confusion over where the default comes from, as has been expressed in the past by some users of the custom-variable facility. * Refactor GUC code a bit to ensure that a custom variable gets initialized to something valid (like its default value) even if the placeholder value was invalid.
2008-11-19 02:10:24 +01:00
gconf->stack = NULL;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
gconf->extra = NULL;
Some infrastructure changes for the upcoming auto-explain contrib module: * Refactor explain.c slightly to export a convenient-to-use subroutine for printing EXPLAIN results. * Provide hooks for plugins to get control at ExecutorStart and ExecutorEnd as well as ExecutorRun. * Add some minimal support for tracking the total runtime of ExecutorRun. This code won't actually do anything unless a plugin prods it to. * Change the API of the DefineCustomXXXVariable functions to allow nonzero "flags" to be specified for a custom GUC variable. While at it, also make the "bootstrap" default value for custom GUCs be explicitly specified as a parameter to these functions. This is to eliminate confusion over where the default comes from, as has been expressed in the past by some users of the custom-variable facility. * Refactor GUC code a bit to ensure that a custom variable gets initialized to something valid (like its default value) even if the placeholder value was invalid.
2008-11-19 02:10:24 +01:00
gconf->sourcefile = NULL;
gconf->sourceline = 0;
switch (gconf->vartype)
{
case PGC_BOOL:
{
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
struct config_bool *conf = (struct config_bool *) gconf;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
bool newval = conf->boot_val;
void *extra = NULL;
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
if (!call_bool_check_hook(conf, &newval, &extra,
PGC_S_DEFAULT, LOG))
elog(FATAL, "failed to initialize %s to %d",
conf->gen.name, (int) newval);
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
if (conf->assign_hook)
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
(*conf->assign_hook) (newval, extra);
*conf->variable = conf->reset_val = newval;
conf->gen.extra = conf->reset_extra = extra;
Some infrastructure changes for the upcoming auto-explain contrib module: * Refactor explain.c slightly to export a convenient-to-use subroutine for printing EXPLAIN results. * Provide hooks for plugins to get control at ExecutorStart and ExecutorEnd as well as ExecutorRun. * Add some minimal support for tracking the total runtime of ExecutorRun. This code won't actually do anything unless a plugin prods it to. * Change the API of the DefineCustomXXXVariable functions to allow nonzero "flags" to be specified for a custom GUC variable. While at it, also make the "bootstrap" default value for custom GUCs be explicitly specified as a parameter to these functions. This is to eliminate confusion over where the default comes from, as has been expressed in the past by some users of the custom-variable facility. * Refactor GUC code a bit to ensure that a custom variable gets initialized to something valid (like its default value) even if the placeholder value was invalid.
2008-11-19 02:10:24 +01:00
break;
}
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
case PGC_INT:
{
struct config_int *conf = (struct config_int *) gconf;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
int newval = conf->boot_val;
void *extra = NULL;
Assert(newval >= conf->min);
Assert(newval <= conf->max);
if (!call_int_check_hook(conf, &newval, &extra,
PGC_S_DEFAULT, LOG))
elog(FATAL, "failed to initialize %s to %d",
conf->gen.name, newval);
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
if (conf->assign_hook)
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
(*conf->assign_hook) (newval, extra);
*conf->variable = conf->reset_val = newval;
conf->gen.extra = conf->reset_extra = extra;
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
break;
}
case PGC_REAL:
{
struct config_real *conf = (struct config_real *) gconf;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
double newval = conf->boot_val;
void *extra = NULL;
Assert(newval >= conf->min);
Assert(newval <= conf->max);
if (!call_real_check_hook(conf, &newval, &extra,
PGC_S_DEFAULT, LOG))
elog(FATAL, "failed to initialize %s to %g",
conf->gen.name, newval);
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
if (conf->assign_hook)
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
(*conf->assign_hook) (newval, extra);
*conf->variable = conf->reset_val = newval;
conf->gen.extra = conf->reset_extra = extra;
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
break;
}
case PGC_STRING:
Some infrastructure changes for the upcoming auto-explain contrib module: * Refactor explain.c slightly to export a convenient-to-use subroutine for printing EXPLAIN results. * Provide hooks for plugins to get control at ExecutorStart and ExecutorEnd as well as ExecutorRun. * Add some minimal support for tracking the total runtime of ExecutorRun. This code won't actually do anything unless a plugin prods it to. * Change the API of the DefineCustomXXXVariable functions to allow nonzero "flags" to be specified for a custom GUC variable. While at it, also make the "bootstrap" default value for custom GUCs be explicitly specified as a parameter to these functions. This is to eliminate confusion over where the default comes from, as has been expressed in the past by some users of the custom-variable facility. * Refactor GUC code a bit to ensure that a custom variable gets initialized to something valid (like its default value) even if the placeholder value was invalid.
2008-11-19 02:10:24 +01:00
{
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
struct config_string *conf = (struct config_string *) gconf;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
char *newval;
void *extra = NULL;
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
/* non-NULL boot_val must always get strdup'd */
if (conf->boot_val != NULL)
newval = guc_strdup(FATAL, conf->boot_val);
else
newval = NULL;
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
if (!call_string_check_hook(conf, &newval, &extra,
PGC_S_DEFAULT, LOG))
elog(FATAL, "failed to initialize %s to \"%s\"",
conf->gen.name, newval ? newval : "");
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
if (conf->assign_hook)
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
(*conf->assign_hook) (newval, extra);
*conf->variable = conf->reset_val = newval;
conf->gen.extra = conf->reset_extra = extra;
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
break;
Some infrastructure changes for the upcoming auto-explain contrib module: * Refactor explain.c slightly to export a convenient-to-use subroutine for printing EXPLAIN results. * Provide hooks for plugins to get control at ExecutorStart and ExecutorEnd as well as ExecutorRun. * Add some minimal support for tracking the total runtime of ExecutorRun. This code won't actually do anything unless a plugin prods it to. * Change the API of the DefineCustomXXXVariable functions to allow nonzero "flags" to be specified for a custom GUC variable. While at it, also make the "bootstrap" default value for custom GUCs be explicitly specified as a parameter to these functions. This is to eliminate confusion over where the default comes from, as has been expressed in the past by some users of the custom-variable facility. * Refactor GUC code a bit to ensure that a custom variable gets initialized to something valid (like its default value) even if the placeholder value was invalid.
2008-11-19 02:10:24 +01:00
}
case PGC_ENUM:
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
{
struct config_enum *conf = (struct config_enum *) gconf;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
int newval = conf->boot_val;
void *extra = NULL;
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
if (!call_enum_check_hook(conf, &newval, &extra,
PGC_S_DEFAULT, LOG))
elog(FATAL, "failed to initialize %s to %d",
conf->gen.name, newval);
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
if (conf->assign_hook)
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
(*conf->assign_hook) (newval, extra);
*conf->variable = conf->reset_val = newval;
conf->gen.extra = conf->reset_extra = extra;
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
break;
}
Some infrastructure changes for the upcoming auto-explain contrib module: * Refactor explain.c slightly to export a convenient-to-use subroutine for printing EXPLAIN results. * Provide hooks for plugins to get control at ExecutorStart and ExecutorEnd as well as ExecutorRun. * Add some minimal support for tracking the total runtime of ExecutorRun. This code won't actually do anything unless a plugin prods it to. * Change the API of the DefineCustomXXXVariable functions to allow nonzero "flags" to be specified for a custom GUC variable. While at it, also make the "bootstrap" default value for custom GUCs be explicitly specified as a parameter to these functions. This is to eliminate confusion over where the default comes from, as has been expressed in the past by some users of the custom-variable facility. * Refactor GUC code a bit to ensure that a custom variable gets initialized to something valid (like its default value) even if the placeholder value was invalid.
2008-11-19 02:10:24 +01:00
}
}
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
Whack some sense into the configuration-file-location patch. Refactor code into something reasonably understandable, cause use of the feature to not fail in standalone backends or in EXEC_BACKEND case, fix sloppy guc.c table entries, make the documentation minimally usable.
2004-10-08 03:36:36 +02:00
/*
* Select the configuration files and data directory to be used, and
* do the initial read of postgresql.conf.
*
* This is called after processing command-line switches.
* userDoption is the -D switch value if any (NULL if unspecified).
* progname is just for use in error messages.
*
* Returns true on success; on failure, prints a suitable error message
* to stderr and returns false.
*/
bool
SelectConfigFiles(const char *userDoption, const char *progname)
{
Adjust configuration-files GUC behavior as per my recent proposal. The vars are renamed to data_directory, config_file, hba_file, and ident_file, and are guaranteed to be set to accurate absolute paths during postmaster startup. This commit does not yet do anything about hiding path values from non-superusers.
2004-10-10 01:13:22 +02:00
char *configdir;
char *fname;
Whack some sense into the configuration-file-location patch. Refactor code into something reasonably understandable, cause use of the feature to not fail in standalone backends or in EXEC_BACKEND case, fix sloppy guc.c table entries, make the documentation minimally usable.
2004-10-08 03:36:36 +02:00
struct stat stat_buf;
Adjust configuration-files GUC behavior as per my recent proposal. The vars are renamed to data_directory, config_file, hba_file, and ident_file, and are guaranteed to be set to accurate absolute paths during postmaster startup. This commit does not yet do anything about hiding path values from non-superusers.
2004-10-10 01:13:22 +02:00
/* configdir is -D option, or $PGDATA if no -D */
if (userDoption)
configdir = make_absolute_path(userDoption);
else
configdir = make_absolute_path(getenv("PGDATA"));
Whack some sense into the configuration-file-location patch. Refactor code into something reasonably understandable, cause use of the feature to not fail in standalone backends or in EXEC_BACKEND case, fix sloppy guc.c table entries, make the documentation minimally usable.
2004-10-08 03:36:36 +02:00
/*
Adjust configuration-files GUC behavior as per my recent proposal. The vars are renamed to data_directory, config_file, hba_file, and ident_file, and are guaranteed to be set to accurate absolute paths during postmaster startup. This commit does not yet do anything about hiding path values from non-superusers.
2004-10-10 01:13:22 +02:00
* Find the configuration file: if config_file was specified on the
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
* command line, use it, else use configdir/postgresql.conf. In any case
* ensure the result is an absolute path, so that it will be interpreted
* the same way by future backends.
Whack some sense into the configuration-file-location patch. Refactor code into something reasonably understandable, cause use of the feature to not fail in standalone backends or in EXEC_BACKEND case, fix sloppy guc.c table entries, make the documentation minimally usable.
2004-10-08 03:36:36 +02:00
*/
Adjust configuration-files GUC behavior as per my recent proposal. The vars are renamed to data_directory, config_file, hba_file, and ident_file, and are guaranteed to be set to accurate absolute paths during postmaster startup. This commit does not yet do anything about hiding path values from non-superusers.
2004-10-10 01:13:22 +02:00
if (ConfigFileName)
fname = make_absolute_path(ConfigFileName);
Re-order 'else' clause for clarity.
2005-05-27 20:33:30 +02:00
else if (configdir)
{
fname = guc_malloc(FATAL,
strlen(configdir) + strlen(CONFIG_FILENAME) + 2);
sprintf(fname, "%s/%s", configdir, CONFIG_FILENAME);
}
else
Whack some sense into the configuration-file-location patch. Refactor code into something reasonably understandable, cause use of the feature to not fail in standalone backends or in EXEC_BACKEND case, fix sloppy guc.c table entries, make the documentation minimally usable.
2004-10-08 03:36:36 +02:00
{
Adjust configuration-files GUC behavior as per my recent proposal. The vars are renamed to data_directory, config_file, hba_file, and ident_file, and are guaranteed to be set to accurate absolute paths during postmaster startup. This commit does not yet do anything about hiding path values from non-superusers.
2004-10-10 01:13:22 +02:00
write_stderr("%s does not know where to find the server configuration file.\n"
Message style revisions
2004-10-12 23:54:45 +02:00
"You must specify the --config-file or -D invocation "
Adjust configuration-files GUC behavior as per my recent proposal. The vars are renamed to data_directory, config_file, hba_file, and ident_file, and are guaranteed to be set to accurate absolute paths during postmaster startup. This commit does not yet do anything about hiding path values from non-superusers.
2004-10-10 01:13:22 +02:00
"option or set the PGDATA environment variable.\n",
progname);
return false;
Whack some sense into the configuration-file-location patch. Refactor code into something reasonably understandable, cause use of the feature to not fail in standalone backends or in EXEC_BACKEND case, fix sloppy guc.c table entries, make the documentation minimally usable.
2004-10-08 03:36:36 +02:00
}
Adjust configuration-files GUC behavior as per my recent proposal. The vars are renamed to data_directory, config_file, hba_file, and ident_file, and are guaranteed to be set to accurate absolute paths during postmaster startup. This commit does not yet do anything about hiding path values from non-superusers.
2004-10-10 01:13:22 +02:00
/*
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
* Set the ConfigFileName GUC variable to its final value, ensuring that
* it can't be overridden later.
Adjust configuration-files GUC behavior as per my recent proposal. The vars are renamed to data_directory, config_file, hba_file, and ident_file, and are guaranteed to be set to accurate absolute paths during postmaster startup. This commit does not yet do anything about hiding path values from non-superusers.
2004-10-10 01:13:22 +02:00
*/
SetConfigOption("config_file", fname, PGC_POSTMASTER, PGC_S_OVERRIDE);
free(fname);
/*
* Now read the config file for the first time.
*/
Whack some sense into the configuration-file-location patch. Refactor code into something reasonably understandable, cause use of the feature to not fail in standalone backends or in EXEC_BACKEND case, fix sloppy guc.c table entries, make the documentation minimally usable.
2004-10-08 03:36:36 +02:00
if (stat(ConfigFileName, &stat_buf) != 0)
{
Adjust configuration-files GUC behavior as per my recent proposal. The vars are renamed to data_directory, config_file, hba_file, and ident_file, and are guaranteed to be set to accurate absolute paths during postmaster startup. This commit does not yet do anything about hiding path values from non-superusers.
2004-10-10 01:13:22 +02:00
write_stderr("%s cannot access the server configuration file \"%s\": %s\n",
Whack some sense into the configuration-file-location patch. Refactor code into something reasonably understandable, cause use of the feature to not fail in standalone backends or in EXEC_BACKEND case, fix sloppy guc.c table entries, make the documentation minimally usable.
2004-10-08 03:36:36 +02:00
progname, ConfigFileName, strerror(errno));
return false;
}
ProcessConfigFile(PGC_POSTMASTER);
/*
Adjust configuration-files GUC behavior as per my recent proposal. The vars are renamed to data_directory, config_file, hba_file, and ident_file, and are guaranteed to be set to accurate absolute paths during postmaster startup. This commit does not yet do anything about hiding path values from non-superusers.
2004-10-10 01:13:22 +02:00
* If the data_directory GUC variable has been set, use that as DataDir;
* otherwise use configdir if set; else punt.
Whack some sense into the configuration-file-location patch. Refactor code into something reasonably understandable, cause use of the feature to not fail in standalone backends or in EXEC_BACKEND case, fix sloppy guc.c table entries, make the documentation minimally usable.
2004-10-08 03:36:36 +02:00
*
Re-run pgindent, fixing a problem where comment lines after a blank comment line where output as too long, and update typedefs for /lib directory. Also fix case where identifiers were used as variable names in the backend, but as typedefs in ecpg (favor the backend for indenting). Backpatch to 8.1.X.
2005-11-22 19:17:34 +01:00
* Note: SetDataDir will copy and absolute-ize its argument, so we don't
* have to.
Whack some sense into the configuration-file-location patch. Refactor code into something reasonably understandable, cause use of the feature to not fail in standalone backends or in EXEC_BACKEND case, fix sloppy guc.c table entries, make the documentation minimally usable.
2004-10-08 03:36:36 +02:00
*/
Adjust configuration-files GUC behavior as per my recent proposal. The vars are renamed to data_directory, config_file, hba_file, and ident_file, and are guaranteed to be set to accurate absolute paths during postmaster startup. This commit does not yet do anything about hiding path values from non-superusers.
2004-10-10 01:13:22 +02:00
if (data_directory)
SetDataDir(data_directory);
else if (configdir)
SetDataDir(configdir);
Whack some sense into the configuration-file-location patch. Refactor code into something reasonably understandable, cause use of the feature to not fail in standalone backends or in EXEC_BACKEND case, fix sloppy guc.c table entries, make the documentation minimally usable.
2004-10-08 03:36:36 +02:00
else
{
write_stderr("%s does not know where to find the database system data.\n"
Adjust configuration-files GUC behavior as per my recent proposal. The vars are renamed to data_directory, config_file, hba_file, and ident_file, and are guaranteed to be set to accurate absolute paths during postmaster startup. This commit does not yet do anything about hiding path values from non-superusers.
2004-10-10 01:13:22 +02:00
"This can be specified as \"data_directory\" in \"%s\", "
"or by the -D invocation option, or by the "
"PGDATA environment variable.\n",
progname, ConfigFileName);
return false;
}
/*
* Reflect the final DataDir value back into the data_directory GUC var.
* (If you are wondering why we don't just make them a single variable,
* it's because the EXEC_BACKEND case needs DataDir to be transmitted to
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
* child backends specially. XXX is that still true? Given that we now
* chdir to DataDir, EXEC_BACKEND can read the config file without knowing
* DataDir in advance.)
Adjust configuration-files GUC behavior as per my recent proposal. The vars are renamed to data_directory, config_file, hba_file, and ident_file, and are guaranteed to be set to accurate absolute paths during postmaster startup. This commit does not yet do anything about hiding path values from non-superusers.
2004-10-10 01:13:22 +02:00
*/
SetConfigOption("data_directory", DataDir, PGC_POSTMASTER, PGC_S_OVERRIDE);
Simplify handling of the timezone GUC by making initdb choose the default. We were doing some amazingly complicated things in order to avoid running the very expensive identify_system_timezone() procedure during GUC initialization. But there is an obvious fix for that, which is to do it once during initdb and have initdb install the system-specific default into postgresql.conf, as it already does for most other GUC variables that need system-environment-dependent defaults. This means that the timezone (and log_timezone) settings no longer have any magic behavior in the server. Per discussion.
2011-09-09 23:59:11 +02:00
/*
* If timezone_abbreviations wasn't set in the configuration file, install
* the default value. We do it this way because we can't safely install
* a "real" value until my_exec_path is set, which may not have happened
* when InitializeGUCOptions runs, so the bootstrap default value cannot
* be the real desired default.
*/
pg_timezone_abbrev_initialize();
Adjust configuration-files GUC behavior as per my recent proposal. The vars are renamed to data_directory, config_file, hba_file, and ident_file, and are guaranteed to be set to accurate absolute paths during postmaster startup. This commit does not yet do anything about hiding path values from non-superusers.
2004-10-10 01:13:22 +02:00
/*
* Figure out where pg_hba.conf is, and make sure the path is absolute.
*/
if (HbaFileName)
fname = make_absolute_path(HbaFileName);
Re-order 'else' clause for clarity.
2005-05-27 20:33:30 +02:00
else if (configdir)
{
fname = guc_malloc(FATAL,
strlen(configdir) + strlen(HBA_FILENAME) + 2);
sprintf(fname, "%s/%s", configdir, HBA_FILENAME);
}
else
Adjust configuration-files GUC behavior as per my recent proposal. The vars are renamed to data_directory, config_file, hba_file, and ident_file, and are guaranteed to be set to accurate absolute paths during postmaster startup. This commit does not yet do anything about hiding path values from non-superusers.
2004-10-10 01:13:22 +02:00
{
write_stderr("%s does not know where to find the \"hba\" configuration file.\n"
"This can be specified as \"hba_file\" in \"%s\", "
"or by the -D invocation option, or by the "
"PGDATA environment variable.\n",
Whack some sense into the configuration-file-location patch. Refactor code into something reasonably understandable, cause use of the feature to not fail in standalone backends or in EXEC_BACKEND case, fix sloppy guc.c table entries, make the documentation minimally usable.
2004-10-08 03:36:36 +02:00
progname, ConfigFileName);
return false;
}
Adjust configuration-files GUC behavior as per my recent proposal. The vars are renamed to data_directory, config_file, hba_file, and ident_file, and are guaranteed to be set to accurate absolute paths during postmaster startup. This commit does not yet do anything about hiding path values from non-superusers.
2004-10-10 01:13:22 +02:00
SetConfigOption("hba_file", fname, PGC_POSTMASTER, PGC_S_OVERRIDE);
free(fname);
Whack some sense into the configuration-file-location patch. Refactor code into something reasonably understandable, cause use of the feature to not fail in standalone backends or in EXEC_BACKEND case, fix sloppy guc.c table entries, make the documentation minimally usable.
2004-10-08 03:36:36 +02:00
/*
Adjust configuration-files GUC behavior as per my recent proposal. The vars are renamed to data_directory, config_file, hba_file, and ident_file, and are guaranteed to be set to accurate absolute paths during postmaster startup. This commit does not yet do anything about hiding path values from non-superusers.
2004-10-10 01:13:22 +02:00
* Likewise for pg_ident.conf.
Whack some sense into the configuration-file-location patch. Refactor code into something reasonably understandable, cause use of the feature to not fail in standalone backends or in EXEC_BACKEND case, fix sloppy guc.c table entries, make the documentation minimally usable.
2004-10-08 03:36:36 +02:00
*/
Adjust configuration-files GUC behavior as per my recent proposal. The vars are renamed to data_directory, config_file, hba_file, and ident_file, and are guaranteed to be set to accurate absolute paths during postmaster startup. This commit does not yet do anything about hiding path values from non-superusers.
2004-10-10 01:13:22 +02:00
if (IdentFileName)
fname = make_absolute_path(IdentFileName);
Re-order 'else' clause for clarity.
2005-05-27 20:33:30 +02:00
else if (configdir)
{
fname = guc_malloc(FATAL,
strlen(configdir) + strlen(IDENT_FILENAME) + 2);
sprintf(fname, "%s/%s", configdir, IDENT_FILENAME);
}
else
Adjust configuration-files GUC behavior as per my recent proposal. The vars are renamed to data_directory, config_file, hba_file, and ident_file, and are guaranteed to be set to accurate absolute paths during postmaster startup. This commit does not yet do anything about hiding path values from non-superusers.
2004-10-10 01:13:22 +02:00
{
write_stderr("%s does not know where to find the \"ident\" configuration file.\n"
"This can be specified as \"ident_file\" in \"%s\", "
"or by the -D invocation option, or by the "
"PGDATA environment variable.\n",
progname, ConfigFileName);
return false;
}
SetConfigOption("ident_file", fname, PGC_POSTMASTER, PGC_S_OVERRIDE);
free(fname);
free(configdir);
Whack some sense into the configuration-file-location patch. Refactor code into something reasonably understandable, cause use of the feature to not fail in standalone backends or in EXEC_BACKEND case, fix sloppy guc.c table entries, make the documentation minimally usable.
2004-10-08 03:36:36 +02:00
return true;
}
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
/*
* Reset all options to their saved default values (implements RESET ALL)
*/
void
ResetAllOptions(void)
{
int i;
for (i = 0; i < num_guc_variables; i++)
{
struct config_generic *gconf = guc_variables[i];
/* Don't reset non-SET-able values */
Add new USERLIMIT GUC source level so certain options can be disabled or increased only by super-users. This fixes problems caused by making certain variables SUSET for security reasons.
2003-07-09 08:47:34 +02:00
if (gconf->context != PGC_SUSET &&
gconf->context != PGC_USERSET)
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
continue;
/* Don't reset if special exclusion from RESET ALL */
if (gconf->flags & GUC_NO_RESET_ALL)
continue;
/* No need to reset if wasn't SET */
if (gconf->source <= PGC_S_OVERRIDE)
continue;
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
/* Save old value to support transaction abort */
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
push_old_value(gconf, GUC_ACTION_SET);
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
switch (gconf->vartype)
{
case PGC_BOOL:
pgindent run.
2002-09-04 22:31:48 +02:00
{
struct config_bool *conf = (struct config_bool *) gconf;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
pgindent run.
2002-09-04 22:31:48 +02:00
if (conf->assign_hook)
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
(*conf->assign_hook) (conf->reset_val,
conf->reset_extra);
pgindent run.
2002-09-04 22:31:48 +02:00
*conf->variable = conf->reset_val;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
set_extra_field(&conf->gen, &conf->gen.extra,
conf->reset_extra);
pgindent run.
2002-09-04 22:31:48 +02:00
break;
}
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
case PGC_INT:
pgindent run.
2002-09-04 22:31:48 +02:00
{
struct config_int *conf = (struct config_int *) gconf;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
pgindent run.
2002-09-04 22:31:48 +02:00
if (conf->assign_hook)
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
(*conf->assign_hook) (conf->reset_val,
conf->reset_extra);
pgindent run.
2002-09-04 22:31:48 +02:00
*conf->variable = conf->reset_val;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
set_extra_field(&conf->gen, &conf->gen.extra,
conf->reset_extra);
pgindent run.
2002-09-04 22:31:48 +02:00
break;
}
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
case PGC_REAL:
{
pgindent run.
2002-09-04 22:31:48 +02:00
struct config_real *conf = (struct config_real *) gconf;
if (conf->assign_hook)
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
(*conf->assign_hook) (conf->reset_val,
conf->reset_extra);
pgindent run.
2002-09-04 22:31:48 +02:00
*conf->variable = conf->reset_val;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
set_extra_field(&conf->gen, &conf->gen.extra,
conf->reset_extra);
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
break;
}
pgindent run.
2002-09-04 22:31:48 +02:00
case PGC_STRING:
{
struct config_string *conf = (struct config_string *) gconf;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
if (conf->assign_hook)
(*conf->assign_hook) (conf->reset_val,
conf->reset_extra);
set_string_field(conf, conf->variable, conf->reset_val);
set_extra_field(&conf->gen, &conf->gen.extra,
conf->reset_extra);
pgindent run.
2002-09-04 22:31:48 +02:00
break;
}
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
case PGC_ENUM:
{
struct config_enum *conf = (struct config_enum *) gconf;
if (conf->assign_hook)
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
(*conf->assign_hook) (conf->reset_val,
conf->reset_extra);
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
*conf->variable = conf->reset_val;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
set_extra_field(&conf->gen, &conf->gen.extra,
conf->reset_extra);
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
break;
}
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
}
In the continuing saga of FE/BE protocol revisions, add reporting of initial values and runtime changes in selected parameters. This gets rid of the need for an initial 'select pg_client_encoding()' query in libpq, bringing us back to one message transmitted in each direction for a standard connection startup. To allow server version to be sent using the same GUC mechanism that handles other parameters, invent the concept of a never-settable GUC parameter: you can 'show server_version' but it's not settable by any GUC input source. Create 'lc_collate' and 'lc_ctype' never-settable parameters so that people can find out these settings without need for pg_controldata. (These side ideas were all discussed some time ago in pgsql-hackers, but not yet implemented.)
2003-04-25 21:45:10 +02:00
Add "source file" and "source line" information to each GUC variable. initdb forced due to changes in the pg_settings view. Magnus Hagander and Alvaro Herrera.
2008-09-10 20:09:20 +02:00
gconf->source = gconf->reset_source;
Remember the source GucContext for each GUC parameter. We used to just remember the GucSource, but saving GucContext too provides a little more information --- notably, whether a SET was done by a superuser or regular user. This allows us to rip out the fairly dodgy code that define_custom_variable used to use to try to infer the context to re-install a pre-existing setting with. In particular, it now works for a superuser to SET a extension's SUSET custom variable before loading the associated extension, because GUC can remember whether the SET was done as a superuser or not. The plperl regression tests contain an example where this is useful.
2011-10-04 22:13:16 +02:00
gconf->scontext = gconf->reset_scontext;
Add "source file" and "source line" information to each GUC variable. initdb forced due to changes in the pg_settings view. Magnus Hagander and Alvaro Herrera.
2008-09-10 20:09:20 +02:00
In the continuing saga of FE/BE protocol revisions, add reporting of initial values and runtime changes in selected parameters. This gets rid of the need for an initial 'select pg_client_encoding()' query in libpq, bringing us back to one message transmitted in each direction for a standard connection startup. To allow server version to be sent using the same GUC mechanism that handles other parameters, invent the concept of a never-settable GUC parameter: you can 'show server_version' but it's not settable by any GUC input source. Create 'lc_collate' and 'lc_ctype' never-settable parameters so that people can find out these settings without need for pg_controldata. (These side ideas were all discussed some time ago in pgsql-hackers, but not yet implemented.)
2003-04-25 21:45:10 +02:00
if (gconf->flags & GUC_REPORT)
ReportGUCOption(gconf);
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
}
}
/*
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
* push_old_value
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
* Push previous state during transactional assignment to a GUC variable.
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
*/
static void
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
push_old_value(struct config_generic * gconf, GucAction action)
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
{
GucStack *stack;
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
/* If we're not inside a nest level, do nothing */
Implement function-local GUC parameter settings, as per recent discussion. There are still some loose ends: I didn't do anything about the SET FROM CURRENT idea yet, and it's not real clear whether we are happy with the interaction of SET LOCAL with function-local settings. The documentation is a bit spartan, too.
2007-09-03 02:39:26 +02:00
if (GUCNestLevel == 0)
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
return;
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
/* Do we already have a stack entry of the current nest level? */
stack = gconf->stack;
if (stack && stack->nest_level >= GUCNestLevel)
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
{
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
/* Yes, so adjust its state if necessary */
Assert(stack->nest_level == GUCNestLevel);
switch (action)
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
{
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
case GUC_ACTION_SET:
/* SET overrides any prior action at same nest level */
if (stack->state == GUC_SET_LOCAL)
{
/* must discard old masked value */
discard_stack_value(gconf, &stack->masked);
}
stack->state = GUC_SET;
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
break;
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
case GUC_ACTION_LOCAL:
if (stack->state == GUC_SET)
{
/* SET followed by SET LOCAL, remember SET's value */
Remember the source GucContext for each GUC parameter. We used to just remember the GucSource, but saving GucContext too provides a little more information --- notably, whether a SET was done by a superuser or regular user. This allows us to rip out the fairly dodgy code that define_custom_variable used to use to try to infer the context to re-install a pre-existing setting with. In particular, it now works for a superuser to SET a extension's SUSET custom variable before loading the associated extension, because GUC can remember whether the SET was done as a superuser or not. The plperl regression tests contain an example where this is useful.
2011-10-04 22:13:16 +02:00
stack->masked_scontext = gconf->scontext;
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
set_stack_value(gconf, &stack->masked);
stack->state = GUC_SET_LOCAL;
}
/* in all other cases, no change to stack entry */
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
break;
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
case GUC_ACTION_SAVE:
/* Could only have a prior SAVE of same variable */
Assert(stack->state == GUC_SAVE);
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
break;
}
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
Assert(guc_dirty); /* must be set already */
return;
}
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
/*
* Push a new stack entry
*
* We keep all the stack entries in TopTransactionContext for simplicity.
*/
stack = (GucStack *) MemoryContextAllocZero(TopTransactionContext,
sizeof(GucStack));
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
stack->prev = gconf->stack;
stack->nest_level = GUCNestLevel;
switch (action)
{
case GUC_ACTION_SET:
stack->state = GUC_SET;
break;
case GUC_ACTION_LOCAL:
stack->state = GUC_LOCAL;
break;
case GUC_ACTION_SAVE:
stack->state = GUC_SAVE;
break;
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
}
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
stack->source = gconf->source;
Remember the source GucContext for each GUC parameter. We used to just remember the GucSource, but saving GucContext too provides a little more information --- notably, whether a SET was done by a superuser or regular user. This allows us to rip out the fairly dodgy code that define_custom_variable used to use to try to infer the context to re-install a pre-existing setting with. In particular, it now works for a superuser to SET a extension's SUSET custom variable before loading the associated extension, because GUC can remember whether the SET was done as a superuser or not. The plperl regression tests contain an example where this is useful.
2011-10-04 22:13:16 +02:00
stack->scontext = gconf->scontext;
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
set_stack_value(gconf, &stack->prior);
gconf->stack = stack;
/* Ensure we remember to pop at end of xact */
guc_dirty = true;
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
}
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
Reverted waiting for further fixes: Make configuration parameters fall back to their default values when they are removed from the configuration file. Joachim Wieland
2007-03-13 15:32:25 +01:00
/*
Implement function-local GUC parameter settings, as per recent discussion. There are still some loose ends: I didn't do anything about the SET FROM CURRENT idea yet, and it's not real clear whether we are happy with the interaction of SET LOCAL with function-local settings. The documentation is a bit spartan, too.
2007-09-03 02:39:26 +02:00
* Do GUC processing at main transaction start.
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
*/
void
Implement function-local GUC parameter settings, as per recent discussion. There are still some loose ends: I didn't do anything about the SET FROM CURRENT idea yet, and it's not real clear whether we are happy with the interaction of SET LOCAL with function-local settings. The documentation is a bit spartan, too.
2007-09-03 02:39:26 +02:00
AtStart_GUC(void)
{
/*
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
* The nest level should be 0 between transactions; if it isn't, somebody
* didn't call AtEOXact_GUC, or called it with the wrong nestLevel. We
* throw a warning but make no other effort to clean up.
Implement function-local GUC parameter settings, as per recent discussion. There are still some loose ends: I didn't do anything about the SET FROM CURRENT idea yet, and it's not real clear whether we are happy with the interaction of SET LOCAL with function-local settings. The documentation is a bit spartan, too.
2007-09-03 02:39:26 +02:00
*/
if (GUCNestLevel != 0)
elog(WARNING, "GUC nest level = %d at transaction start",
GUCNestLevel);
GUCNestLevel = 1;
}
/*
* Enter a new nesting level for GUC values. This is called at subtransaction
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
* start and when entering a function that has proconfig settings. NOTE that
Implement function-local GUC parameter settings, as per recent discussion. There are still some loose ends: I didn't do anything about the SET FROM CURRENT idea yet, and it's not real clear whether we are happy with the interaction of SET LOCAL with function-local settings. The documentation is a bit spartan, too.
2007-09-03 02:39:26 +02:00
* we must not risk error here, else subtransaction start will be unhappy.
*/
int
NewGUCNestLevel(void)
{
return ++GUCNestLevel;
}
/*
* Do GUC processing at transaction or subtransaction commit or abort, or
* when exiting a function that has proconfig settings. (The name is thus
* a bit of a misnomer; perhaps it should be ExitGUCNestLevel or some such.)
* During abort, we discard all GUC settings that were applied at nesting
* levels >= nestLevel. nestLevel == 1 corresponds to the main transaction.
*/
void
AtEOXact_GUC(bool isCommit, int nestLevel)
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
{
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
bool still_dirty;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
int i;
Fix assorted core dumps and Assert failures that could occur during AbortTransaction or AbortSubTransaction, when trying to clean up after an error that prevented (sub)transaction start from completing: * access to TopTransactionResourceOwner that might not exist * assert failure in AtEOXact_GUC, if AtStart_GUC not called yet * assert failure or core dump in AfterTriggerEndSubXact, if AfterTriggerBeginSubXact not called yet Per testing by injecting elog(ERROR) at successive steps in StartTransaction and StartSubTransaction. It's not clear whether all of these cases could really occur in the field, but at least one of them is easily exposed by simple stress testing, as per my accidental discovery yesterday.
2010-01-24 22:49:17 +01:00
/*
* Note: it's possible to get here with GUCNestLevel == nestLevel-1 during
* abort, if there is a failure during transaction start before
* AtStart_GUC is called.
*/
Assert(nestLevel > 0 &&
(nestLevel <= GUCNestLevel ||
(nestLevel == GUCNestLevel + 1 && !isCommit)));
Implement function-local GUC parameter settings, as per recent discussion. There are still some loose ends: I didn't do anything about the SET FROM CURRENT idea yet, and it's not real clear whether we are happy with the interaction of SET LOCAL with function-local settings. The documentation is a bit spartan, too.
2007-09-03 02:39:26 +02:00
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
/* Quick exit if nothing's changed in this transaction */
if (!guc_dirty)
Implement function-local GUC parameter settings, as per recent discussion. There are still some loose ends: I didn't do anything about the SET FROM CURRENT idea yet, and it's not real clear whether we are happy with the interaction of SET LOCAL with function-local settings. The documentation is a bit spartan, too.
2007-09-03 02:39:26 +02:00
{
GUCNestLevel = nestLevel - 1;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
return;
Implement function-local GUC parameter settings, as per recent discussion. There are still some loose ends: I didn't do anything about the SET FROM CURRENT idea yet, and it's not real clear whether we are happy with the interaction of SET LOCAL with function-local settings. The documentation is a bit spartan, too.
2007-09-03 02:39:26 +02:00
}
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
still_dirty = false;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
for (i = 0; i < num_guc_variables; i++)
{
struct config_generic *gconf = guc_variables[i];
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
GucStack *stack;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
/*
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
* Process and pop each stack entry within the nest level. To
* simplify fmgr_security_definer(), we allow failure exit from a
* function-with-SET-options to be recovered at the surrounding
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
* transaction or subtransaction abort; so there could be more than
* one stack entry to pop.
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
*/
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
while ((stack = gconf->stack) != NULL &&
stack->nest_level >= nestLevel)
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
{
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
GucStack *prev = stack->prev;
bool restorePrior = false;
bool restoreMasked = false;
bool changed;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
/*
* In this next bit, if we don't set either restorePrior or
* restoreMasked, we must "discard" any unwanted fields of the
* stack entries to avoid leaking memory. If we do set one of
* those flags, unused fields will be cleaned up after restoring.
*/
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
if (!isCommit) /* if abort, always restore prior value */
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
restorePrior = true;
else if (stack->state == GUC_SAVE)
restorePrior = true;
else if (stack->nest_level == 1)
{
/* transaction commit */
if (stack->state == GUC_SET_LOCAL)
restoreMasked = true;
else if (stack->state == GUC_SET)
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
{
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
/* we keep the current active value */
discard_stack_value(gconf, &stack->prior);
}
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
else /* must be GUC_LOCAL */
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
restorePrior = true;
}
else if (prev == NULL ||
prev->nest_level < stack->nest_level - 1)
{
/* decrement entry's level and do not pop it */
stack->nest_level--;
continue;
}
else
{
/*
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
* We have to merge this stack entry into prev. See README for
* discussion of this bit.
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
*/
switch (stack->state)
{
case GUC_SAVE:
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
Assert(false); /* can't get here */
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
case GUC_SET:
/* next level always becomes SET */
discard_stack_value(gconf, &stack->prior);
if (prev->state == GUC_SET_LOCAL)
discard_stack_value(gconf, &prev->masked);
prev->state = GUC_SET;
break;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
case GUC_LOCAL:
if (prev->state == GUC_SET)
{
/* LOCAL migrates down */
Remember the source GucContext for each GUC parameter. We used to just remember the GucSource, but saving GucContext too provides a little more information --- notably, whether a SET was done by a superuser or regular user. This allows us to rip out the fairly dodgy code that define_custom_variable used to use to try to infer the context to re-install a pre-existing setting with. In particular, it now works for a superuser to SET a extension's SUSET custom variable before loading the associated extension, because GUC can remember whether the SET was done as a superuser or not. The plperl regression tests contain an example where this is useful.
2011-10-04 22:13:16 +02:00
prev->masked_scontext = stack->scontext;
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
prev->masked = stack->prior;
prev->state = GUC_SET_LOCAL;
}
else
{
/* else just forget this stack level */
discard_stack_value(gconf, &stack->prior);
}
break;
pgindent run.
2002-09-04 22:31:48 +02:00
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
case GUC_SET_LOCAL:
/* prior state at this level no longer wanted */
discard_stack_value(gconf, &stack->prior);
/* copy down the masked state */
Remember the source GucContext for each GUC parameter. We used to just remember the GucSource, but saving GucContext too provides a little more information --- notably, whether a SET was done by a superuser or regular user. This allows us to rip out the fairly dodgy code that define_custom_variable used to use to try to infer the context to re-install a pre-existing setting with. In particular, it now works for a superuser to SET a extension's SUSET custom variable before loading the associated extension, because GUC can remember whether the SET was done as a superuser or not. The plperl regression tests contain an example where this is useful.
2011-10-04 22:13:16 +02:00
prev->masked_scontext = stack->masked_scontext;
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
if (prev->state == GUC_SET_LOCAL)
discard_stack_value(gconf, &prev->masked);
prev->masked = stack->masked;
prev->state = GUC_SET_LOCAL;
break;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
}
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
}
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
changed = false;
pgindent run.
2002-09-04 22:31:48 +02:00
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
if (restorePrior || restoreMasked)
{
/* Perform appropriate restoration of the stacked value */
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
config_var_value newvalue;
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
GucSource newsource;
Remember the source GucContext for each GUC parameter. We used to just remember the GucSource, but saving GucContext too provides a little more information --- notably, whether a SET was done by a superuser or regular user. This allows us to rip out the fairly dodgy code that define_custom_variable used to use to try to infer the context to re-install a pre-existing setting with. In particular, it now works for a superuser to SET a extension's SUSET custom variable before loading the associated extension, because GUC can remember whether the SET was done as a superuser or not. The plperl regression tests contain an example where this is useful.
2011-10-04 22:13:16 +02:00
GucContext newscontext;
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
if (restoreMasked)
{
newvalue = stack->masked;
newsource = PGC_S_SESSION;
Remember the source GucContext for each GUC parameter. We used to just remember the GucSource, but saving GucContext too provides a little more information --- notably, whether a SET was done by a superuser or regular user. This allows us to rip out the fairly dodgy code that define_custom_variable used to use to try to infer the context to re-install a pre-existing setting with. In particular, it now works for a superuser to SET a extension's SUSET custom variable before loading the associated extension, because GUC can remember whether the SET was done as a superuser or not. The plperl regression tests contain an example where this is useful.
2011-10-04 22:13:16 +02:00
newscontext = stack->masked_scontext;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
}
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
else
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
{
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
newvalue = stack->prior;
newsource = stack->source;
Remember the source GucContext for each GUC parameter. We used to just remember the GucSource, but saving GucContext too provides a little more information --- notably, whether a SET was done by a superuser or regular user. This allows us to rip out the fairly dodgy code that define_custom_variable used to use to try to infer the context to re-install a pre-existing setting with. In particular, it now works for a superuser to SET a extension's SUSET custom variable before loading the associated extension, because GUC can remember whether the SET was done as a superuser or not. The plperl regression tests contain an example where this is useful.
2011-10-04 22:13:16 +02:00
newscontext = stack->scontext;
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
}
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
switch (gconf->vartype)
{
case PGC_BOOL:
{
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
struct config_bool *conf = (struct config_bool *) gconf;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
bool newval = newvalue.val.boolval;
void *newextra = newvalue.extra;
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
if (*conf->variable != newval ||
conf->gen.extra != newextra)
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
{
if (conf->assign_hook)
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
(*conf->assign_hook) (newval, newextra);
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
*conf->variable = newval;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
set_extra_field(&conf->gen, &conf->gen.extra,
newextra);
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
changed = true;
}
break;
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
}
case PGC_INT:
{
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
struct config_int *conf = (struct config_int *) gconf;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
int newval = newvalue.val.intval;
void *newextra = newvalue.extra;
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
if (*conf->variable != newval ||
conf->gen.extra != newextra)
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
{
if (conf->assign_hook)
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
(*conf->assign_hook) (newval, newextra);
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
*conf->variable = newval;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
set_extra_field(&conf->gen, &conf->gen.extra,
newextra);
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
changed = true;
}
break;
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
}
case PGC_REAL:
{
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
struct config_real *conf = (struct config_real *) gconf;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
double newval = newvalue.val.realval;
void *newextra = newvalue.extra;
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
if (*conf->variable != newval ||
conf->gen.extra != newextra)
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
{
if (conf->assign_hook)
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
(*conf->assign_hook) (newval, newextra);
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
*conf->variable = newval;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
set_extra_field(&conf->gen, &conf->gen.extra,
newextra);
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
changed = true;
}
break;
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
}
case PGC_STRING:
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
{
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
struct config_string *conf = (struct config_string *) gconf;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
char *newval = newvalue.val.stringval;
void *newextra = newvalue.extra;
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
if (*conf->variable != newval ||
conf->gen.extra != newextra)
pgindent run.
2002-09-04 22:31:48 +02:00
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
if (conf->assign_hook)
(*conf->assign_hook) (newval, newextra);
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
set_string_field(conf, conf->variable, newval);
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
set_extra_field(&conf->gen, &conf->gen.extra,
newextra);
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
changed = true;
pgindent run.
2002-09-04 22:31:48 +02:00
}
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
/*
* Release stacked values if not used anymore. We
* could use discard_stack_value() here, but since
* we have type-specific code anyway, might as
* well inline it.
*/
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
set_string_field(conf, &stack->prior.val.stringval, NULL);
set_string_field(conf, &stack->masked.val.stringval, NULL);
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
break;
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
}
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
case PGC_ENUM:
{
struct config_enum *conf = (struct config_enum *) gconf;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
int newval = newvalue.val.enumval;
void *newextra = newvalue.extra;
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
if (*conf->variable != newval ||
conf->gen.extra != newextra)
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
{
if (conf->assign_hook)
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
(*conf->assign_hook) (newval, newextra);
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
*conf->variable = newval;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
set_extra_field(&conf->gen, &conf->gen.extra,
newextra);
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
changed = true;
}
break;
}
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
}
In the continuing saga of FE/BE protocol revisions, add reporting of initial values and runtime changes in selected parameters. This gets rid of the need for an initial 'select pg_client_encoding()' query in libpq, bringing us back to one message transmitted in each direction for a standard connection startup. To allow server version to be sent using the same GUC mechanism that handles other parameters, invent the concept of a never-settable GUC parameter: you can 'show server_version' but it's not settable by any GUC input source. Create 'lc_collate' and 'lc_ctype' never-settable parameters so that people can find out these settings without need for pg_controldata. (These side ideas were all discussed some time ago in pgsql-hackers, but not yet implemented.)
2003-04-25 21:45:10 +02:00
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
/*
* Release stacked extra values if not used anymore.
*/
set_extra_field(gconf, &(stack->prior.extra), NULL);
set_extra_field(gconf, &(stack->masked.extra), NULL);
Remember the source GucContext for each GUC parameter. We used to just remember the GucSource, but saving GucContext too provides a little more information --- notably, whether a SET was done by a superuser or regular user. This allows us to rip out the fairly dodgy code that define_custom_variable used to use to try to infer the context to re-install a pre-existing setting with. In particular, it now works for a superuser to SET a extension's SUSET custom variable before loading the associated extension, because GUC can remember whether the SET was done as a superuser or not. The plperl regression tests contain an example where this is useful.
2011-10-04 22:13:16 +02:00
/* And restore source information */
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
gconf->source = newsource;
Remember the source GucContext for each GUC parameter. We used to just remember the GucSource, but saving GucContext too provides a little more information --- notably, whether a SET was done by a superuser or regular user. This allows us to rip out the fairly dodgy code that define_custom_variable used to use to try to infer the context to re-install a pre-existing setting with. In particular, it now works for a superuser to SET a extension's SUSET custom variable before loading the associated extension, because GUC can remember whether the SET was done as a superuser or not. The plperl regression tests contain an example where this is useful.
2011-10-04 22:13:16 +02:00
gconf->scontext = newscontext;
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
}
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
/* Finish popping the state stack */
gconf->stack = prev;
pfree(stack);
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
/* Report new value if we changed it */
if (changed && (gconf->flags & GUC_REPORT))
ReportGUCOption(gconf);
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
} /* end of stack-popping loop */
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
if (stack != NULL)
still_dirty = true;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
}
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
/* If there are no remaining stack entries, we can reset guc_dirty */
guc_dirty = still_dirty;
Implement function-local GUC parameter settings, as per recent discussion. There are still some loose ends: I didn't do anything about the SET FROM CURRENT idea yet, and it's not real clear whether we are happy with the interaction of SET LOCAL with function-local settings. The documentation is a bit spartan, too.
2007-09-03 02:39:26 +02:00
/* Update nesting level */
GUCNestLevel = nestLevel - 1;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
}
In the continuing saga of FE/BE protocol revisions, add reporting of initial values and runtime changes in selected parameters. This gets rid of the need for an initial 'select pg_client_encoding()' query in libpq, bringing us back to one message transmitted in each direction for a standard connection startup. To allow server version to be sent using the same GUC mechanism that handles other parameters, invent the concept of a never-settable GUC parameter: you can 'show server_version' but it's not settable by any GUC input source. Create 'lc_collate' and 'lc_ctype' never-settable parameters so that people can find out these settings without need for pg_controldata. (These side ideas were all discussed some time ago in pgsql-hackers, but not yet implemented.)
2003-04-25 21:45:10 +02:00
/*
* Start up automatic reporting of changes to variables marked GUC_REPORT.
* This is executed at completion of backend startup.
*/
void
BeginReportingGUCOptions(void)
{
int i;
/*
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
* Don't do anything unless talking to an interactive frontend of protocol
* 3.0 or later.
In the continuing saga of FE/BE protocol revisions, add reporting of initial values and runtime changes in selected parameters. This gets rid of the need for an initial 'select pg_client_encoding()' query in libpq, bringing us back to one message transmitted in each direction for a standard connection startup. To allow server version to be sent using the same GUC mechanism that handles other parameters, invent the concept of a never-settable GUC parameter: you can 'show server_version' but it's not settable by any GUC input source. Create 'lc_collate' and 'lc_ctype' never-settable parameters so that people can find out these settings without need for pg_controldata. (These side ideas were all discussed some time ago in pgsql-hackers, but not yet implemented.)
2003-04-25 21:45:10 +02:00
*/
Rename the members of CommandDest enum so they don't collide with other uses of those names. (Debug and None were pretty bad names anyway.) I hope I catched all uses of the names in comments too.
2005-11-03 18:11:40 +01:00
if (whereToSendOutput != DestRemote ||
In the continuing saga of FE/BE protocol revisions, add reporting of initial values and runtime changes in selected parameters. This gets rid of the need for an initial 'select pg_client_encoding()' query in libpq, bringing us back to one message transmitted in each direction for a standard connection startup. To allow server version to be sent using the same GUC mechanism that handles other parameters, invent the concept of a never-settable GUC parameter: you can 'show server_version' but it's not settable by any GUC input source. Create 'lc_collate' and 'lc_ctype' never-settable parameters so that people can find out these settings without need for pg_controldata. (These side ideas were all discussed some time ago in pgsql-hackers, but not yet implemented.)
2003-04-25 21:45:10 +02:00
PG_PROTOCOL_MAJOR(FrontendProtocol) < 3)
return;
reporting_enabled = true;
/* Transmit initial values of interesting variables */
for (i = 0; i < num_guc_variables; i++)
{
struct config_generic *conf = guc_variables[i];
if (conf->flags & GUC_REPORT)
ReportGUCOption(conf);
}
}
/*
* ReportGUCOption: if appropriate, transmit option value to frontend
*/
static void
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
ReportGUCOption(struct config_generic * record)
In the continuing saga of FE/BE protocol revisions, add reporting of initial values and runtime changes in selected parameters. This gets rid of the need for an initial 'select pg_client_encoding()' query in libpq, bringing us back to one message transmitted in each direction for a standard connection startup. To allow server version to be sent using the same GUC mechanism that handles other parameters, invent the concept of a never-settable GUC parameter: you can 'show server_version' but it's not settable by any GUC input source. Create 'lc_collate' and 'lc_ctype' never-settable parameters so that people can find out these settings without need for pg_controldata. (These side ideas were all discussed some time ago in pgsql-hackers, but not yet implemented.)
2003-04-25 21:45:10 +02:00
{
if (reporting_enabled && (record->flags & GUC_REPORT))
{
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
char *val = _ShowOption(record, false);
In the continuing saga of FE/BE protocol revisions, add reporting of initial values and runtime changes in selected parameters. This gets rid of the need for an initial 'select pg_client_encoding()' query in libpq, bringing us back to one message transmitted in each direction for a standard connection startup. To allow server version to be sent using the same GUC mechanism that handles other parameters, invent the concept of a never-settable GUC parameter: you can 'show server_version' but it's not settable by any GUC input source. Create 'lc_collate' and 'lc_ctype' never-settable parameters so that people can find out these settings without need for pg_controldata. (These side ideas were all discussed some time ago in pgsql-hackers, but not yet implemented.)
2003-04-25 21:45:10 +02:00
StringInfoData msgbuf;
pq_beginmessage(&msgbuf, 'S');
pq_sendstring(&msgbuf, record->name);
pq_sendstring(&msgbuf, val);
pq_endmessage(&msgbuf);
pfree(val);
}
}
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
/*
* Try to parse value as an integer. The accepted formats are the
Provide a HINT listing the allowed unit names when a GUC variable seems to contain a wrong unit specification, per discussion. In passing, fix the code to avoid unnecessary integer overflows when converting units, and to detect overflows when they do occur.
2007-06-21 20:14:21 +02:00
* usual decimal, octal, or hexadecimal formats, optionally followed by
* a unit name if "flags" indicates a unit is allowed.
*
* If the string parses okay, return true, else false.
* If okay and result is not NULL, return the value in *result.
* If not okay and hintmsg is not NULL, *hintmsg is set to a suitable
* HINT message, or NULL if no hint provided.
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
*/
Use guc.c's parse_int() instead of pg_atoi() to parse fillfactor in default_reloptions(). The previous coding was really a bug because pg_atoi() will always throw elog on bad input data, whereas default_reloptions is not supposed to complain about bad input unless its validate parameter is true. Right now you could only expose the problem by hand-modifying pg_class.reloptions into an invalid state, so it doesn't seem worth back-patching; but we should get it right in HEAD because there might be other situations in future. Noted while studying GIN fast-update patch.
2008-07-23 19:29:53 +02:00
bool
Provide a HINT listing the allowed unit names when a GUC variable seems to contain a wrong unit specification, per discussion. In passing, fix the code to avoid unnecessary integer overflows when converting units, and to detect overflows when they do occur.
2007-06-21 20:14:21 +02:00
parse_int(const char *value, int *result, int flags, const char **hintmsg)
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
{
Provide a HINT listing the allowed unit names when a GUC variable seems to contain a wrong unit specification, per discussion. In passing, fix the code to avoid unnecessary integer overflows when converting units, and to detect overflows when they do occur.
2007-06-21 20:14:21 +02:00
int64 val;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
char *endptr;
Provide a HINT listing the allowed unit names when a GUC variable seems to contain a wrong unit specification, per discussion. In passing, fix the code to avoid unnecessary integer overflows when converting units, and to detect overflows when they do occur.
2007-06-21 20:14:21 +02:00
/* To suppress compiler warnings, always set output params */
if (result)
*result = 0;
if (hintmsg)
*hintmsg = NULL;
/* We assume here that int64 is at least as wide as long */
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
errno = 0;
val = strtol(value, &endptr, 0);
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
Provide a HINT listing the allowed unit names when a GUC variable seems to contain a wrong unit specification, per discussion. In passing, fix the code to avoid unnecessary integer overflows when converting units, and to detect overflows when they do occur.
2007-06-21 20:14:21 +02:00
if (endptr == value)
return false; /* no HINT for integer syntax error */
if (errno == ERANGE || val != (int64) ((int32) val))
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
{
Provide a HINT listing the allowed unit names when a GUC variable seems to contain a wrong unit specification, per discussion. In passing, fix the code to avoid unnecessary integer overflows when converting units, and to detect overflows when they do occur.
2007-06-21 20:14:21 +02:00
if (hintmsg)
*hintmsg = gettext_noop("Value exceeds integer range.");
return false;
}
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
Provide a HINT listing the allowed unit names when a GUC variable seems to contain a wrong unit specification, per discussion. In passing, fix the code to avoid unnecessary integer overflows when converting units, and to detect overflows when they do occur.
2007-06-21 20:14:21 +02:00
/* allow whitespace between integer and unit */
while (isspace((unsigned char) *endptr))
endptr++;
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
Provide a HINT listing the allowed unit names when a GUC variable seems to contain a wrong unit specification, per discussion. In passing, fix the code to avoid unnecessary integer overflows when converting units, and to detect overflows when they do occur.
2007-06-21 20:14:21 +02:00
/* Handle possible unit */
if (*endptr != '\0')
{
/*
* Note: the multiple-switch coding technique here is a bit tedious,
* but seems necessary to avoid intermediate-value overflows.
*/
if (flags & GUC_UNIT_MEMORY)
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
{
Provide a HINT listing the allowed unit names when a GUC variable seems to contain a wrong unit specification, per discussion. In passing, fix the code to avoid unnecessary integer overflows when converting units, and to detect overflows when they do occur.
2007-06-21 20:14:21 +02:00
/* Set hint for use if no match or trailing garbage */
if (hintmsg)
*hintmsg = gettext_noop("Valid units for this parameter are \"kB\", \"MB\", and \"GB\".");
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
Provide a HINT listing the allowed unit names when a GUC variable seems to contain a wrong unit specification, per discussion. In passing, fix the code to avoid unnecessary integer overflows when converting units, and to detect overflows when they do occur.
2007-06-21 20:14:21 +02:00
#if BLCKSZ < 1024 || BLCKSZ > (1024*1024)
#error BLCKSZ must be between 1KB and 1MB
#endif
#if XLOG_BLCKSZ < 1024 || XLOG_BLCKSZ > (1024*1024)
#error XLOG_BLCKSZ must be between 1KB and 1MB
Move BLCKSZ < 1024 check to guc.c.
2007-02-23 22:36:19 +01:00
#endif
Provide a HINT listing the allowed unit names when a GUC variable seems to contain a wrong unit specification, per discussion. In passing, fix the code to avoid unnecessary integer overflows when converting units, and to detect overflows when they do occur.
2007-06-21 20:14:21 +02:00
if (strncmp(endptr, "kB", 2) == 0)
The attached patch changes units of the some default values in postgresql.conf. - shared_buffers = 32000kB => 32MB - temp_buffers = 8000kB => 8MB - wal_buffers = 8 => 64kB The code of initdb was a bit modified to write MB-unit values. Values greater than 8000kB are rounded out to MB. GUC_UNIT_XBLOCKS is added for wal_buffers. It is like GUC_UNIT_BLOCKS, but uses XLOG_BLCKSZ instead of BLCKSZ. Also, I cleaned up the test of GUC_UNIT_* flags in preparation to add more unit flags in less bits. ITAGAKI Takahiro
2006-10-03 23:11:55 +02:00
{
Provide a HINT listing the allowed unit names when a GUC variable seems to contain a wrong unit specification, per discussion. In passing, fix the code to avoid unnecessary integer overflows when converting units, and to detect overflows when they do occur.
2007-06-21 20:14:21 +02:00
endptr += 2;
switch (flags & GUC_UNIT_MEMORY)
{
case GUC_UNIT_BLOCKS:
val /= (BLCKSZ / 1024);
break;
case GUC_UNIT_XBLOCKS:
val /= (XLOG_BLCKSZ / 1024);
break;
}
}
else if (strncmp(endptr, "MB", 2) == 0)
{
endptr += 2;
switch (flags & GUC_UNIT_MEMORY)
{
case GUC_UNIT_KB:
val *= KB_PER_MB;
break;
case GUC_UNIT_BLOCKS:
val *= KB_PER_MB / (BLCKSZ / 1024);
break;
case GUC_UNIT_XBLOCKS:
val *= KB_PER_MB / (XLOG_BLCKSZ / 1024);
break;
}
}
else if (strncmp(endptr, "GB", 2) == 0)
{
endptr += 2;
switch (flags & GUC_UNIT_MEMORY)
{
case GUC_UNIT_KB:
val *= KB_PER_GB;
break;
case GUC_UNIT_BLOCKS:
val *= KB_PER_GB / (BLCKSZ / 1024);
break;
case GUC_UNIT_XBLOCKS:
val *= KB_PER_GB / (XLOG_BLCKSZ / 1024);
break;
}
The attached patch changes units of the some default values in postgresql.conf. - shared_buffers = 32000kB => 32MB - temp_buffers = 8000kB => 8MB - wal_buffers = 8 => 64kB The code of initdb was a bit modified to write MB-unit values. Values greater than 8000kB are rounded out to MB. GUC_UNIT_XBLOCKS is added for wal_buffers. It is like GUC_UNIT_BLOCKS, but uses XLOG_BLCKSZ instead of BLCKSZ. Also, I cleaned up the test of GUC_UNIT_* flags in preparation to add more unit flags in less bits. ITAGAKI Takahiro
2006-10-03 23:11:55 +02:00
}
}
Provide a HINT listing the allowed unit names when a GUC variable seems to contain a wrong unit specification, per discussion. In passing, fix the code to avoid unnecessary integer overflows when converting units, and to detect overflows when they do occur.
2007-06-21 20:14:21 +02:00
else if (flags & GUC_UNIT_TIME)
{
/* Set hint for use if no match or trailing garbage */
if (hintmsg)
*hintmsg = gettext_noop("Valid units for this parameter are \"ms\", \"s\", \"min\", \"h\", and \"d\".");
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
Provide a HINT listing the allowed unit names when a GUC variable seems to contain a wrong unit specification, per discussion. In passing, fix the code to avoid unnecessary integer overflows when converting units, and to detect overflows when they do occur.
2007-06-21 20:14:21 +02:00
if (strncmp(endptr, "ms", 2) == 0)
{
endptr += 2;
switch (flags & GUC_UNIT_TIME)
{
case GUC_UNIT_S:
val /= MS_PER_S;
break;
case GUC_UNIT_MIN:
val /= MS_PER_MIN;
break;
}
}
else if (strncmp(endptr, "s", 1) == 0)
{
endptr += 1;
switch (flags & GUC_UNIT_TIME)
{
case GUC_UNIT_MS:
val *= MS_PER_S;
break;
case GUC_UNIT_MIN:
val /= S_PER_MIN;
break;
}
}
else if (strncmp(endptr, "min", 3) == 0)
{
endptr += 3;
switch (flags & GUC_UNIT_TIME)
{
case GUC_UNIT_MS:
val *= MS_PER_MIN;
break;
case GUC_UNIT_S:
val *= S_PER_MIN;
break;
}
}
else if (strncmp(endptr, "h", 1) == 0)
{
endptr += 1;
switch (flags & GUC_UNIT_TIME)
{
case GUC_UNIT_MS:
val *= MS_PER_H;
break;
case GUC_UNIT_S:
val *= S_PER_H;
break;
case GUC_UNIT_MIN:
val *= MIN_PER_H;
break;
}
}
else if (strncmp(endptr, "d", 1) == 0)
{
endptr += 1;
switch (flags & GUC_UNIT_TIME)
{
case GUC_UNIT_MS:
val *= MS_PER_D;
break;
case GUC_UNIT_S:
val *= S_PER_D;
break;
case GUC_UNIT_MIN:
val *= MIN_PER_D;
break;
}
}
}
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
Provide a HINT listing the allowed unit names when a GUC variable seems to contain a wrong unit specification, per discussion. In passing, fix the code to avoid unnecessary integer overflows when converting units, and to detect overflows when they do occur.
2007-06-21 20:14:21 +02:00
/* allow whitespace after unit */
while (isspace((unsigned char) *endptr))
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
endptr++;
Provide a HINT listing the allowed unit names when a GUC variable seems to contain a wrong unit specification, per discussion. In passing, fix the code to avoid unnecessary integer overflows when converting units, and to detect overflows when they do occur.
2007-06-21 20:14:21 +02:00
if (*endptr != '\0')
return false; /* appropriate hint, if any, already set */
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
Provide a HINT listing the allowed unit names when a GUC variable seems to contain a wrong unit specification, per discussion. In passing, fix the code to avoid unnecessary integer overflows when converting units, and to detect overflows when they do occur.
2007-06-21 20:14:21 +02:00
/* Check for overflow due to units conversion */
if (val != (int64) ((int32) val))
The attached patch changes units of the some default values in postgresql.conf. - shared_buffers = 32000kB => 32MB - temp_buffers = 8000kB => 8MB - wal_buffers = 8 => 64kB The code of initdb was a bit modified to write MB-unit values. Values greater than 8000kB are rounded out to MB. GUC_UNIT_XBLOCKS is added for wal_buffers. It is like GUC_UNIT_BLOCKS, but uses XLOG_BLCKSZ instead of BLCKSZ. Also, I cleaned up the test of GUC_UNIT_* flags in preparation to add more unit flags in less bits. ITAGAKI Takahiro
2006-10-03 23:11:55 +02:00
{
Provide a HINT listing the allowed unit names when a GUC variable seems to contain a wrong unit specification, per discussion. In passing, fix the code to avoid unnecessary integer overflows when converting units, and to detect overflows when they do occur.
2007-06-21 20:14:21 +02:00
if (hintmsg)
*hintmsg = gettext_noop("Value exceeds integer range.");
return false;
The attached patch changes units of the some default values in postgresql.conf. - shared_buffers = 32000kB => 32MB - temp_buffers = 8000kB => 8MB - wal_buffers = 8 => 64kB The code of initdb was a bit modified to write MB-unit values. Values greater than 8000kB are rounded out to MB. GUC_UNIT_XBLOCKS is added for wal_buffers. It is like GUC_UNIT_BLOCKS, but uses XLOG_BLCKSZ instead of BLCKSZ. Also, I cleaned up the test of GUC_UNIT_* flags in preparation to add more unit flags in less bits. ITAGAKI Takahiro
2006-10-03 23:11:55 +02:00
}
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
}
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
if (result)
*result = (int) val;
return true;
}
/*
Allow trailing whitespace in parse_real(), for consistency with parse_int() and with itself (strtod allows leading whitespace, so it seems odd not to allow trailing whitespace). parse_bool remains not-whitespace-friendly, but this is generically true for non-numeric GUC variables, so I'll desist from changing it.
2007-06-22 00:59:12 +02:00
* Try to parse value as a floating point number in the usual format.
* If the string parses okay, return true, else false.
* If okay and result is not NULL, return the value in *result.
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
*/
Use guc.c's parse_int() instead of pg_atoi() to parse fillfactor in default_reloptions(). The previous coding was really a bug because pg_atoi() will always throw elog on bad input data, whereas default_reloptions is not supposed to complain about bad input unless its validate parameter is true. Right now you could only expose the problem by hand-modifying pg_class.reloptions into an invalid state, so it doesn't seem worth back-patching; but we should get it right in HEAD because there might be other situations in future. Noted while studying GIN fast-update patch.
2008-07-23 19:29:53 +02:00
bool
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
parse_real(const char *value, double *result)
{
double val;
char *endptr;
Allow trailing whitespace in parse_real(), for consistency with parse_int() and with itself (strtod allows leading whitespace, so it seems odd not to allow trailing whitespace). parse_bool remains not-whitespace-friendly, but this is generically true for non-numeric GUC variables, so I'll desist from changing it.
2007-06-22 00:59:12 +02:00
if (result)
*result = 0; /* suppress compiler warning */
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
errno = 0;
val = strtod(value, &endptr);
Allow trailing whitespace in parse_real(), for consistency with parse_int() and with itself (strtod allows leading whitespace, so it seems odd not to allow trailing whitespace). parse_bool remains not-whitespace-friendly, but this is generically true for non-numeric GUC variables, so I'll desist from changing it.
2007-06-22 00:59:12 +02:00
if (endptr == value || errno == ERANGE)
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
return false;
Allow trailing whitespace in parse_real(), for consistency with parse_int() and with itself (strtod allows leading whitespace, so it seems odd not to allow trailing whitespace). parse_bool remains not-whitespace-friendly, but this is generically true for non-numeric GUC variables, so I'll desist from changing it.
2007-06-22 00:59:12 +02:00
/* allow whitespace after number */
while (isspace((unsigned char) *endptr))
endptr++;
if (*endptr != '\0')
return false;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
if (result)
*result = val;
return true;
}
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
/*
* Lookup the name for an enum option with the selected value.
* Should only ever be called with known-valid values, so throws
* an elog(ERROR) if the enum option is not found.
*
* The returned string is a pointer to static data and not
* allocated for modification.
*/
Fix postgres --describe-config for guc enums, breakage noted by Alvaro. While at it, rename option lookup functions to make names clearer, per discussion with Tom.
2008-03-17 18:45:09 +01:00
const char *
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
config_enum_lookup_by_value(struct config_enum * record, int val)
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
{
Minor code embellishments.
2008-12-02 03:00:32 +01:00
const struct config_enum_entry *entry;
for (entry = record->options; entry && entry->name; entry++)
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
{
if (entry->val == val)
return entry->name;
}
Minor code embellishments.
2008-12-02 03:00:32 +01:00
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
elog(ERROR, "could not find enum option %d for %s",
val, record->gen.name);
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
return NULL; /* silence compiler */
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
}
/*
* Lookup the value for an enum option with the selected name
* (case-insensitive).
* If the enum option is found, sets the retval value and returns
Some cleanups of enum-guc code, per comments from Tom.
2008-03-16 17:42:44 +01:00
* true. If it's not found, return FALSE and retval is set to 0.
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
*/
Fix postgres --describe-config for guc enums, breakage noted by Alvaro. While at it, rename option lookup functions to make names clearer, per discussion with Tom.
2008-03-17 18:45:09 +01:00
bool
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
config_enum_lookup_by_name(struct config_enum * record, const char *value,
Minor code embellishments.
2008-12-02 03:00:32 +01:00
int *retval)
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
{
Minor code embellishments.
2008-12-02 03:00:32 +01:00
const struct config_enum_entry *entry;
for (entry = record->options; entry && entry->name; entry++)
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
{
Some cleanups of enum-guc code, per comments from Tom.
2008-03-16 17:42:44 +01:00
if (pg_strcasecmp(value, entry->name) == 0)
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
{
*retval = entry->val;
return TRUE;
}
}
Minor code embellishments.
2008-12-02 03:00:32 +01:00
*retval = 0;
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
return FALSE;
}
/*
Add a field to guc enums to allow hiding of values from display while still accepting them as input, used to allow alternate syntax for the same setting. Alex Hunsaker
2008-05-28 11:04:06 +02:00
* Return a list of all available options for an enum, excluding
Minor code embellishments.
2008-12-02 03:00:32 +01:00
* hidden ones, separated by the given separator.
Some cleanups of enum-guc code, per comments from Tom.
2008-03-16 17:42:44 +01:00
* If prefix is non-NULL, it is added before the first enum value.
* If suffix is non-NULL, it is added to the end of the string.
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
*/
static char *
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
config_enum_get_options(struct config_enum * record, const char *prefix,
Make the enumvals column of pg_settings be text[] instead of just a comma separated string.
2008-11-21 19:49:24 +01:00
const char *suffix, const char *separator)
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
{
Minor code embellishments.
2008-12-02 03:00:32 +01:00
const struct config_enum_entry *entry;
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
StringInfoData retstr;
Minor code embellishments.
2008-12-02 03:00:32 +01:00
int seplen;
Add a field to guc enums to allow hiding of values from display while still accepting them as input, used to allow alternate syntax for the same setting. Alex Hunsaker
2008-05-28 11:04:06 +02:00
Minor code embellishments.
2008-12-02 03:00:32 +01:00
initStringInfo(&retstr);
appendStringInfoString(&retstr, prefix);
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
Minor code embellishments.
2008-12-02 03:00:32 +01:00
seplen = strlen(separator);
for (entry = record->options; entry && entry->name; entry++)
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
{
Add a field to guc enums to allow hiding of values from display while still accepting them as input, used to allow alternate syntax for the same setting. Alex Hunsaker
2008-05-28 11:04:06 +02:00
if (!entry->hidden)
{
Minor code embellishments.
2008-12-02 03:00:32 +01:00
appendStringInfoString(&retstr, entry->name);
appendBinaryStringInfo(&retstr, separator, seplen);
Add a field to guc enums to allow hiding of values from display while still accepting them as input, used to allow alternate syntax for the same setting. Alex Hunsaker
2008-05-28 11:04:06 +02:00
}
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
}
Add a field to guc enums to allow hiding of values from display while still accepting them as input, used to allow alternate syntax for the same setting. Alex Hunsaker
2008-05-28 11:04:06 +02:00
/*
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
* All the entries may have been hidden, leaving the string empty if no
* prefix was given. This indicates a broken GUC setup, since there is no
* use for an enum without any values, so we just check to make sure we
* don't write to invalid memory instead of actually trying to do
* something smart with it.
Add a field to guc enums to allow hiding of values from display while still accepting them as input, used to allow alternate syntax for the same setting. Alex Hunsaker
2008-05-28 11:04:06 +02:00
*/
Minor code embellishments.
2008-12-02 03:00:32 +01:00
if (retstr.len >= seplen)
{
Make the enumvals column of pg_settings be text[] instead of just a comma separated string.
2008-11-21 19:49:24 +01:00
/* Replace final separator */
Minor code embellishments.
2008-12-02 03:00:32 +01:00
retstr.data[retstr.len - seplen] = '\0';
retstr.len -= seplen;
}
Add a field to guc enums to allow hiding of values from display while still accepting them as input, used to allow alternate syntax for the same setting. Alex Hunsaker
2008-05-28 11:04:06 +02:00
Minor code embellishments.
2008-12-02 03:00:32 +01:00
appendStringInfoString(&retstr, suffix);
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
Minor code embellishments.
2008-12-02 03:00:32 +01:00
return retstr.data;
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
}
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
/*
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
* Sets option `name' to given value.
*
* The value should be a string, which will be parsed and converted to
* the appropriate data type. The context and source parameters indicate
* in which context this function is being called, so that it can apply the
* access restrictions properly.
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
*
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
* If value is NULL, set the option to its default value (normally the
* reset_val, but if source == PGC_S_DEFAULT we instead use the boot_val).
*
* action indicates whether to set the value globally in the session, locally
* to the current top transaction, or just for the duration of a function call.
*
* If changeVal is false then don't really set the option but do all
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
* the checks to see if it would work.
*
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
* Return value:
* +1: the value is valid and was successfully applied.
* 0: the name or value is invalid (but see below).
* -1: the value was not applied because of context, priority, or changeVal.
*
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
* If there is an error (non-existing option, invalid value) then an
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
* ereport(ERROR) is thrown *unless* this is called for a source for which
* we don't want an ERROR (currently, those are defaults, the config file,
* and per-database or per-user settings). In those cases we write a
* suitable error message via ereport() and return 0.
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
*
* See also SetConfigOption for an external interface.
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
*/
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
int
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
set_config_option(const char *name, const char *value,
GucContext context, GucSource source,
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
GucAction action, bool changeVal)
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
{
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
struct config_generic *record;
int elevel;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
bool prohibitValueChange = false;
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
bool makeDefault;
Re-apply guc cleanup patch, with memory allocation bugs fixed.
2006-08-13 03:30:17 +02:00
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
if (source == PGC_S_DEFAULT || source == PGC_S_FILE)
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
{
/*
* To avoid cluttering the log, only the postmaster bleats loudly
* about problems with the config file.
*/
Improve consistency of error reporting in GUC assign_hook routines. Some were reporting ERROR for interactive assignments and LOG for other cases, some were saying nothing for non-interactive cases, and a few did yet other things. Make them use a new function GUC_complaint_elevel() to establish a reasonably uniform policy about how to report. There are still a few edge cases such as assign_search_path(), but it's much better than before. Per gripe from Devrim Gunduz and subsequent discussion. As noted by Alvaro, it'd be better to fold these custom messages into the standard "invalid parameter value" complaint from guc.c, perhaps as the DETAIL field. However that will require more redesign than seems prudent for 8.3. This is a relatively safe, low-impact change that we can afford to risk now.
2007-12-28 01:23:23 +01:00
elevel = IsUnderPostmaster ? DEBUG3 : LOG;
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
}
Add new PGC_S_DATABASE_USER enum value to several places missed by my patch last week. Per note and patch from Jeff Davis.
2009-10-13 16:18:40 +02:00
else if (source == PGC_S_DATABASE || source == PGC_S_USER ||
source == PGC_S_DATABASE_USER)
Change a couple of ill-advised uses of INFO elog level to WARNINGs; in particular this allows EmitWarningsOnPlaceholders messages to show up in the postmaster log by default. Update elog.h comment to make it clearer what INFO is for, and fix one example in the SGML docs that was misusing it. Per my gripe of yesterday.
2009-01-06 17:39:52 +01:00
elevel = WARNING;
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
else
elevel = ERROR;
Re-apply guc cleanup patch, with memory allocation bugs fixed.
2006-08-13 03:30:17 +02:00
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
record = find_option(name, true, elevel);
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
if (record == NULL)
{
ereport(elevel,
(errcode(ERRCODE_UNDEFINED_OBJECT),
errmsg("unrecognized configuration parameter \"%s\"", name)));
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
return 0;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
}
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
/*
* Check if the option can be set at this time. See guc.h for the precise
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
* rules.
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
*/
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
switch (record->context)
{
In the continuing saga of FE/BE protocol revisions, add reporting of initial values and runtime changes in selected parameters. This gets rid of the need for an initial 'select pg_client_encoding()' query in libpq, bringing us back to one message transmitted in each direction for a standard connection startup. To allow server version to be sent using the same GUC mechanism that handles other parameters, invent the concept of a never-settable GUC parameter: you can 'show server_version' but it's not settable by any GUC input source. Create 'lc_collate' and 'lc_ctype' never-settable parameters so that people can find out these settings without need for pg_controldata. (These side ideas were all discussed some time ago in pgsql-hackers, but not yet implemented.)
2003-04-25 21:45:10 +02:00
case PGC_INTERNAL:
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
if (context != PGC_INTERNAL)
In the continuing saga of FE/BE protocol revisions, add reporting of initial values and runtime changes in selected parameters. This gets rid of the need for an initial 'select pg_client_encoding()' query in libpq, bringing us back to one message transmitted in each direction for a standard connection startup. To allow server version to be sent using the same GUC mechanism that handles other parameters, invent the concept of a never-settable GUC parameter: you can 'show server_version' but it's not settable by any GUC input source. Create 'lc_collate' and 'lc_ctype' never-settable parameters so that people can find out these settings without need for pg_controldata. (These side ideas were all discussed some time ago in pgsql-hackers, but not yet implemented.)
2003-04-25 21:45:10 +02:00
{
Error message editing in backend/utils (except /adt).
2003-07-25 22:18:01 +02:00
ereport(elevel,
(errcode(ERRCODE_CANT_CHANGE_RUNTIME_PARAM),
Message editing: remove gratuitous variations in message wording, standardize terms, add some clarifications, fix some untranslatable attempts at dynamic message building.
2003-09-25 08:58:07 +02:00
errmsg("parameter \"%s\" cannot be changed",
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
name)));
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
return 0;
In the continuing saga of FE/BE protocol revisions, add reporting of initial values and runtime changes in selected parameters. This gets rid of the need for an initial 'select pg_client_encoding()' query in libpq, bringing us back to one message transmitted in each direction for a standard connection startup. To allow server version to be sent using the same GUC mechanism that handles other parameters, invent the concept of a never-settable GUC parameter: you can 'show server_version' but it's not settable by any GUC input source. Create 'lc_collate' and 'lc_ctype' never-settable parameters so that people can find out these settings without need for pg_controldata. (These side ideas were all discussed some time ago in pgsql-hackers, but not yet implemented.)
2003-04-25 21:45:10 +02:00
}
break;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
case PGC_POSTMASTER:
if (context == PGC_SIGHUP)
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
{
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
/*
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
* We are re-reading a PGC_POSTMASTER variable from
* postgresql.conf. We can't change the setting, so we should
* give a warning if the DBA tries to change it. However,
* because of variant formats, canonicalization by check
* hooks, etc, we can't just compare the given string directly
* to what's stored. Set a flag to check below after we have
* the final storable value.
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
*/
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
prohibitValueChange = true;
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
}
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
else if (context != PGC_POSTMASTER)
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
{
Error message editing in backend/utils (except /adt).
2003-07-25 22:18:01 +02:00
ereport(elevel,
(errcode(ERRCODE_CANT_CHANGE_RUNTIME_PARAM),
pgindent run for 9.0
2010-02-26 03:01:40 +01:00
errmsg("parameter \"%s\" cannot be changed without restarting the server",
name)));
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
return 0;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
}
break;
case PGC_SIGHUP:
if (context != PGC_SIGHUP && context != PGC_POSTMASTER)
{
Error message editing in backend/utils (except /adt).
2003-07-25 22:18:01 +02:00
ereport(elevel,
(errcode(ERRCODE_CANT_CHANGE_RUNTIME_PARAM),
Message editing: remove gratuitous variations in message wording, standardize terms, add some clarifications, fix some untranslatable attempts at dynamic message building.
2003-09-25 08:58:07 +02:00
errmsg("parameter \"%s\" cannot be changed now",
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
name)));
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
return 0;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
}
/*
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
* Hmm, the idea of the SIGHUP context is "ought to be global, but
* can be changed after postmaster start". But there's nothing
* that prevents a crafty administrator from sending SIGHUP
* signals to individual backends only.
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
*/
break;
case PGC_BACKEND:
if (context == PGC_SIGHUP)
{
/*
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
* If a PGC_BACKEND parameter is changed in the config file,
* we want to accept the new value in the postmaster (whence
* it will propagate to subsequently-started backends), but
* ignore it in existing backends. This is a tad klugy, but
* necessary because we don't re-read the config file during
* backend start.
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
*/
if (IsUnderPostmaster)
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
return -1;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
}
Remove the use of the pg_auth flat file for client authentication. (That flat file is now completely useless, but removal will come later.) To do this, postpone client authentication into the startup transaction that's run by InitPostgres. We still collect the startup packet and do SSL initialization (if needed) at the same time we did before. The AuthenticationTimeout is applied separately to startup packet collection and the actual authentication cycle. (This is a bit annoying, since it means a couple extra syscalls; but the signal handling requirements inside and outside a transaction are sufficiently different that it seems best to treat the timeouts as completely independent.) A small security disadvantage is that if the given database name is invalid, this will be reported to the client before any authentication happens. We could work around that by connecting to database "postgres" instead, but consensus seems to be that it's not worth introducing such surprising behavior. Processing of all command-line switches and GUC options received from the client is now postponed until after authentication. This means that PostAuthDelay is much less useful than it used to be --- if you need to investigate problems during InitPostgres you'll have to set PreAuthDelay instead. However, allowing an unauthenticated user to set any GUC options whatever seems a bit too risky, so we'll live with that.
2009-08-29 21:26:52 +02:00
else if (context != PGC_POSTMASTER && context != PGC_BACKEND &&
source != PGC_S_CLIENT)
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
{
Error message editing in backend/utils (except /adt).
2003-07-25 22:18:01 +02:00
ereport(elevel,
(errcode(ERRCODE_CANT_CHANGE_RUNTIME_PARAM),
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
errmsg("parameter \"%s\" cannot be set after connection start",
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
name)));
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
return 0;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
}
break;
case PGC_SUSET:
if (context == PGC_USERSET || context == PGC_BACKEND)
{
Error message editing in backend/utils (except /adt).
2003-07-25 22:18:01 +02:00
ereport(elevel,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
errmsg("permission denied to set parameter \"%s\"",
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
name)));
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
return 0;
Allow the postmaster to accept changes in PGC_BACKEND GUC variables from the config file, so that these changes will propagate to backends started later. Already-started backends continue to ignore changes in these variables.
2001-09-30 22:16:21 +02:00
}
break;
case PGC_USERSET:
/* always okay */
break;
Second pass over run-time configuration system. Adjust priorities on some option settings. Sort out SIGHUP vs BACKEND -- there is no total ordering here, so make explicit checks. Add comments explaining all of this. Removed permissions check on SHOW command. Add examine_subclass to the game, rename to SQL_inheritance to fit the official data model better. Adjust documentation. Standalone backend needs to reset all options before it starts. To facilitate that, have IsUnderPostmaster be set by the postmaster itself, don't wait for the magic -p switch. Also make sure that all environment variables and argv's survive init_ps_display(). Use strdup where necessary. Have initdb make configuration files (postgresql.conf, pg_hba.conf) mode 0600 -- having configuration files is no fun if you can't edit them.
2000-06-23 00:31:24 +02:00
}
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
Disallow RESET ROLE and RESET SESSION AUTHORIZATION inside security-definer functions. This extends the previous patch that forbade SETting these variables inside security-definer functions. RESET is equally a security hole, since it would allow regaining privileges of the caller; furthermore it can trigger Assert failures and perhaps other internal errors, since the code is not expecting these variables to change in such contexts. The previous patch did not cover this case because assign hooks don't really have enough information, so move the responsibility for preventing this into guc.c. Problem discovered by Heikki Linnakangas. Security: no CVE assigned yet, extends CVE-2007-6600
2009-09-04 00:08:05 +02:00
/*
Prevent indirect security attacks via changing session-local state within an allegedly immutable index function. It was previously recognized that we had to prevent such a function from executing SET/RESET ROLE/SESSION AUTHORIZATION, or it could trivially obtain the privileges of the session user. However, since there is in general no privilege checking for changes of session-local state, it is also possible for such a function to change settings in a way that might subvert later operations in the same session. Examples include changing search_path to cause an unexpected function to be called, or replacing an existing prepared statement with another one that will execute a function of the attacker's choosing. The present patch secures VACUUM, ANALYZE, and CREATE INDEX/REINDEX against these threats, which are the same places previously deemed to need protection against the SET ROLE issue. GUC changes are still allowed, since there are many useful cases for that, but we prevent security problems by forcing a rollback of any GUC change after completing the operation. Other cases are handled by throwing an error if any change is attempted; these include temp table creation, closing a cursor, and creating or deleting a prepared statement. (In 7.4, the infrastructure to roll back GUC changes doesn't exist, so we settle for rejecting changes of "search_path" in these contexts.) Original report and patch by Gurjeet Singh, additional analysis by Tom Lane. Security: CVE-2009-4136
2009-12-09 22:57:51 +01:00
* Disallow changing GUC_NOT_WHILE_SEC_REST values if we are inside a
pgindent run for 9.0
2010-02-26 03:01:40 +01:00
* security restriction context. We can reject this regardless of the GUC
* context or source, mainly because sources that it might be reasonable
* to override for won't be seen while inside a function.
Disallow RESET ROLE and RESET SESSION AUTHORIZATION inside security-definer functions. This extends the previous patch that forbade SETting these variables inside security-definer functions. RESET is equally a security hole, since it would allow regaining privileges of the caller; furthermore it can trigger Assert failures and perhaps other internal errors, since the code is not expecting these variables to change in such contexts. The previous patch did not cover this case because assign hooks don't really have enough information, so move the responsibility for preventing this into guc.c. Problem discovered by Heikki Linnakangas. Security: no CVE assigned yet, extends CVE-2007-6600
2009-09-04 00:08:05 +02:00
*
Prevent indirect security attacks via changing session-local state within an allegedly immutable index function. It was previously recognized that we had to prevent such a function from executing SET/RESET ROLE/SESSION AUTHORIZATION, or it could trivially obtain the privileges of the session user. However, since there is in general no privilege checking for changes of session-local state, it is also possible for such a function to change settings in a way that might subvert later operations in the same session. Examples include changing search_path to cause an unexpected function to be called, or replacing an existing prepared statement with another one that will execute a function of the attacker's choosing. The present patch secures VACUUM, ANALYZE, and CREATE INDEX/REINDEX against these threats, which are the same places previously deemed to need protection against the SET ROLE issue. GUC changes are still allowed, since there are many useful cases for that, but we prevent security problems by forcing a rollback of any GUC change after completing the operation. Other cases are handled by throwing an error if any change is attempted; these include temp table creation, closing a cursor, and creating or deleting a prepared statement. (In 7.4, the infrastructure to roll back GUC changes doesn't exist, so we settle for rejecting changes of "search_path" in these contexts.) Original report and patch by Gurjeet Singh, additional analysis by Tom Lane. Security: CVE-2009-4136
2009-12-09 22:57:51 +01:00
* Note: variables marked GUC_NOT_WHILE_SEC_REST should usually be marked
Disallow RESET ROLE and RESET SESSION AUTHORIZATION inside security-definer functions. This extends the previous patch that forbade SETting these variables inside security-definer functions. RESET is equally a security hole, since it would allow regaining privileges of the caller; furthermore it can trigger Assert failures and perhaps other internal errors, since the code is not expecting these variables to change in such contexts. The previous patch did not cover this case because assign hooks don't really have enough information, so move the responsibility for preventing this into guc.c. Problem discovered by Heikki Linnakangas. Security: no CVE assigned yet, extends CVE-2007-6600
2009-09-04 00:08:05 +02:00
* GUC_NO_RESET_ALL as well, because ResetAllOptions() doesn't check this.
Prevent indirect security attacks via changing session-local state within an allegedly immutable index function. It was previously recognized that we had to prevent such a function from executing SET/RESET ROLE/SESSION AUTHORIZATION, or it could trivially obtain the privileges of the session user. However, since there is in general no privilege checking for changes of session-local state, it is also possible for such a function to change settings in a way that might subvert later operations in the same session. Examples include changing search_path to cause an unexpected function to be called, or replacing an existing prepared statement with another one that will execute a function of the attacker's choosing. The present patch secures VACUUM, ANALYZE, and CREATE INDEX/REINDEX against these threats, which are the same places previously deemed to need protection against the SET ROLE issue. GUC changes are still allowed, since there are many useful cases for that, but we prevent security problems by forcing a rollback of any GUC change after completing the operation. Other cases are handled by throwing an error if any change is attempted; these include temp table creation, closing a cursor, and creating or deleting a prepared statement. (In 7.4, the infrastructure to roll back GUC changes doesn't exist, so we settle for rejecting changes of "search_path" in these contexts.) Original report and patch by Gurjeet Singh, additional analysis by Tom Lane. Security: CVE-2009-4136
2009-12-09 22:57:51 +01:00
* An exception might be made if the reset value is assumed to be "safe".
Disallow RESET ROLE and RESET SESSION AUTHORIZATION inside security-definer functions. This extends the previous patch that forbade SETting these variables inside security-definer functions. RESET is equally a security hole, since it would allow regaining privileges of the caller; furthermore it can trigger Assert failures and perhaps other internal errors, since the code is not expecting these variables to change in such contexts. The previous patch did not cover this case because assign hooks don't really have enough information, so move the responsibility for preventing this into guc.c. Problem discovered by Heikki Linnakangas. Security: no CVE assigned yet, extends CVE-2007-6600
2009-09-04 00:08:05 +02:00
*
* Note: this flag is currently used for "session_authorization" and
pgindent run for 9.0
2010-02-26 03:01:40 +01:00
* "role". We need to prohibit changing these inside a local userid
Prevent indirect security attacks via changing session-local state within an allegedly immutable index function. It was previously recognized that we had to prevent such a function from executing SET/RESET ROLE/SESSION AUTHORIZATION, or it could trivially obtain the privileges of the session user. However, since there is in general no privilege checking for changes of session-local state, it is also possible for such a function to change settings in a way that might subvert later operations in the same session. Examples include changing search_path to cause an unexpected function to be called, or replacing an existing prepared statement with another one that will execute a function of the attacker's choosing. The present patch secures VACUUM, ANALYZE, and CREATE INDEX/REINDEX against these threats, which are the same places previously deemed to need protection against the SET ROLE issue. GUC changes are still allowed, since there are many useful cases for that, but we prevent security problems by forcing a rollback of any GUC change after completing the operation. Other cases are handled by throwing an error if any change is attempted; these include temp table creation, closing a cursor, and creating or deleting a prepared statement. (In 7.4, the infrastructure to roll back GUC changes doesn't exist, so we settle for rejecting changes of "search_path" in these contexts.) Original report and patch by Gurjeet Singh, additional analysis by Tom Lane. Security: CVE-2009-4136
2009-12-09 22:57:51 +01:00
* context because when we exit it, GUC won't be notified, leaving things
* out of sync. (This could be fixed by forcing a new GUC nesting level,
pgindent run for 9.0
2010-02-26 03:01:40 +01:00
* but that would change behavior in possibly-undesirable ways.) Also, we
* prohibit changing these in a security-restricted operation because
Prevent indirect security attacks via changing session-local state within an allegedly immutable index function. It was previously recognized that we had to prevent such a function from executing SET/RESET ROLE/SESSION AUTHORIZATION, or it could trivially obtain the privileges of the session user. However, since there is in general no privilege checking for changes of session-local state, it is also possible for such a function to change settings in a way that might subvert later operations in the same session. Examples include changing search_path to cause an unexpected function to be called, or replacing an existing prepared statement with another one that will execute a function of the attacker's choosing. The present patch secures VACUUM, ANALYZE, and CREATE INDEX/REINDEX against these threats, which are the same places previously deemed to need protection against the SET ROLE issue. GUC changes are still allowed, since there are many useful cases for that, but we prevent security problems by forcing a rollback of any GUC change after completing the operation. Other cases are handled by throwing an error if any change is attempted; these include temp table creation, closing a cursor, and creating or deleting a prepared statement. (In 7.4, the infrastructure to roll back GUC changes doesn't exist, so we settle for rejecting changes of "search_path" in these contexts.) Original report and patch by Gurjeet Singh, additional analysis by Tom Lane. Security: CVE-2009-4136
2009-12-09 22:57:51 +01:00
* otherwise RESET could be used to regain the session user's privileges.
Disallow RESET ROLE and RESET SESSION AUTHORIZATION inside security-definer functions. This extends the previous patch that forbade SETting these variables inside security-definer functions. RESET is equally a security hole, since it would allow regaining privileges of the caller; furthermore it can trigger Assert failures and perhaps other internal errors, since the code is not expecting these variables to change in such contexts. The previous patch did not cover this case because assign hooks don't really have enough information, so move the responsibility for preventing this into guc.c. Problem discovered by Heikki Linnakangas. Security: no CVE assigned yet, extends CVE-2007-6600
2009-09-04 00:08:05 +02:00
*/
Prevent indirect security attacks via changing session-local state within an allegedly immutable index function. It was previously recognized that we had to prevent such a function from executing SET/RESET ROLE/SESSION AUTHORIZATION, or it could trivially obtain the privileges of the session user. However, since there is in general no privilege checking for changes of session-local state, it is also possible for such a function to change settings in a way that might subvert later operations in the same session. Examples include changing search_path to cause an unexpected function to be called, or replacing an existing prepared statement with another one that will execute a function of the attacker's choosing. The present patch secures VACUUM, ANALYZE, and CREATE INDEX/REINDEX against these threats, which are the same places previously deemed to need protection against the SET ROLE issue. GUC changes are still allowed, since there are many useful cases for that, but we prevent security problems by forcing a rollback of any GUC change after completing the operation. Other cases are handled by throwing an error if any change is attempted; these include temp table creation, closing a cursor, and creating or deleting a prepared statement. (In 7.4, the infrastructure to roll back GUC changes doesn't exist, so we settle for rejecting changes of "search_path" in these contexts.) Original report and patch by Gurjeet Singh, additional analysis by Tom Lane. Security: CVE-2009-4136
2009-12-09 22:57:51 +01:00
if (record->flags & GUC_NOT_WHILE_SEC_REST)
Disallow RESET ROLE and RESET SESSION AUTHORIZATION inside security-definer functions. This extends the previous patch that forbade SETting these variables inside security-definer functions. RESET is equally a security hole, since it would allow regaining privileges of the caller; furthermore it can trigger Assert failures and perhaps other internal errors, since the code is not expecting these variables to change in such contexts. The previous patch did not cover this case because assign hooks don't really have enough information, so move the responsibility for preventing this into guc.c. Problem discovered by Heikki Linnakangas. Security: no CVE assigned yet, extends CVE-2007-6600
2009-09-04 00:08:05 +02:00
{
Prevent indirect security attacks via changing session-local state within an allegedly immutable index function. It was previously recognized that we had to prevent such a function from executing SET/RESET ROLE/SESSION AUTHORIZATION, or it could trivially obtain the privileges of the session user. However, since there is in general no privilege checking for changes of session-local state, it is also possible for such a function to change settings in a way that might subvert later operations in the same session. Examples include changing search_path to cause an unexpected function to be called, or replacing an existing prepared statement with another one that will execute a function of the attacker's choosing. The present patch secures VACUUM, ANALYZE, and CREATE INDEX/REINDEX against these threats, which are the same places previously deemed to need protection against the SET ROLE issue. GUC changes are still allowed, since there are many useful cases for that, but we prevent security problems by forcing a rollback of any GUC change after completing the operation. Other cases are handled by throwing an error if any change is attempted; these include temp table creation, closing a cursor, and creating or deleting a prepared statement. (In 7.4, the infrastructure to roll back GUC changes doesn't exist, so we settle for rejecting changes of "search_path" in these contexts.) Original report and patch by Gurjeet Singh, additional analysis by Tom Lane. Security: CVE-2009-4136
2009-12-09 22:57:51 +01:00
if (InLocalUserIdChange())
{
/*
pgindent run for 9.0
2010-02-26 03:01:40 +01:00
* Phrasing of this error message is historical, but it's the most
* common case.
Prevent indirect security attacks via changing session-local state within an allegedly immutable index function. It was previously recognized that we had to prevent such a function from executing SET/RESET ROLE/SESSION AUTHORIZATION, or it could trivially obtain the privileges of the session user. However, since there is in general no privilege checking for changes of session-local state, it is also possible for such a function to change settings in a way that might subvert later operations in the same session. Examples include changing search_path to cause an unexpected function to be called, or replacing an existing prepared statement with another one that will execute a function of the attacker's choosing. The present patch secures VACUUM, ANALYZE, and CREATE INDEX/REINDEX against these threats, which are the same places previously deemed to need protection against the SET ROLE issue. GUC changes are still allowed, since there are many useful cases for that, but we prevent security problems by forcing a rollback of any GUC change after completing the operation. Other cases are handled by throwing an error if any change is attempted; these include temp table creation, closing a cursor, and creating or deleting a prepared statement. (In 7.4, the infrastructure to roll back GUC changes doesn't exist, so we settle for rejecting changes of "search_path" in these contexts.) Original report and patch by Gurjeet Singh, additional analysis by Tom Lane. Security: CVE-2009-4136
2009-12-09 22:57:51 +01:00
*/
ereport(elevel,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
errmsg("cannot set parameter \"%s\" within security-definer function",
name)));
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
return 0;
Prevent indirect security attacks via changing session-local state within an allegedly immutable index function. It was previously recognized that we had to prevent such a function from executing SET/RESET ROLE/SESSION AUTHORIZATION, or it could trivially obtain the privileges of the session user. However, since there is in general no privilege checking for changes of session-local state, it is also possible for such a function to change settings in a way that might subvert later operations in the same session. Examples include changing search_path to cause an unexpected function to be called, or replacing an existing prepared statement with another one that will execute a function of the attacker's choosing. The present patch secures VACUUM, ANALYZE, and CREATE INDEX/REINDEX against these threats, which are the same places previously deemed to need protection against the SET ROLE issue. GUC changes are still allowed, since there are many useful cases for that, but we prevent security problems by forcing a rollback of any GUC change after completing the operation. Other cases are handled by throwing an error if any change is attempted; these include temp table creation, closing a cursor, and creating or deleting a prepared statement. (In 7.4, the infrastructure to roll back GUC changes doesn't exist, so we settle for rejecting changes of "search_path" in these contexts.) Original report and patch by Gurjeet Singh, additional analysis by Tom Lane. Security: CVE-2009-4136
2009-12-09 22:57:51 +01:00
}
if (InSecurityRestrictedOperation())
{
ereport(elevel,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
errmsg("cannot set parameter \"%s\" within security-restricted operation",
name)));
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
return 0;
Prevent indirect security attacks via changing session-local state within an allegedly immutable index function. It was previously recognized that we had to prevent such a function from executing SET/RESET ROLE/SESSION AUTHORIZATION, or it could trivially obtain the privileges of the session user. However, since there is in general no privilege checking for changes of session-local state, it is also possible for such a function to change settings in a way that might subvert later operations in the same session. Examples include changing search_path to cause an unexpected function to be called, or replacing an existing prepared statement with another one that will execute a function of the attacker's choosing. The present patch secures VACUUM, ANALYZE, and CREATE INDEX/REINDEX against these threats, which are the same places previously deemed to need protection against the SET ROLE issue. GUC changes are still allowed, since there are many useful cases for that, but we prevent security problems by forcing a rollback of any GUC change after completing the operation. Other cases are handled by throwing an error if any change is attempted; these include temp table creation, closing a cursor, and creating or deleting a prepared statement. (In 7.4, the infrastructure to roll back GUC changes doesn't exist, so we settle for rejecting changes of "search_path" in these contexts.) Original report and patch by Gurjeet Singh, additional analysis by Tom Lane. Security: CVE-2009-4136
2009-12-09 22:57:51 +01:00
}
Disallow RESET ROLE and RESET SESSION AUTHORIZATION inside security-definer functions. This extends the previous patch that forbade SETting these variables inside security-definer functions. RESET is equally a security hole, since it would allow regaining privileges of the caller; furthermore it can trigger Assert failures and perhaps other internal errors, since the code is not expecting these variables to change in such contexts. The previous patch did not cover this case because assign hooks don't really have enough information, so move the responsibility for preventing this into guc.c. Problem discovered by Heikki Linnakangas. Security: no CVE assigned yet, extends CVE-2007-6600
2009-09-04 00:08:05 +02:00
}
Dept. of second thoughts: interaction between DoIt and makeDepend in set_config_option wasn't quite right. Also clean up a couple other things that could have been done better.
2002-05-17 22:32:29 +02:00
/*
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
* Should we set reset/stacked values? (If so, the behavior is not
* transactional.) This is done either when we get a default value from
* the database's/user's/client's default settings or when we reset a
* value to its default.
Dept. of second thoughts: interaction between DoIt and makeDepend in set_config_option wasn't quite right. Also clean up a couple other things that could have been done better.
2002-05-17 22:32:29 +02:00
*/
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
makeDefault = changeVal && (source <= PGC_S_OVERRIDE) &&
((value != NULL) || source == PGC_S_DEFAULT);
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
/*
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
* Ignore attempted set if overridden by previously processed setting.
* However, if changeVal is false then plow ahead anyway since we are
* trying to find out if the value is potentially good, not actually use
* it. Also keep going if makeDefault is true, since we may want to set
* the reset/stacked values even if we can't set the variable itself.
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
*/
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
if (record->source > source)
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
{
Rename DoIt to changeVar, for clarity.
2003-09-01 06:15:51 +02:00
if (changeVal && !makeDefault)
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
{
Error message editing in backend/utils (except /adt).
2003-07-25 22:18:01 +02:00
elog(DEBUG3, "\"%s\": setting ignored because previous source is higher priority",
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
name);
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
return -1;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
}
Remove GUC USERLIMIT variable category, making the affected variables plain SUSET instead. Also delay processing of options received in client connection request until after we know if the user is a superuser, so that SUSET values can be set that way by legitimate superusers. Per recent discussion.
2004-11-14 20:35:35 +01:00
changeVal = false;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
}
Second pass over run-time configuration system. Adjust priorities on some option settings. Sort out SIGHUP vs BACKEND -- there is no total ordering here, so make explicit checks. Add comments explaining all of this. Removed permissions check on SHOW command. Add examine_subclass to the game, rename to SQL_inheritance to fit the official data model better. Adjust documentation. Standalone backend needs to reset all options before it starts. To facilitate that, have IsUnderPostmaster be set by the postmaster itself, don't wait for the magic -p switch. Also make sure that all environment variables and argv's survive init_ps_display(). Use strdup where necessary. Have initdb make configuration files (postgresql.conf, pg_hba.conf) mode 0600 -- having configuration files is no fun if you can't edit them.
2000-06-23 00:31:24 +02:00
/*
Remove GUC USERLIMIT variable category, making the affected variables plain SUSET instead. Also delay processing of options received in client connection request until after we know if the user is a superuser, so that SUSET values can be set that way by legitimate superusers. Per recent discussion.
2004-11-14 20:35:35 +01:00
* Evaluate value and set variable.
Second pass over run-time configuration system. Adjust priorities on some option settings. Sort out SIGHUP vs BACKEND -- there is no total ordering here, so make explicit checks. Add comments explaining all of this. Removed permissions check on SHOW command. Add examine_subclass to the game, rename to SQL_inheritance to fit the official data model better. Adjust documentation. Standalone backend needs to reset all options before it starts. To facilitate that, have IsUnderPostmaster be set by the postmaster itself, don't wait for the magic -p switch. Also make sure that all environment variables and argv's survive init_ps_display(). Use strdup where necessary. Have initdb make configuration files (postgresql.conf, pg_hba.conf) mode 0600 -- having configuration files is no fun if you can't edit them.
2000-06-23 00:31:24 +02:00
*/
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
switch (record->vartype)
Add all possible config file options.
2001-01-24 19:37:31 +01:00
{
case PGC_BOOL:
{
pgindent run. Make it all clean.
2001-03-22 05:01:46 +01:00
struct config_bool *conf = (struct config_bool *) record;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
bool newval;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
void *newextra = NULL;
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
if (value)
{
if (!parse_bool(value, &newval))
{
ereport(elevel,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("parameter \"%s\" requires a Boolean value",
name)));
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
return 0;
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
}
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
if (!call_bool_check_hook(conf, &newval, &newextra,
source, elevel))
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
return 0;
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
}
Make configuration parameters fall back to their default values when they are removed from the configuration file. Joachim Wieland
2007-04-21 22:02:41 +02:00
else if (source == PGC_S_DEFAULT)
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
{
Make configuration parameters fall back to their default values when they are removed from the configuration file. Joachim Wieland
2007-04-21 22:02:41 +02:00
newval = conf->boot_val;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
if (!call_bool_check_hook(conf, &newval, &newextra,
source, elevel))
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
return 0;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
}
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
else
{
newval = conf->reset_val;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
newextra = conf->reset_extra;
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
source = conf->gen.reset_source;
Remember the source GucContext for each GUC parameter. We used to just remember the GucSource, but saving GucContext too provides a little more information --- notably, whether a SET was done by a superuser or regular user. This allows us to rip out the fairly dodgy code that define_custom_variable used to use to try to infer the context to re-install a pre-existing setting with. In particular, it now works for a superuser to SET a extension's SUSET custom variable before loading the associated extension, because GUC can remember whether the SET was done as a superuser or not. The plperl regression tests contain an example where this is useful.
2011-10-04 22:13:16 +02:00
context = conf->gen.reset_scontext;
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
}
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
if (prohibitValueChange)
{
if (*conf->variable != newval)
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
{
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
ereport(elevel,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
(errcode(ERRCODE_CANT_CHANGE_RUNTIME_PARAM),
errmsg("parameter \"%s\" cannot be changed without restarting the server",
name)));
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
return 0;
}
return -1;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
}
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
if (changeVal)
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
/* Save old value to support transaction abort */
if (!makeDefault)
push_old_value(&conf->gen, action);
if (conf->assign_hook)
(*conf->assign_hook) (newval, newextra);
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
*conf->variable = newval;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
set_extra_field(&conf->gen, &conf->gen.extra,
newextra);
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
conf->gen.source = source;
Remember the source GucContext for each GUC parameter. We used to just remember the GucSource, but saving GucContext too provides a little more information --- notably, whether a SET was done by a superuser or regular user. This allows us to rip out the fairly dodgy code that define_custom_variable used to use to try to infer the context to re-install a pre-existing setting with. In particular, it now works for a superuser to SET a extension's SUSET custom variable before loading the associated extension, because GUC can remember whether the SET was done as a superuser or not. The plperl regression tests contain an example where this is useful.
2011-10-04 22:13:16 +02:00
conf->gen.scontext = context;
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
}
if (makeDefault)
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
{
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
GucStack *stack;
if (conf->gen.reset_source <= source)
reset all: command line and .conf options change defaults on RESET ALL those are restored. show all: GUC + non-GUC. SHOW ALL, RESET ALL Marko Kreen
2001-06-07 06:50:57 +02:00
{
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
conf->reset_val = newval;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
set_extra_field(&conf->gen, &conf->reset_extra,
newextra);
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
conf->gen.reset_source = source;
Remember the source GucContext for each GUC parameter. We used to just remember the GucSource, but saving GucContext too provides a little more information --- notably, whether a SET was done by a superuser or regular user. This allows us to rip out the fairly dodgy code that define_custom_variable used to use to try to infer the context to re-install a pre-existing setting with. In particular, it now works for a superuser to SET a extension's SUSET custom variable before loading the associated extension, because GUC can remember whether the SET was done as a superuser or not. The plperl regression tests contain an example where this is useful.
2011-10-04 22:13:16 +02:00
conf->gen.reset_scontext = context;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
}
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
for (stack = conf->gen.stack; stack; stack = stack->prev)
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
{
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
if (stack->source <= source)
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
stack->prior.val.boolval = newval;
set_extra_field(&conf->gen, &stack->prior.extra,
newextra);
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
stack->source = source;
Remember the source GucContext for each GUC parameter. We used to just remember the GucSource, but saving GucContext too provides a little more information --- notably, whether a SET was done by a superuser or regular user. This allows us to rip out the fairly dodgy code that define_custom_variable used to use to try to infer the context to re-install a pre-existing setting with. In particular, it now works for a superuser to SET a extension's SUSET custom variable before loading the associated extension, because GUC can remember whether the SET was done as a superuser or not. The plperl regression tests contain an example where this is useful.
2011-10-04 22:13:16 +02:00
stack->scontext = context;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
}
}
Extend GUC concepts of parse_hook and assign_hook to all four supported datatypes, not only strings. parse_hook is useless for bool, I suppose, but it seems possibly useful for int and double to apply variable-specific constraints that are more complex than simple range limits. assign_hook is definitely useful for all datatypes --- we need it right now for bool to support date cache reset when changing Australian timezone rule setting. Also, clean up some residual problems with the reset all/show all patch, including memory leaks and mistaken reset of PostPortNumber. It seems best that RESET ALL not touch variables that don't have SUSET or USERSET context.
2001-06-13 00:54:06 +02:00
}
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
/* Perhaps we didn't install newextra anywhere */
if (newextra && !extra_field_used(&conf->gen, newextra))
free(newextra);
pgindent run. Make it all clean.
2001-03-22 05:01:46 +01:00
break;
Add all possible config file options.
2001-01-24 19:37:31 +01:00
}
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
case PGC_INT:
Add all possible config file options.
2001-01-24 19:37:31 +01:00
{
pgindent run. Make it all clean.
2001-03-22 05:01:46 +01:00
struct config_int *conf = (struct config_int *) record;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
int newval;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
void *newextra = NULL;
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
if (value)
{
Provide a HINT listing the allowed unit names when a GUC variable seems to contain a wrong unit specification, per discussion. In passing, fix the code to avoid unnecessary integer overflows when converting units, and to detect overflows when they do occur.
2007-06-21 20:14:21 +02:00
const char *hintmsg;
if (!parse_int(value, &newval, conf->gen.flags, &hintmsg))
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
{
ereport(elevel,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
errmsg("invalid value for parameter \"%s\": \"%s\"",
name, value),
Make some strings translatable again that were accidentally removed in earlier patch to fix "printf-arguments".
2009-01-21 10:28:26 +01:00
hintmsg ? errhint("%s", _(hintmsg)) : 0));
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
return 0;
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
}
if (newval < conf->min || newval > conf->max)
{
ereport(elevel,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("%d is outside the valid range for parameter \"%s\" (%d .. %d)",
newval, name, conf->min, conf->max)));
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
return 0;
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
}
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
if (!call_int_check_hook(conf, &newval, &newextra,
source, elevel))
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
return 0;
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
}
Make configuration parameters fall back to their default values when they are removed from the configuration file. Joachim Wieland
2007-04-21 22:02:41 +02:00
else if (source == PGC_S_DEFAULT)
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
{
Make configuration parameters fall back to their default values when they are removed from the configuration file. Joachim Wieland
2007-04-21 22:02:41 +02:00
newval = conf->boot_val;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
if (!call_int_check_hook(conf, &newval, &newextra,
source, elevel))
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
return 0;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
}
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
else
{
newval = conf->reset_val;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
newextra = conf->reset_extra;
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
source = conf->gen.reset_source;
Remember the source GucContext for each GUC parameter. We used to just remember the GucSource, but saving GucContext too provides a little more information --- notably, whether a SET was done by a superuser or regular user. This allows us to rip out the fairly dodgy code that define_custom_variable used to use to try to infer the context to re-install a pre-existing setting with. In particular, it now works for a superuser to SET a extension's SUSET custom variable before loading the associated extension, because GUC can remember whether the SET was done as a superuser or not. The plperl regression tests contain an example where this is useful.
2011-10-04 22:13:16 +02:00
context = conf->gen.reset_scontext;
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
}
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
if (prohibitValueChange)
{
if (*conf->variable != newval)
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
{
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
ereport(elevel,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
(errcode(ERRCODE_CANT_CHANGE_RUNTIME_PARAM),
errmsg("parameter \"%s\" cannot be changed without restarting the server",
name)));
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
return 0;
}
return -1;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
}
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
if (changeVal)
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
/* Save old value to support transaction abort */
if (!makeDefault)
push_old_value(&conf->gen, action);
if (conf->assign_hook)
(*conf->assign_hook) (newval, newextra);
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
*conf->variable = newval;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
set_extra_field(&conf->gen, &conf->gen.extra,
newextra);
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
conf->gen.source = source;
Remember the source GucContext for each GUC parameter. We used to just remember the GucSource, but saving GucContext too provides a little more information --- notably, whether a SET was done by a superuser or regular user. This allows us to rip out the fairly dodgy code that define_custom_variable used to use to try to infer the context to re-install a pre-existing setting with. In particular, it now works for a superuser to SET a extension's SUSET custom variable before loading the associated extension, because GUC can remember whether the SET was done as a superuser or not. The plperl regression tests contain an example where this is useful.
2011-10-04 22:13:16 +02:00
conf->gen.scontext = context;
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
}
if (makeDefault)
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
{
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
GucStack *stack;
if (conf->gen.reset_source <= source)
reset all: command line and .conf options change defaults on RESET ALL those are restored. show all: GUC + non-GUC. SHOW ALL, RESET ALL Marko Kreen
2001-06-07 06:50:57 +02:00
{
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
conf->reset_val = newval;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
set_extra_field(&conf->gen, &conf->reset_extra,
newextra);
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
conf->gen.reset_source = source;
Remember the source GucContext for each GUC parameter. We used to just remember the GucSource, but saving GucContext too provides a little more information --- notably, whether a SET was done by a superuser or regular user. This allows us to rip out the fairly dodgy code that define_custom_variable used to use to try to infer the context to re-install a pre-existing setting with. In particular, it now works for a superuser to SET a extension's SUSET custom variable before loading the associated extension, because GUC can remember whether the SET was done as a superuser or not. The plperl regression tests contain an example where this is useful.
2011-10-04 22:13:16 +02:00
conf->gen.reset_scontext = context;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
}
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
for (stack = conf->gen.stack; stack; stack = stack->prev)
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
{
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
if (stack->source <= source)
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
stack->prior.val.intval = newval;
set_extra_field(&conf->gen, &stack->prior.extra,
newextra);
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
stack->source = source;
Remember the source GucContext for each GUC parameter. We used to just remember the GucSource, but saving GucContext too provides a little more information --- notably, whether a SET was done by a superuser or regular user. This allows us to rip out the fairly dodgy code that define_custom_variable used to use to try to infer the context to re-install a pre-existing setting with. In particular, it now works for a superuser to SET a extension's SUSET custom variable before loading the associated extension, because GUC can remember whether the SET was done as a superuser or not. The plperl regression tests contain an example where this is useful.
2011-10-04 22:13:16 +02:00
stack->scontext = context;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
}
}
Extend GUC concepts of parse_hook and assign_hook to all four supported datatypes, not only strings. parse_hook is useless for bool, I suppose, but it seems possibly useful for int and double to apply variable-specific constraints that are more complex than simple range limits. assign_hook is definitely useful for all datatypes --- we need it right now for bool to support date cache reset when changing Australian timezone rule setting. Also, clean up some residual problems with the reset all/show all patch, including memory leaks and mistaken reset of PostPortNumber. It seems best that RESET ALL not touch variables that don't have SUSET or USERSET context.
2001-06-13 00:54:06 +02:00
}
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
/* Perhaps we didn't install newextra anywhere */
if (newextra && !extra_field_used(&conf->gen, newextra))
free(newextra);
pgindent run. Make it all clean.
2001-03-22 05:01:46 +01:00
break;
Add all possible config file options.
2001-01-24 19:37:31 +01:00
}
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
case PGC_REAL:
Add all possible config file options.
2001-01-24 19:37:31 +01:00
{
pgindent run. Make it all clean.
2001-03-22 05:01:46 +01:00
struct config_real *conf = (struct config_real *) record;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
double newval;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
void *newextra = NULL;
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
if (value)
{
if (!parse_real(value, &newval))
{
ereport(elevel,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("parameter \"%s\" requires a numeric value",
name)));
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
return 0;
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
}
if (newval < conf->min || newval > conf->max)
{
ereport(elevel,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("%g is outside the valid range for parameter \"%s\" (%g .. %g)",
newval, name, conf->min, conf->max)));
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
return 0;
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
}
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
if (!call_real_check_hook(conf, &newval, &newextra,
source, elevel))
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
return 0;
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
}
Make configuration parameters fall back to their default values when they are removed from the configuration file. Joachim Wieland
2007-04-21 22:02:41 +02:00
else if (source == PGC_S_DEFAULT)
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
{
Make configuration parameters fall back to their default values when they are removed from the configuration file. Joachim Wieland
2007-04-21 22:02:41 +02:00
newval = conf->boot_val;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
if (!call_real_check_hook(conf, &newval, &newextra,
source, elevel))
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
return 0;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
}
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
else
{
newval = conf->reset_val;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
newextra = conf->reset_extra;
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
source = conf->gen.reset_source;
Remember the source GucContext for each GUC parameter. We used to just remember the GucSource, but saving GucContext too provides a little more information --- notably, whether a SET was done by a superuser or regular user. This allows us to rip out the fairly dodgy code that define_custom_variable used to use to try to infer the context to re-install a pre-existing setting with. In particular, it now works for a superuser to SET a extension's SUSET custom variable before loading the associated extension, because GUC can remember whether the SET was done as a superuser or not. The plperl regression tests contain an example where this is useful.
2011-10-04 22:13:16 +02:00
context = conf->gen.reset_scontext;
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
}
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
if (prohibitValueChange)
{
if (*conf->variable != newval)
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
{
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
ereport(elevel,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
(errcode(ERRCODE_CANT_CHANGE_RUNTIME_PARAM),
errmsg("parameter \"%s\" cannot be changed without restarting the server",
name)));
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
return 0;
}
return -1;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
}
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
if (changeVal)
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
/* Save old value to support transaction abort */
if (!makeDefault)
push_old_value(&conf->gen, action);
if (conf->assign_hook)
(*conf->assign_hook) (newval, newextra);
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
*conf->variable = newval;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
set_extra_field(&conf->gen, &conf->gen.extra,
newextra);
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
conf->gen.source = source;
Remember the source GucContext for each GUC parameter. We used to just remember the GucSource, but saving GucContext too provides a little more information --- notably, whether a SET was done by a superuser or regular user. This allows us to rip out the fairly dodgy code that define_custom_variable used to use to try to infer the context to re-install a pre-existing setting with. In particular, it now works for a superuser to SET a extension's SUSET custom variable before loading the associated extension, because GUC can remember whether the SET was done as a superuser or not. The plperl regression tests contain an example where this is useful.
2011-10-04 22:13:16 +02:00
conf->gen.scontext = context;
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
}
if (makeDefault)
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
{
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
GucStack *stack;
if (conf->gen.reset_source <= source)
reset all: command line and .conf options change defaults on RESET ALL those are restored. show all: GUC + non-GUC. SHOW ALL, RESET ALL Marko Kreen
2001-06-07 06:50:57 +02:00
{
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
conf->reset_val = newval;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
set_extra_field(&conf->gen, &conf->reset_extra,
newextra);
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
conf->gen.reset_source = source;
Remember the source GucContext for each GUC parameter. We used to just remember the GucSource, but saving GucContext too provides a little more information --- notably, whether a SET was done by a superuser or regular user. This allows us to rip out the fairly dodgy code that define_custom_variable used to use to try to infer the context to re-install a pre-existing setting with. In particular, it now works for a superuser to SET a extension's SUSET custom variable before loading the associated extension, because GUC can remember whether the SET was done as a superuser or not. The plperl regression tests contain an example where this is useful.
2011-10-04 22:13:16 +02:00
conf->gen.reset_scontext = context;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
}
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
for (stack = conf->gen.stack; stack; stack = stack->prev)
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
{
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
if (stack->source <= source)
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
stack->prior.val.realval = newval;
set_extra_field(&conf->gen, &stack->prior.extra,
newextra);
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
stack->source = source;
Remember the source GucContext for each GUC parameter. We used to just remember the GucSource, but saving GucContext too provides a little more information --- notably, whether a SET was done by a superuser or regular user. This allows us to rip out the fairly dodgy code that define_custom_variable used to use to try to infer the context to re-install a pre-existing setting with. In particular, it now works for a superuser to SET a extension's SUSET custom variable before loading the associated extension, because GUC can remember whether the SET was done as a superuser or not. The plperl regression tests contain an example where this is useful.
2011-10-04 22:13:16 +02:00
stack->scontext = context;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
}
}
Extend GUC concepts of parse_hook and assign_hook to all four supported datatypes, not only strings. parse_hook is useless for bool, I suppose, but it seems possibly useful for int and double to apply variable-specific constraints that are more complex than simple range limits. assign_hook is definitely useful for all datatypes --- we need it right now for bool to support date cache reset when changing Australian timezone rule setting. Also, clean up some residual problems with the reset all/show all patch, including memory leaks and mistaken reset of PostPortNumber. It seems best that RESET ALL not touch variables that don't have SUSET or USERSET context.
2001-06-13 00:54:06 +02:00
}
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
/* Perhaps we didn't install newextra anywhere */
if (newextra && !extra_field_used(&conf->gen, newextra))
free(newextra);
pgindent run. Make it all clean.
2001-03-22 05:01:46 +01:00
break;
Add all possible config file options.
2001-01-24 19:37:31 +01:00
}
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
case PGC_STRING:
{
pgindent run. Make it all clean.
2001-03-22 05:01:46 +01:00
struct config_string *conf = (struct config_string *) record;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
char *newval;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
void *newextra = NULL;
pgindent run. Make it all clean.
2001-03-22 05:01:46 +01:00
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
if (value)
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
/*
pgindent run before PG 9.1 beta 1.
2011-04-10 17:42:00 +02:00
* The value passed by the caller could be transient, so
* we always strdup it.
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
*/
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
newval = guc_strdup(elevel, value);
if (newval == NULL)
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
return 0;
pgindent run for 8.2.
2006-10-04 02:30:14 +02:00
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
/*
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
* The only built-in "parsing" check we have is to apply
pgindent run for 8.2.
2006-10-04 02:30:14 +02:00
* truncation if GUC_IS_NAME.
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
*/
if (conf->gen.flags & GUC_IS_NAME)
truncate_identifier(newval, strlen(newval), true);
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
if (!call_string_check_hook(conf, &newval, &newextra,
source, elevel))
{
free(newval);
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
return 0;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
}
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
}
Make configuration parameters fall back to their default values when they are removed from the configuration file. Joachim Wieland
2007-04-21 22:02:41 +02:00
else if (source == PGC_S_DEFAULT)
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
/* non-NULL boot_val must always get strdup'd */
if (conf->boot_val != NULL)
Make configuration parameters fall back to their default values when they are removed from the configuration file. Joachim Wieland
2007-04-21 22:02:41 +02:00
{
newval = guc_strdup(elevel, conf->boot_val);
if (newval == NULL)
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
return 0;
Make configuration parameters fall back to their default values when they are removed from the configuration file. Joachim Wieland
2007-04-21 22:02:41 +02:00
}
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
else
newval = NULL;
if (!call_string_check_hook(conf, &newval, &newextra,
source, elevel))
{
free(newval);
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
return 0;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
}
Make configuration parameters fall back to their default values when they are removed from the configuration file. Joachim Wieland
2007-04-21 22:02:41 +02:00
}
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
else
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
{
/*
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
* strdup not needed, since reset_val is already under
* guc.c's control
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
*/
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
newval = conf->reset_val;
newextra = conf->reset_extra;
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
source = conf->gen.reset_source;
Remember the source GucContext for each GUC parameter. We used to just remember the GucSource, but saving GucContext too provides a little more information --- notably, whether a SET was done by a superuser or regular user. This allows us to rip out the fairly dodgy code that define_custom_variable used to use to try to infer the context to re-install a pre-existing setting with. In particular, it now works for a superuser to SET a extension's SUSET custom variable before loading the associated extension, because GUC can remember whether the SET was done as a superuser or not. The plperl regression tests contain an example where this is useful.
2011-10-04 22:13:16 +02:00
context = conf->gen.reset_scontext;
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
}
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
if (prohibitValueChange)
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
/* newval shouldn't be NULL, so we're a bit sloppy here */
if (*conf->variable == NULL || newval == NULL ||
strcmp(*conf->variable, newval) != 0)
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
{
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
ereport(elevel,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
(errcode(ERRCODE_CANT_CHANGE_RUNTIME_PARAM),
errmsg("parameter \"%s\" cannot be changed without restarting the server",
name)));
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
return 0;
}
return -1;
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
}
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
if (changeVal)
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
/* Save old value to support transaction abort */
if (!makeDefault)
push_old_value(&conf->gen, action);
if (conf->assign_hook)
(*conf->assign_hook) (newval, newextra);
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
set_string_field(conf, conf->variable, newval);
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
set_extra_field(&conf->gen, &conf->gen.extra,
newextra);
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
conf->gen.source = source;
Remember the source GucContext for each GUC parameter. We used to just remember the GucSource, but saving GucContext too provides a little more information --- notably, whether a SET was done by a superuser or regular user. This allows us to rip out the fairly dodgy code that define_custom_variable used to use to try to infer the context to re-install a pre-existing setting with. In particular, it now works for a superuser to SET a extension's SUSET custom variable before loading the associated extension, because GUC can remember whether the SET was done as a superuser or not. The plperl regression tests contain an example where this is useful.
2011-10-04 22:13:16 +02:00
conf->gen.scontext = context;
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
}
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
if (makeDefault)
{
GucStack *stack;
if (conf->gen.reset_source <= source)
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
{
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
set_string_field(conf, &conf->reset_val, newval);
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
set_extra_field(&conf->gen, &conf->reset_extra,
newextra);
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
conf->gen.reset_source = source;
Remember the source GucContext for each GUC parameter. We used to just remember the GucSource, but saving GucContext too provides a little more information --- notably, whether a SET was done by a superuser or regular user. This allows us to rip out the fairly dodgy code that define_custom_variable used to use to try to infer the context to re-install a pre-existing setting with. In particular, it now works for a superuser to SET a extension's SUSET custom variable before loading the associated extension, because GUC can remember whether the SET was done as a superuser or not. The plperl regression tests contain an example where this is useful.
2011-10-04 22:13:16 +02:00
conf->gen.reset_scontext = context;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
}
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
for (stack = conf->gen.stack; stack; stack = stack->prev)
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
{
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
if (stack->source <= source)
reset all: command line and .conf options change defaults on RESET ALL those are restored. show all: GUC + non-GUC. SHOW ALL, RESET ALL Marko Kreen
2001-06-07 06:50:57 +02:00
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
set_string_field(conf, &stack->prior.val.stringval,
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
newval);
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
set_extra_field(&conf->gen, &stack->prior.extra,
newextra);
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
stack->source = source;
Remember the source GucContext for each GUC parameter. We used to just remember the GucSource, but saving GucContext too provides a little more information --- notably, whether a SET was done by a superuser or regular user. This allows us to rip out the fairly dodgy code that define_custom_variable used to use to try to infer the context to re-install a pre-existing setting with. In particular, it now works for a superuser to SET a extension's SUSET custom variable before loading the associated extension, because GUC can remember whether the SET was done as a superuser or not. The plperl regression tests contain an example where this is useful.
2011-10-04 22:13:16 +02:00
stack->scontext = context;
reset all: command line and .conf options change defaults on RESET ALL those are restored. show all: GUC + non-GUC. SHOW ALL, RESET ALL Marko Kreen
2001-06-07 06:50:57 +02:00
}
pgindent run. Make it all clean.
2001-03-22 05:01:46 +01:00
}
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
}
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
/* Perhaps we didn't install newval anywhere */
if (newval && !string_field_used(conf, newval))
Back out patch to reorganize guc processing. Was causing regression failures.
2006-08-12 06:12:41 +02:00
free(newval);
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
/* Perhaps we didn't install newextra anywhere */
if (newextra && !extra_field_used(&conf->gen, newextra))
free(newextra);
pgindent run. Make it all clean.
2001-03-22 05:01:46 +01:00
break;
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
}
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
case PGC_ENUM:
{
struct config_enum *conf = (struct config_enum *) record;
int newval;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
void *newextra = NULL;
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
if (value)
{
Fix postgres --describe-config for guc enums, breakage noted by Alvaro. While at it, rename option lookup functions to make names clearer, per discussion with Tom.
2008-03-17 18:45:09 +01:00
if (!config_enum_lookup_by_name(conf, value, &newval))
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
{
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
char *hintmsg;
hintmsg = config_enum_get_options(conf,
"Available values: ",
Minor code embellishments.
2008-12-02 03:00:32 +01:00
".", ", ");
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
ereport(elevel,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
errmsg("invalid value for parameter \"%s\": \"%s\"",
name, value),
Make some strings translatable again that were accidentally removed in earlier patch to fix "printf-arguments".
2009-01-21 10:28:26 +01:00
hintmsg ? errhint("%s", _(hintmsg)) : 0));
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
if (hintmsg)
pfree(hintmsg);
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
return 0;
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
}
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
if (!call_enum_check_hook(conf, &newval, &newextra,
source, elevel))
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
return 0;
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
}
else if (source == PGC_S_DEFAULT)
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
{
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
newval = conf->boot_val;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
if (!call_enum_check_hook(conf, &newval, &newextra,
source, elevel))
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
return 0;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
}
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
else
{
newval = conf->reset_val;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
newextra = conf->reset_extra;
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
source = conf->gen.reset_source;
Remember the source GucContext for each GUC parameter. We used to just remember the GucSource, but saving GucContext too provides a little more information --- notably, whether a SET was done by a superuser or regular user. This allows us to rip out the fairly dodgy code that define_custom_variable used to use to try to infer the context to re-install a pre-existing setting with. In particular, it now works for a superuser to SET a extension's SUSET custom variable before loading the associated extension, because GUC can remember whether the SET was done as a superuser or not. The plperl regression tests contain an example where this is useful.
2011-10-04 22:13:16 +02:00
context = conf->gen.reset_scontext;
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
}
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
if (prohibitValueChange)
{
if (*conf->variable != newval)
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
{
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
ereport(elevel,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
(errcode(ERRCODE_CANT_CHANGE_RUNTIME_PARAM),
errmsg("parameter \"%s\" cannot be changed without restarting the server",
name)));
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
return 0;
}
return -1;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
}
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
if (changeVal)
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
/* Save old value to support transaction abort */
if (!makeDefault)
push_old_value(&conf->gen, action);
if (conf->assign_hook)
(*conf->assign_hook) (newval, newextra);
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
*conf->variable = newval;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
set_extra_field(&conf->gen, &conf->gen.extra,
newextra);
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
conf->gen.source = source;
Remember the source GucContext for each GUC parameter. We used to just remember the GucSource, but saving GucContext too provides a little more information --- notably, whether a SET was done by a superuser or regular user. This allows us to rip out the fairly dodgy code that define_custom_variable used to use to try to infer the context to re-install a pre-existing setting with. In particular, it now works for a superuser to SET a extension's SUSET custom variable before loading the associated extension, because GUC can remember whether the SET was done as a superuser or not. The plperl regression tests contain an example where this is useful.
2011-10-04 22:13:16 +02:00
conf->gen.scontext = context;
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
}
if (makeDefault)
{
GucStack *stack;
if (conf->gen.reset_source <= source)
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
{
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
conf->reset_val = newval;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
set_extra_field(&conf->gen, &conf->reset_extra,
newextra);
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
conf->gen.reset_source = source;
Remember the source GucContext for each GUC parameter. We used to just remember the GucSource, but saving GucContext too provides a little more information --- notably, whether a SET was done by a superuser or regular user. This allows us to rip out the fairly dodgy code that define_custom_variable used to use to try to infer the context to re-install a pre-existing setting with. In particular, it now works for a superuser to SET a extension's SUSET custom variable before loading the associated extension, because GUC can remember whether the SET was done as a superuser or not. The plperl regression tests contain an example where this is useful.
2011-10-04 22:13:16 +02:00
conf->gen.reset_scontext = context;
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
}
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
for (stack = conf->gen.stack; stack; stack = stack->prev)
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
{
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
if (stack->source <= source)
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
stack->prior.val.enumval = newval;
set_extra_field(&conf->gen, &stack->prior.extra,
newextra);
Fix an old corner-case bug in set_config_option: push_old_value has to be called before, not after, calling the assign_hook if any. This is because push_old_value might fail (due to palloc out-of-memory), and in that case there would be no stack entry to tell transaction abort to undo the GUC assignment. Of course the actual assignment to the GUC variable hasn't happened yet --- but the assign_hook might have altered subsidiary state. Without a stack entry we won't call it again to make it undo such actions. So this is necessary to make the world safe for assign_hooks with side effects. Per a discussion a couple weeks ago with Magnus. Back-patch to 8.0. 7.x did not have the problem because it did not have allocatable stacks of GUC values.
2008-05-26 20:54:29 +02:00
stack->source = source;
Remember the source GucContext for each GUC parameter. We used to just remember the GucSource, but saving GucContext too provides a little more information --- notably, whether a SET was done by a superuser or regular user. This allows us to rip out the fairly dodgy code that define_custom_variable used to use to try to infer the context to re-install a pre-existing setting with. In particular, it now works for a superuser to SET a extension's SUSET custom variable before loading the associated extension, because GUC can remember whether the SET was done as a superuser or not. The plperl regression tests contain an example where this is useful.
2011-10-04 22:13:16 +02:00
stack->scontext = context;
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
}
}
}
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
/* Perhaps we didn't install newextra anywhere */
if (newextra && !extra_field_used(&conf->gen, newextra))
free(newextra);
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
break;
}
Add all possible config file options.
2001-01-24 19:37:31 +01:00
}
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
Rename DoIt to changeVar, for clarity.
2003-09-01 06:15:51 +02:00
if (changeVal && (record->flags & GUC_REPORT))
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
ReportGUCOption(record);
In the continuing saga of FE/BE protocol revisions, add reporting of initial values and runtime changes in selected parameters. This gets rid of the need for an initial 'select pg_client_encoding()' query in libpq, bringing us back to one message transmitted in each direction for a standard connection startup. To allow server version to be sent using the same GUC mechanism that handles other parameters, invent the concept of a never-settable GUC parameter: you can 'show server_version' but it's not settable by any GUC input source. Create 'lc_collate' and 'lc_ctype' never-settable parameters so that people can find out these settings without need for pg_controldata. (These side ideas were all discussed some time ago in pgsql-hackers, but not yet implemented.)
2003-04-25 21:45:10 +02:00
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
return changeVal ? 1 : -1;
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
}
Add "source file" and "source line" information to each GUC variable. initdb forced due to changes in the pg_settings view. Magnus Hagander and Alvaro Herrera.
2008-09-10 20:09:20 +02:00
/*
* Set the fields for source file and line number the setting came from.
*/
static void
set_config_sourcefile(const char *name, char *sourcefile, int sourceline)
{
struct config_generic *record;
int elevel;
/*
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
* To avoid cluttering the log, only the postmaster bleats loudly about
* problems with the config file.
Add "source file" and "source line" information to each GUC variable. initdb forced due to changes in the pg_settings view. Magnus Hagander and Alvaro Herrera.
2008-09-10 20:09:20 +02:00
*/
elevel = IsUnderPostmaster ? DEBUG3 : LOG;
record = find_option(name, true, elevel);
/* should not happen */
if (record == NULL)
elog(ERROR, "unrecognized configuration parameter \"%s\"", name);
Tweak newly added set_config_sourcefile() so that the target record isn't left corrupt if guc_strdup should fail.
2008-09-10 21:16:22 +02:00
sourcefile = guc_strdup(elevel, sourcefile);
Add "source file" and "source line" information to each GUC variable. initdb forced due to changes in the pg_settings view. Magnus Hagander and Alvaro Herrera.
2008-09-10 20:09:20 +02:00
if (record->sourcefile)
free(record->sourcefile);
Tweak newly added set_config_sourcefile() so that the target record isn't left corrupt if guc_strdup should fail.
2008-09-10 21:16:22 +02:00
record->sourcefile = sourcefile;
Add "source file" and "source line" information to each GUC variable. initdb forced due to changes in the pg_settings view. Magnus Hagander and Alvaro Herrera.
2008-09-10 20:09:20 +02:00
record->sourceline = sourceline;
}
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
/*
* Set a config option to the given value. See also set_config_option,
pgindent run.
2002-09-04 22:31:48 +02:00
* this is just the wrapper to be called from outside GUC. NB: this
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
* is used only for non-transactional operations.
Add "source file" and "source line" information to each GUC variable. initdb forced due to changes in the pg_settings view. Magnus Hagander and Alvaro Herrera.
2008-09-10 20:09:20 +02:00
*
* Note: there is no support here for setting source file/line, as it
* is currently not needed.
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
*/
void
Extend GUC concepts of parse_hook and assign_hook to all four supported datatypes, not only strings. parse_hook is useless for bool, I suppose, but it seems possibly useful for int and double to apply variable-specific constraints that are more complex than simple range limits. assign_hook is definitely useful for all datatypes --- we need it right now for bool to support date cache reset when changing Australian timezone rule setting. Also, clean up some residual problems with the reset all/show all patch, including memory leaks and mistaken reset of PostPortNumber. It seems best that RESET ALL not touch variables that don't have SUSET or USERSET context.
2001-06-13 00:54:06 +02:00
SetConfigOption(const char *name, const char *value,
Save source of GUC settings, allowing different sources to be processed in any order without affecting results.
2002-02-23 02:31:37 +01:00
GucContext context, GucSource source)
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
{
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
(void) set_config_option(name, value, context, source,
GUC_ACTION_SET, true);
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
}
/*
Fix another oversight in logging of changes in postgresql.conf settings. We were using GetConfigOption to collect the old value of each setting, overlooking the possibility that it didn't exist yet. This does happen in the case of adding a new entry within a custom variable class, as exhibited in bug #6097 from Maxim Boguk. To fix, add a missing_ok parameter to GetConfigOption, but only in 9.1 and HEAD --- it seems possible that some third-party code is using that function, so changing its API in a minor release would cause problems. In 9.0, create a near-duplicate function instead.
2011-07-08 23:02:58 +02:00
* Fetch the current value of the option `name', as a string.
*
* If the option doesn't exist, return NULL if missing_ok is true (NOTE that
* this cannot be distinguished from a string variable with a NULL value!),
* otherwise throw an ereport and don't return.
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
*
Fix a couple of issues in recent patch to print updates to postgresql.conf settings: avoid calling superuser() in contexts where it's not defined, don't leak the transient copies of GetConfigOption output, and avoid the whole exercise in postmaster child processes. I found that actually no current caller of GetConfigOption has any use for its internal check of GUC_SUPERUSER_ONLY. But rather than just remove that entirely, it seemed better to add a parameter indicating whether to enforce the check. Per report from Simon and subsequent testing.
2009-10-03 20:04:57 +02:00
* If restrict_superuser is true, we also enforce that only superusers can
* see GUC_SUPERUSER_ONLY variables. This should only be passed as true
* in user-driven calls.
*
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
* The string is *not* allocated for modification and is really only
* valid until the next call to configuration related functions.
*/
const char *
Fix another oversight in logging of changes in postgresql.conf settings. We were using GetConfigOption to collect the old value of each setting, overlooking the possibility that it didn't exist yet. This does happen in the case of adding a new entry within a custom variable class, as exhibited in bug #6097 from Maxim Boguk. To fix, add a missing_ok parameter to GetConfigOption, but only in 9.1 and HEAD --- it seems possible that some third-party code is using that function, so changing its API in a minor release would cause problems. In 9.0, create a near-duplicate function instead.
2011-07-08 23:02:58 +02:00
GetConfigOption(const char *name, bool missing_ok, bool restrict_superuser)
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
{
pgindent run. Make it all clean.
2001-03-22 05:01:46 +01:00
struct config_generic *record;
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
static char buffer[256];
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
record = find_option(name, false, ERROR);
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
if (record == NULL)
Fix another oversight in logging of changes in postgresql.conf settings. We were using GetConfigOption to collect the old value of each setting, overlooking the possibility that it didn't exist yet. This does happen in the case of adding a new entry within a custom variable class, as exhibited in bug #6097 from Maxim Boguk. To fix, add a missing_ok parameter to GetConfigOption, but only in 9.1 and HEAD --- it seems possible that some third-party code is using that function, so changing its API in a minor release would cause problems. In 9.0, create a near-duplicate function instead.
2011-07-08 23:02:58 +02:00
{
if (missing_ok)
return NULL;
Error message editing in backend/utils (except /adt).
2003-07-25 22:18:01 +02:00
ereport(ERROR,
(errcode(ERRCODE_UNDEFINED_OBJECT),
Fix another oversight in logging of changes in postgresql.conf settings. We were using GetConfigOption to collect the old value of each setting, overlooking the possibility that it didn't exist yet. This does happen in the case of adding a new entry within a custom variable class, as exhibited in bug #6097 from Maxim Boguk. To fix, add a missing_ok parameter to GetConfigOption, but only in 9.1 and HEAD --- it seems possible that some third-party code is using that function, so changing its API in a minor release would cause problems. In 9.0, create a near-duplicate function instead.
2011-07-08 23:02:58 +02:00
errmsg("unrecognized configuration parameter \"%s\"",
name)));
}
Fix a couple of issues in recent patch to print updates to postgresql.conf settings: avoid calling superuser() in contexts where it's not defined, don't leak the transient copies of GetConfigOption output, and avoid the whole exercise in postmaster child processes. I found that actually no current caller of GetConfigOption has any use for its internal check of GUC_SUPERUSER_ONLY. But rather than just remove that entirely, it seemed better to add a parameter indicating whether to enforce the check. Per report from Simon and subsequent testing.
2009-10-03 20:04:57 +02:00
if (restrict_superuser &&
(record->flags & GUC_SUPERUSER_ONLY) &&
!superuser())
Add a GUC_SUPERUSER_ONLY flag to mark GUC variables that should not be examinable by non-superusers, and use it to protect the recently-added GUC variables for data directory and config files. For now I have only flagged those variables that could be used to deduce something about the server's filesystem layout, but possibly we should also mark vars related to logging settings and other admin-only information?
2004-10-22 21:48:19 +02:00
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
errmsg("must be superuser to examine \"%s\"", name)));
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
switch (record->vartype)
Add all possible config file options.
2001-01-24 19:37:31 +01:00
{
case PGC_BOOL:
pgindent run. Make it all clean.
2001-03-22 05:01:46 +01:00
return *((struct config_bool *) record)->variable ? "on" : "off";
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
Add all possible config file options.
2001-01-24 19:37:31 +01:00
case PGC_INT:
Extend GUC concepts of parse_hook and assign_hook to all four supported datatypes, not only strings. parse_hook is useless for bool, I suppose, but it seems possibly useful for int and double to apply variable-specific constraints that are more complex than simple range limits. assign_hook is definitely useful for all datatypes --- we need it right now for bool to support date cache reset when changing Australian timezone rule setting. Also, clean up some residual problems with the reset all/show all patch, including memory leaks and mistaken reset of PostPortNumber. It seems best that RESET ALL not touch variables that don't have SUSET or USERSET context.
2001-06-13 00:54:06 +02:00
snprintf(buffer, sizeof(buffer), "%d",
*((struct config_int *) record)->variable);
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
return buffer;
Add all possible config file options.
2001-01-24 19:37:31 +01:00
case PGC_REAL:
Extend GUC concepts of parse_hook and assign_hook to all four supported datatypes, not only strings. parse_hook is useless for bool, I suppose, but it seems possibly useful for int and double to apply variable-specific constraints that are more complex than simple range limits. assign_hook is definitely useful for all datatypes --- we need it right now for bool to support date cache reset when changing Australian timezone rule setting. Also, clean up some residual problems with the reset all/show all patch, including memory leaks and mistaken reset of PostPortNumber. It seems best that RESET ALL not touch variables that don't have SUSET or USERSET context.
2001-06-13 00:54:06 +02:00
snprintf(buffer, sizeof(buffer), "%g",
*((struct config_real *) record)->variable);
The heralded `Grand Unified Configuration scheme' (GUC) That means you can now set your options in either or all of $PGDATA/configuration, some postmaster option (--enable-fsync=off), or set a SET command. The list of options is in backend/utils/misc/guc.c, documentation will be written post haste. pg_options is gone, so is that pq_geqo config file. Also removed were backend -K, -Q, and -T options (no longer applicable, although -d0 does the same as -Q). Added to configure an --enable-syslog option. changed all callers from TPRINTF to elog(DEBUG)
2000-05-31 02:28:42 +02:00
return buffer;
case PGC_STRING:
pgindent run. Make it all clean.
2001-03-22 05:01:46 +01:00
return *((struct config_string *) record)->variable;
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
case PGC_ENUM:
Fix postgres --describe-config for guc enums, breakage noted by Alvaro. While at it, rename option lookup functions to make names clearer, per discussion with Tom.
2008-03-17 18:45:09 +01:00
return config_enum_lookup_by_value((struct config_enum *) record,
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
*((struct config_enum *) record)->variable);
Add all possible config file options.
2001-01-24 19:37:31 +01:00
}
return NULL;
}
Move PGPORT envar handling to ResetAllOptions(). Improve long options parsing to not clobber the optarg string -- so that we can bring SetOptsFile() up to speed.
2000-07-03 22:46:10 +02:00
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
/*
* Get the RESET value associated with the given option.
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
*
* Note: this is not re-entrant, due to use of static result buffer;
* not to mention that a string variable could have its reset_val changed.
* Beware of assuming the result value is good for very long.
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
*/
const char *
GetConfigOptionResetString(const char *name)
reset all: command line and .conf options change defaults on RESET ALL those are restored. show all: GUC + non-GUC. SHOW ALL, RESET ALL Marko Kreen
2001-06-07 06:50:57 +02:00
{
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
struct config_generic *record;
static char buffer[256];
reset all: command line and .conf options change defaults on RESET ALL those are restored. show all: GUC + non-GUC. SHOW ALL, RESET ALL Marko Kreen
2001-06-07 06:50:57 +02:00
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
record = find_option(name, false, ERROR);
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
if (record == NULL)
Error message editing in backend/utils (except /adt).
2003-07-25 22:18:01 +02:00
ereport(ERROR,
(errcode(ERRCODE_UNDEFINED_OBJECT),
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
errmsg("unrecognized configuration parameter \"%s\"", name)));
Add a GUC_SUPERUSER_ONLY flag to mark GUC variables that should not be examinable by non-superusers, and use it to protect the recently-added GUC variables for data directory and config files. For now I have only flagged those variables that could be used to deduce something about the server's filesystem layout, but possibly we should also mark vars related to logging settings and other admin-only information?
2004-10-22 21:48:19 +02:00
if ((record->flags & GUC_SUPERUSER_ONLY) && !superuser())
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
errmsg("must be superuser to examine \"%s\"", name)));
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
switch (record->vartype)
reset all: command line and .conf options change defaults on RESET ALL those are restored. show all: GUC + non-GUC. SHOW ALL, RESET ALL Marko Kreen
2001-06-07 06:50:57 +02:00
{
case PGC_BOOL:
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
return ((struct config_bool *) record)->reset_val ? "on" : "off";
Add GUC setting for Australian timezones. Uses new GUC boolean callback functions to clear date cache. Allow regression tests to pass when timezone set.
2001-06-18 18:14:44 +02:00
reset all: command line and .conf options change defaults on RESET ALL those are restored. show all: GUC + non-GUC. SHOW ALL, RESET ALL Marko Kreen
2001-06-07 06:50:57 +02:00
case PGC_INT:
Extend GUC concepts of parse_hook and assign_hook to all four supported datatypes, not only strings. parse_hook is useless for bool, I suppose, but it seems possibly useful for int and double to apply variable-specific constraints that are more complex than simple range limits. assign_hook is definitely useful for all datatypes --- we need it right now for bool to support date cache reset when changing Australian timezone rule setting. Also, clean up some residual problems with the reset all/show all patch, including memory leaks and mistaken reset of PostPortNumber. It seems best that RESET ALL not touch variables that don't have SUSET or USERSET context.
2001-06-13 00:54:06 +02:00
snprintf(buffer, sizeof(buffer), "%d",
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
((struct config_int *) record)->reset_val);
return buffer;
reset all: command line and .conf options change defaults on RESET ALL those are restored. show all: GUC + non-GUC. SHOW ALL, RESET ALL Marko Kreen
2001-06-07 06:50:57 +02:00
case PGC_REAL:
Extend GUC concepts of parse_hook and assign_hook to all four supported datatypes, not only strings. parse_hook is useless for bool, I suppose, but it seems possibly useful for int and double to apply variable-specific constraints that are more complex than simple range limits. assign_hook is definitely useful for all datatypes --- we need it right now for bool to support date cache reset when changing Australian timezone rule setting. Also, clean up some residual problems with the reset all/show all patch, including memory leaks and mistaken reset of PostPortNumber. It seems best that RESET ALL not touch variables that don't have SUSET or USERSET context.
2001-06-13 00:54:06 +02:00
snprintf(buffer, sizeof(buffer), "%g",
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
((struct config_real *) record)->reset_val);
return buffer;
reset all: command line and .conf options change defaults on RESET ALL those are restored. show all: GUC + non-GUC. SHOW ALL, RESET ALL Marko Kreen
2001-06-07 06:50:57 +02:00
case PGC_STRING:
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
return ((struct config_string *) record)->reset_val;
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
case PGC_ENUM:
Fix postgres --describe-config for guc enums, breakage noted by Alvaro. While at it, rename option lookup functions to make names clearer, per discussion with Tom.
2008-03-17 18:45:09 +01:00
return config_enum_lookup_by_value((struct config_enum *) record,
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
((struct config_enum *) record)->reset_val);
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
}
return NULL;
}
reset all: command line and .conf options change defaults on RESET ALL those are restored. show all: GUC + non-GUC. SHOW ALL, RESET ALL Marko Kreen
2001-06-07 06:50:57 +02:00
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
/*
* flatten_set_variable_args
* Given a parsenode List as emitted by the grammar for SET,
* convert to the flat string representation used by GUC.
*
* We need to be told the name of the variable the args are for, because
* the flattening rules vary (ugh).
*
Support SET FROM CURRENT in CREATE/ALTER FUNCTION, ALTER DATABASE, ALTER ROLE. (Actually, it works as a plain statement too, but I didn't document that because it seems a bit useless.) Unify VariableResetStmt with VariableSetStmt, and clean up some ancient cruft in the representation of same.
2007-09-03 20:46:30 +02:00
* The result is NULL if args is NIL (ie, SET ... TO DEFAULT), otherwise
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
* a palloc'd string.
*/
Support SET FROM CURRENT in CREATE/ALTER FUNCTION, ALTER DATABASE, ALTER ROLE. (Actually, it works as a plain statement too, but I didn't document that because it seems a bit useless.) Unify VariableResetStmt with VariableSetStmt, and clean up some ancient cruft in the representation of same.
2007-09-03 20:46:30 +02:00
static char *
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
flatten_set_variable_args(const char *name, List *args)
{
struct config_generic *record;
int flags;
StringInfoData buf;
Reimplement the linked list data structure used throughout the backend. In the past, we used a 'Lispy' linked list implementation: a "list" was merely a pointer to the head node of the list. The problem with that design is that it makes lappend() and length() linear time. This patch fixes that problem (and others) by maintaining a count of the list length and a pointer to the tail node along with each head node pointer. A "list" is now a pointer to a structure containing some meta-data about the list; the head and tail pointers in that structure refer to ListCell structures that maintain the actual linked list of nodes. The function names of the list API have also been changed to, I hope, be more logically consistent. By default, the old function names are still available; they will be disabled-by-default once the rest of the tree has been updated to use the new API names.
2004-05-26 06:41:50 +02:00
ListCell *l;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
Support SET FROM CURRENT in CREATE/ALTER FUNCTION, ALTER DATABASE, ALTER ROLE. (Actually, it works as a plain statement too, but I didn't document that because it seems a bit useless.) Unify VariableResetStmt with VariableSetStmt, and clean up some ancient cruft in the representation of same.
2007-09-03 20:46:30 +02:00
/* Fast path if just DEFAULT */
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
if (args == NIL)
return NULL;
Enforce superuser permissions checks during ALTER ROLE/DATABASE SET, rather than during define_custom_variable(). This entails rejecting an ALTER command if the target variable doesn't have a known (non-placeholder) definition, unless the calling user is superuser. When the variable *is* known, we can correctly apply the rule that only superusers can issue ALTER for SUSET parameters. This allows define_custom_variable to apply ALTER's values for SUSET parameters at module load time, secure in the knowledge that only a superuser could have set the ALTER value. This change fixes a longstanding gotcha in the usage of SUSET-level custom parameters; which is a good thing to fix now that plpgsql defines such a parameter.
2010-04-21 22:54:19 +02:00
/*
* Get flags for the variable; if it's not known, use default flags.
* (Caller might throw error later, but not our business to do so here.)
*/
record = find_option(name, false, WARNING);
if (record)
flags = record->flags;
else
flags = 0;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
/* Complain if list input and non-list variable */
if ((flags & GUC_LIST_INPUT) == 0 &&
Reimplement the linked list data structure used throughout the backend. In the past, we used a 'Lispy' linked list implementation: a "list" was merely a pointer to the head node of the list. The problem with that design is that it makes lappend() and length() linear time. This patch fixes that problem (and others) by maintaining a count of the list length and a pointer to the tail node along with each head node pointer. A "list" is now a pointer to a structure containing some meta-data about the list; the head and tail pointers in that structure refer to ListCell structures that maintain the actual linked list of nodes. The function names of the list API have also been changed to, I hope, be more logically consistent. By default, the old function names are still available; they will be disabled-by-default once the rest of the tree has been updated to use the new API names.
2004-05-26 06:41:50 +02:00
list_length(args) != 1)
Error message editing in backend/utils (except /adt).
2003-07-25 22:18:01 +02:00
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("SET %s takes only one argument", name)));
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
initStringInfo(&buf);
Remove typename from A_Const. Brendan Jurd, minor editorialization by me.
2008-04-29 16:59:17 +02:00
/*
* Each list member may be a plain A_Const node, or an A_Const within a
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
* TypeCast; the latter case is supported only for ConstInterval arguments
* (for SET TIME ZONE).
Remove typename from A_Const. Brendan Jurd, minor editorialization by me.
2008-04-29 16:59:17 +02:00
*/
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
foreach(l, args)
{
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
Node *arg = (Node *) lfirst(l);
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
char *val;
Make backend header files C++ safe This alters various incidental uses of C++ key words to use other similar identifiers, so that a C++ compiler won't choke outright. You still (probably) need extern "C" { }; around the inclusion of backend headers. based on a patch by Kurt Harriman <harriman@acm.org> Also add a script cpluspluscheck to check for C++ compatibility in the future. As of right now, this passes without error for me.
2009-07-16 08:33:46 +02:00
TypeName *typeName = NULL;
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
A_Const *con;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
Reimplement the linked list data structure used throughout the backend. In the past, we used a 'Lispy' linked list implementation: a "list" was merely a pointer to the head node of the list. The problem with that design is that it makes lappend() and length() linear time. This patch fixes that problem (and others) by maintaining a count of the list length and a pointer to the tail node along with each head node pointer. A "list" is now a pointer to a structure containing some meta-data about the list; the head and tail pointers in that structure refer to ListCell structures that maintain the actual linked list of nodes. The function names of the list API have also been changed to, I hope, be more logically consistent. By default, the old function names are still available; they will be disabled-by-default once the rest of the tree has been updated to use the new API names.
2004-05-26 06:41:50 +02:00
if (l != list_head(args))
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
appendStringInfo(&buf, ", ");
Remove typename from A_Const. Brendan Jurd, minor editorialization by me.
2008-04-29 16:59:17 +02:00
if (IsA(arg, TypeCast))
{
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
TypeCast *tc = (TypeCast *) arg;
Remove typename from A_Const. Brendan Jurd, minor editorialization by me.
2008-04-29 16:59:17 +02:00
arg = tc->arg;
Make backend header files C++ safe This alters various incidental uses of C++ key words to use other similar identifiers, so that a C++ compiler won't choke outright. You still (probably) need extern "C" { }; around the inclusion of backend headers. based on a patch by Kurt Harriman <harriman@acm.org> Also add a script cpluspluscheck to check for C++ compatibility in the future. As of right now, this passes without error for me.
2009-07-16 08:33:46 +02:00
typeName = tc->typeName;
Remove typename from A_Const. Brendan Jurd, minor editorialization by me.
2008-04-29 16:59:17 +02:00
}
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
if (!IsA(arg, A_Const))
Error message editing in backend/utils (except /adt).
2003-07-25 22:18:01 +02:00
elog(ERROR, "unrecognized node type: %d", (int) nodeTag(arg));
Remove typename from A_Const. Brendan Jurd, minor editorialization by me.
2008-04-29 16:59:17 +02:00
con = (A_Const *) arg;
Some minor further cleanup around A_Const. Don't attach a typecast in makeFloatConst, and avoid "manual" construction of A_Const nodes in grammar productions, in favor of using makeXXXConst subroutines.
2008-04-29 22:44:49 +02:00
Remove typename from A_Const. Brendan Jurd, minor editorialization by me.
2008-04-29 16:59:17 +02:00
switch (nodeTag(&con->val))
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
{
case T_Integer:
Remove typename from A_Const. Brendan Jurd, minor editorialization by me.
2008-04-29 16:59:17 +02:00
appendStringInfo(&buf, "%ld", intVal(&con->val));
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
break;
case T_Float:
/* represented as a string, so just copy it */
Remove typename from A_Const. Brendan Jurd, minor editorialization by me.
2008-04-29 16:59:17 +02:00
appendStringInfoString(&buf, strVal(&con->val));
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
break;
case T_String:
Remove typename from A_Const. Brendan Jurd, minor editorialization by me.
2008-04-29 16:59:17 +02:00
val = strVal(&con->val);
Make backend header files C++ safe This alters various incidental uses of C++ key words to use other similar identifiers, so that a C++ compiler won't choke outright. You still (probably) need extern "C" { }; around the inclusion of backend headers. based on a patch by Kurt Harriman <harriman@acm.org> Also add a script cpluspluscheck to check for C++ compatibility in the future. As of right now, this passes without error for me.
2009-07-16 08:33:46 +02:00
if (typeName != NULL)
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
{
/*
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
* Must be a ConstInterval argument for TIME ZONE. Coerce
* to interval and back to normalize the value and account
* for any typmod.
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
*/
Ensure that typmod decoration on a datatype name is validated in all cases, even in code paths where we don't pay any subsequent attention to the typmod value. This seems needed in view of the fact that 8.3's generalized typmod support will accept a lot of bogus syntax, such as "timestamp(foo)" or "record(int, 42)" --- if we allow such things to pass without comment, users will get confused. Per a recent example from Greg Stark. To implement this in a way that's not very vulnerable to future bugs-of-omission, refactor the API of parse_type.c's TypeName lookup routines so that typmod validation is folded into the base lookup operation. Callers can still choose not to receive the encoded typmod, but we'll check the decoration anyway if it's present.
2007-11-11 20:22:49 +01:00
Oid typoid;
Support type modifiers for user-defined types, and pull most knowledge about typmod representation for standard types out into type-specific typmod I/O functions. Teodor Sigaev, with some editorialization by Tom Lane.
2006-12-30 22:21:56 +01:00
int32 typmod;
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
Datum interval;
pgindent run.
2002-09-04 22:31:48 +02:00
char *intervalout;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
Refactor typenameTypeId() Split the old typenameTypeId() into two functions: A new typenameTypeId() that returns only a type OID, and typenameTypeIdAndMod() that returns type OID and typmod. This isolates call sites better that actually care about the typmod.
2010-10-25 20:40:46 +02:00
typenameTypeIdAndMod(NULL, typeName, &typoid, &typmod);
Ensure that typmod decoration on a datatype name is validated in all cases, even in code paths where we don't pay any subsequent attention to the typmod value. This seems needed in view of the fact that 8.3's generalized typmod support will accept a lot of bogus syntax, such as "timestamp(foo)" or "record(int, 42)" --- if we allow such things to pass without comment, users will get confused. Per a recent example from Greg Stark. To implement this in a way that's not very vulnerable to future bugs-of-omission, refactor the API of parse_type.c's TypeName lookup routines so that typmod validation is folded into the base lookup operation. Callers can still choose not to receive the encoded typmod, but we'll check the decoration anyway if it's present.
2007-11-11 20:22:49 +01:00
Assert(typoid == INTERVALOID);
Support type modifiers for user-defined types, and pull most knowledge about typmod representation for standard types out into type-specific typmod I/O functions. Teodor Sigaev, with some editorialization by Tom Lane.
2006-12-30 22:21:56 +01:00
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
interval =
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
DirectFunctionCall3(interval_in,
CStringGetDatum(val),
ObjectIdGetDatum(InvalidOid),
Support type modifiers for user-defined types, and pull most knowledge about typmod representation for standard types out into type-specific typmod I/O functions. Teodor Sigaev, with some editorialization by Tom Lane.
2006-12-30 22:21:56 +01:00
Int32GetDatum(typmod));
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
intervalout =
Change CREATE TYPE to require datatype output and send functions to have only one argument. (Per recent discussion, the option to accept multiple arguments is pretty useless for user-defined types, and would be a likely source of security holes if it was used.) Simplify call sites of output/send functions to not bother passing more than one argument.
2005-05-01 20:56:19 +02:00
DatumGetCString(DirectFunctionCall1(interval_out,
interval));
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
appendStringInfo(&buf, "INTERVAL '%s'", intervalout);
}
else
{
Some minor further cleanup around A_Const. Don't attach a typecast in makeFloatConst, and avoid "manual" construction of A_Const nodes in grammar productions, in favor of using makeXXXConst subroutines.
2008-04-29 22:44:49 +02:00
/*
* Plain string literal or identifier. For quote mode,
* quote it if it's not a vanilla identifier.
*/
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
if (flags & GUC_LIST_QUOTE)
Micro-opt: replace calls like appendStringInfo(buf, "%s", str); with appendStringInfoString(buf, str); as the latter form is slightly faster.
2004-01-31 06:09:41 +01:00
appendStringInfoString(&buf, quote_identifier(val));
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
else
Micro-opt: replace calls like appendStringInfo(buf, "%s", str); with appendStringInfoString(buf, str); as the latter form is slightly faster.
2004-01-31 06:09:41 +01:00
appendStringInfoString(&buf, val);
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
}
break;
default:
Error message editing in backend/utils (except /adt).
2003-07-25 22:18:01 +02:00
elog(ERROR, "unrecognized node type: %d",
Remove typename from A_Const. Brendan Jurd, minor editorialization by me.
2008-04-29 16:59:17 +02:00
(int) nodeTag(&con->val));
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
break;
}
reset all: command line and .conf options change defaults on RESET ALL those are restored. show all: GUC + non-GUC. SHOW ALL, RESET ALL Marko Kreen
2001-06-07 06:50:57 +02:00
}
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
return buf.data;
}
/*
* SET command
*/
void
Support SET FROM CURRENT in CREATE/ALTER FUNCTION, ALTER DATABASE, ALTER ROLE. (Actually, it works as a plain statement too, but I didn't document that because it seems a bit useless.) Unify VariableResetStmt with VariableSetStmt, and clean up some ancient cruft in the representation of same.
2007-09-03 20:46:30 +02:00
ExecSetVariableStmt(VariableSetStmt *stmt)
{
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
GucAction action = stmt->is_local ? GUC_ACTION_LOCAL : GUC_ACTION_SET;
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
Support SET FROM CURRENT in CREATE/ALTER FUNCTION, ALTER DATABASE, ALTER ROLE. (Actually, it works as a plain statement too, but I didn't document that because it seems a bit useless.) Unify VariableResetStmt with VariableSetStmt, and clean up some ancient cruft in the representation of same.
2007-09-03 20:46:30 +02:00
switch (stmt->kind)
{
case VAR_SET_VALUE:
case VAR_SET_CURRENT:
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
(void) set_config_option(stmt->name,
ExtractSetVariableArgs(stmt),
(superuser() ? PGC_SUSET : PGC_USERSET),
PGC_S_SESSION,
action,
true);
Support SET FROM CURRENT in CREATE/ALTER FUNCTION, ALTER DATABASE, ALTER ROLE. (Actually, it works as a plain statement too, but I didn't document that because it seems a bit useless.) Unify VariableResetStmt with VariableSetStmt, and clean up some ancient cruft in the representation of same.
2007-09-03 20:46:30 +02:00
break;
case VAR_SET_MULTI:
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
Support SET FROM CURRENT in CREATE/ALTER FUNCTION, ALTER DATABASE, ALTER ROLE. (Actually, it works as a plain statement too, but I didn't document that because it seems a bit useless.) Unify VariableResetStmt with VariableSetStmt, and clean up some ancient cruft in the representation of same.
2007-09-03 20:46:30 +02:00
/*
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
* Special case for special SQL syntax that effectively sets more
* than one variable per statement.
Support SET FROM CURRENT in CREATE/ALTER FUNCTION, ALTER DATABASE, ALTER ROLE. (Actually, it works as a plain statement too, but I didn't document that because it seems a bit useless.) Unify VariableResetStmt with VariableSetStmt, and clean up some ancient cruft in the representation of same.
2007-09-03 20:46:30 +02:00
*/
if (strcmp(stmt->name, "TRANSACTION") == 0)
{
ListCell *head;
foreach(head, stmt->args)
{
DefElem *item = (DefElem *) lfirst(head);
if (strcmp(item->defname, "transaction_isolation") == 0)
SetPGVariable("transaction_isolation",
list_make1(item->arg), stmt->is_local);
else if (strcmp(item->defname, "transaction_read_only") == 0)
SetPGVariable("transaction_read_only",
list_make1(item->arg), stmt->is_local);
Implement genuine serializable isolation level. Until now, our Serializable mode has in fact been what's called Snapshot Isolation, which allows some anomalies that could not occur in any serialized ordering of the transactions. This patch fixes that using a method called Serializable Snapshot Isolation, based on research papers by Michael J. Cahill (see README-SSI for full references). In Serializable Snapshot Isolation, transactions run like they do in Snapshot Isolation, but a predicate lock manager observes the reads and writes performed and aborts transactions if it detects that an anomaly might occur. This method produces some false positives, ie. it sometimes aborts transactions even though there is no anomaly. To track reads we implement predicate locking, see storage/lmgr/predicate.c. Whenever a tuple is read, a predicate lock is acquired on the tuple. Shared memory is finite, so when a transaction takes many tuple-level locks on a page, the locks are promoted to a single page-level lock, and further to a single relation level lock if necessary. To lock key values with no matching tuple, a sequential scan always takes a relation-level lock, and an index scan acquires a page-level lock that covers the search key, whether or not there are any matching keys at the moment. A predicate lock doesn't conflict with any regular locks or with another predicate locks in the normal sense. They're only used by the predicate lock manager to detect the danger of anomalies. Only serializable transactions participate in predicate locking, so there should be no extra overhead for for other transactions. Predicate locks can't be released at commit, but must be remembered until all the transactions that overlapped with it have completed. That means that we need to remember an unbounded amount of predicate locks, so we apply a lossy but conservative method of tracking locks for committed transactions. If we run short of shared memory, we overflow to a new "pg_serial" SLRU pool. We don't currently allow Serializable transactions in Hot Standby mode. That would be hard, because even read-only transactions can cause anomalies that wouldn't otherwise occur. Serializable isolation mode now means the new fully serializable level. Repeatable Read gives you the old Snapshot Isolation level that we have always had. Kevin Grittner and Dan Ports, reviewed by Jeff Davis, Heikki Linnakangas and Anssi Kääriäinen
2011-02-07 22:46:51 +01:00
else if (strcmp(item->defname, "transaction_deferrable") == 0)
SetPGVariable("transaction_deferrable",
list_make1(item->arg), stmt->is_local);
Support SET FROM CURRENT in CREATE/ALTER FUNCTION, ALTER DATABASE, ALTER ROLE. (Actually, it works as a plain statement too, but I didn't document that because it seems a bit useless.) Unify VariableResetStmt with VariableSetStmt, and clean up some ancient cruft in the representation of same.
2007-09-03 20:46:30 +02:00
else
elog(ERROR, "unexpected SET TRANSACTION element: %s",
item->defname);
}
}
else if (strcmp(stmt->name, "SESSION CHARACTERISTICS") == 0)
{
ListCell *head;
foreach(head, stmt->args)
{
DefElem *item = (DefElem *) lfirst(head);
if (strcmp(item->defname, "transaction_isolation") == 0)
SetPGVariable("default_transaction_isolation",
list_make1(item->arg), stmt->is_local);
else if (strcmp(item->defname, "transaction_read_only") == 0)
SetPGVariable("default_transaction_read_only",
list_make1(item->arg), stmt->is_local);
Implement genuine serializable isolation level. Until now, our Serializable mode has in fact been what's called Snapshot Isolation, which allows some anomalies that could not occur in any serialized ordering of the transactions. This patch fixes that using a method called Serializable Snapshot Isolation, based on research papers by Michael J. Cahill (see README-SSI for full references). In Serializable Snapshot Isolation, transactions run like they do in Snapshot Isolation, but a predicate lock manager observes the reads and writes performed and aborts transactions if it detects that an anomaly might occur. This method produces some false positives, ie. it sometimes aborts transactions even though there is no anomaly. To track reads we implement predicate locking, see storage/lmgr/predicate.c. Whenever a tuple is read, a predicate lock is acquired on the tuple. Shared memory is finite, so when a transaction takes many tuple-level locks on a page, the locks are promoted to a single page-level lock, and further to a single relation level lock if necessary. To lock key values with no matching tuple, a sequential scan always takes a relation-level lock, and an index scan acquires a page-level lock that covers the search key, whether or not there are any matching keys at the moment. A predicate lock doesn't conflict with any regular locks or with another predicate locks in the normal sense. They're only used by the predicate lock manager to detect the danger of anomalies. Only serializable transactions participate in predicate locking, so there should be no extra overhead for for other transactions. Predicate locks can't be released at commit, but must be remembered until all the transactions that overlapped with it have completed. That means that we need to remember an unbounded amount of predicate locks, so we apply a lossy but conservative method of tracking locks for committed transactions. If we run short of shared memory, we overflow to a new "pg_serial" SLRU pool. We don't currently allow Serializable transactions in Hot Standby mode. That would be hard, because even read-only transactions can cause anomalies that wouldn't otherwise occur. Serializable isolation mode now means the new fully serializable level. Repeatable Read gives you the old Snapshot Isolation level that we have always had. Kevin Grittner and Dan Ports, reviewed by Jeff Davis, Heikki Linnakangas and Anssi Kääriäinen
2011-02-07 22:46:51 +01:00
else if (strcmp(item->defname, "transaction_deferrable") == 0)
SetPGVariable("default_transaction_deferrable",
list_make1(item->arg), stmt->is_local);
Support SET FROM CURRENT in CREATE/ALTER FUNCTION, ALTER DATABASE, ALTER ROLE. (Actually, it works as a plain statement too, but I didn't document that because it seems a bit useless.) Unify VariableResetStmt with VariableSetStmt, and clean up some ancient cruft in the representation of same.
2007-09-03 20:46:30 +02:00
else
elog(ERROR, "unexpected SET SESSION element: %s",
item->defname);
}
}
else
elog(ERROR, "unexpected SET MULTI element: %s",
stmt->name);
break;
case VAR_SET_DEFAULT:
case VAR_RESET:
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
(void) set_config_option(stmt->name,
NULL,
(superuser() ? PGC_SUSET : PGC_USERSET),
PGC_S_SESSION,
action,
true);
Support SET FROM CURRENT in CREATE/ALTER FUNCTION, ALTER DATABASE, ALTER ROLE. (Actually, it works as a plain statement too, but I didn't document that because it seems a bit useless.) Unify VariableResetStmt with VariableSetStmt, and clean up some ancient cruft in the representation of same.
2007-09-03 20:46:30 +02:00
break;
case VAR_RESET_ALL:
ResetAllOptions();
break;
}
}
/*
* Get the value to assign for a VariableSetStmt, or NULL if it's RESET.
* The result is palloc'd.
*
* This is exported for use by actions such as ALTER ROLE SET.
*/
char *
ExtractSetVariableArgs(VariableSetStmt *stmt)
{
switch (stmt->kind)
{
case VAR_SET_VALUE:
return flatten_set_variable_args(stmt->name, stmt->args);
case VAR_SET_CURRENT:
return GetConfigOptionByName(stmt->name, NULL);
default:
return NULL;
}
}
/*
* SetPGVariable - SET command exported as an easily-C-callable function.
*
* This provides access to SET TO value, as well as SET TO DEFAULT (expressed
* by passing args == NIL), but not SET FROM CURRENT functionality.
*/
void
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
SetPGVariable(const char *name, List *args, bool is_local)
{
char *argstring = flatten_set_variable_args(name, args);
/* Note SET DEFAULT (argstring == NULL) is equivalent to RESET */
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
(void) set_config_option(name,
argstring,
(superuser() ? PGC_SUSET : PGC_USERSET),
PGC_S_SESSION,
is_local ? GUC_ACTION_LOCAL : GUC_ACTION_SET,
true);
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
}
> 2. This patch includes the same Table Function API fixes that I > submitted on July 9: > > http://archives.postgresql.org/pgsql-patches/2002-07/msg00056.php > > Please disregard that one *if* this one is applied. If this one is > rejected please go ahead with the July 9th patch. The July 9th Table Function API patch mentioned above is now in CVS, so here is an updated version of the guc patch which should apply cleanly against CVS tip. Joe Conway
2002-07-20 07:49:28 +02:00
/*
* SET command wrapped as a SQL callable function.
*/
Datum
set_config_by_name(PG_FUNCTION_ARGS)
{
pgindent run.
2002-09-04 22:31:48 +02:00
char *name;
char *value;
char *new_value;
bool is_local;
> 2. This patch includes the same Table Function API fixes that I > submitted on July 9: > > http://archives.postgresql.org/pgsql-patches/2002-07/msg00056.php > > Please disregard that one *if* this one is applied. If this one is > rejected please go ahead with the July 9th patch. The July 9th Table Function API patch mentioned above is now in CVS, so here is an updated version of the guc patch which should apply cleanly against CVS tip. Joe Conway
2002-07-20 07:49:28 +02:00
if (PG_ARGISNULL(0))
Error message editing in backend/utils (except /adt).
2003-07-25 22:18:01 +02:00
ereport(ERROR,
(errcode(ERRCODE_NULL_VALUE_NOT_ALLOWED),
Message editing: remove gratuitous variations in message wording, standardize terms, add some clarifications, fix some untranslatable attempts at dynamic message building.
2003-09-25 08:58:07 +02:00
errmsg("SET requires parameter name")));
> 2. This patch includes the same Table Function API fixes that I > submitted on July 9: > > http://archives.postgresql.org/pgsql-patches/2002-07/msg00056.php > > Please disregard that one *if* this one is applied. If this one is > rejected please go ahead with the July 9th patch. The July 9th Table Function API patch mentioned above is now in CVS, so here is an updated version of the guc patch which should apply cleanly against CVS tip. Joe Conway
2002-07-20 07:49:28 +02:00
/* Get the GUC variable name */
Simplify and standardize conversions between TEXT datums and ordinary C strings. This patch introduces four support functions cstring_to_text, cstring_to_text_with_len, text_to_cstring, and text_to_cstring_buffer, and two macros CStringGetTextDatum and TextDatumGetCString. A number of existing macros that provided variants on these themes were removed. Most of the places that need to make such conversions now require just one function or macro call, in place of the multiple notational layers that used to be needed. There are no longer any direct calls of textout or textin, and we got most of the places that were using handmade conversions via memcpy (there may be a few still lurking, though). This commit doesn't make any serious effort to eliminate transient memory leaks caused by detoasting toasted text objects before they reach text_to_cstring. We changed PG_GETARG_TEXT_P to PG_GETARG_TEXT_PP in a few places where it was easy, but much more could be done. Brendan Jurd and Tom Lane
2008-03-25 23:42:46 +01:00
name = TextDatumGetCString(PG_GETARG_DATUM(0));
> 2. This patch includes the same Table Function API fixes that I > submitted on July 9: > > http://archives.postgresql.org/pgsql-patches/2002-07/msg00056.php > > Please disregard that one *if* this one is applied. If this one is > rejected please go ahead with the July 9th patch. The July 9th Table Function API patch mentioned above is now in CVS, so here is an updated version of the guc patch which should apply cleanly against CVS tip. Joe Conway
2002-07-20 07:49:28 +02:00
/* Get the desired value or set to NULL for a reset request */
if (PG_ARGISNULL(1))
value = NULL;
else
Simplify and standardize conversions between TEXT datums and ordinary C strings. This patch introduces four support functions cstring_to_text, cstring_to_text_with_len, text_to_cstring, and text_to_cstring_buffer, and two macros CStringGetTextDatum and TextDatumGetCString. A number of existing macros that provided variants on these themes were removed. Most of the places that need to make such conversions now require just one function or macro call, in place of the multiple notational layers that used to be needed. There are no longer any direct calls of textout or textin, and we got most of the places that were using handmade conversions via memcpy (there may be a few still lurking, though). This commit doesn't make any serious effort to eliminate transient memory leaks caused by detoasting toasted text objects before they reach text_to_cstring. We changed PG_GETARG_TEXT_P to PG_GETARG_TEXT_PP in a few places where it was easy, but much more could be done. Brendan Jurd and Tom Lane
2008-03-25 23:42:46 +01:00
value = TextDatumGetCString(PG_GETARG_DATUM(1));
> 2. This patch includes the same Table Function API fixes that I > submitted on July 9: > > http://archives.postgresql.org/pgsql-patches/2002-07/msg00056.php > > Please disregard that one *if* this one is applied. If this one is > rejected please go ahead with the July 9th patch. The July 9th Table Function API patch mentioned above is now in CVS, so here is an updated version of the guc patch which should apply cleanly against CVS tip. Joe Conway
2002-07-20 07:49:28 +02:00
/*
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
* Get the desired state of is_local. Default to false if provided value
* is NULL
> 2. This patch includes the same Table Function API fixes that I > submitted on July 9: > > http://archives.postgresql.org/pgsql-patches/2002-07/msg00056.php > > Please disregard that one *if* this one is applied. If this one is > rejected please go ahead with the July 9th patch. The July 9th Table Function API patch mentioned above is now in CVS, so here is an updated version of the guc patch which should apply cleanly against CVS tip. Joe Conway
2002-07-20 07:49:28 +02:00
*/
if (PG_ARGISNULL(2))
is_local = false;
else
is_local = PG_GETARG_BOOL(2);
/* Note SET DEFAULT (argstring == NULL) is equivalent to RESET */
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
(void) set_config_option(name,
value,
(superuser() ? PGC_SUSET : PGC_USERSET),
PGC_S_SESSION,
is_local ? GUC_ACTION_LOCAL : GUC_ACTION_SET,
true);
> 2. This patch includes the same Table Function API fixes that I > submitted on July 9: > > http://archives.postgresql.org/pgsql-patches/2002-07/msg00056.php > > Please disregard that one *if* this one is applied. If this one is > rejected please go ahead with the July 9th patch. The July 9th Table Function API patch mentioned above is now in CVS, so here is an updated version of the guc patch which should apply cleanly against CVS tip. Joe Conway
2002-07-20 07:49:28 +02:00
/* get the new current value */
Code review for SHOW output changes; fix horology expected files for new SHOW output format.
2002-07-20 17:12:56 +02:00
new_value = GetConfigOptionByName(name, NULL);
> 2. This patch includes the same Table Function API fixes that I > submitted on July 9: > > http://archives.postgresql.org/pgsql-patches/2002-07/msg00056.php > > Please disregard that one *if* this one is applied. If this one is > rejected please go ahead with the July 9th patch. The July 9th Table Function API patch mentioned above is now in CVS, so here is an updated version of the guc patch which should apply cleanly against CVS tip. Joe Conway
2002-07-20 07:49:28 +02:00
/* Convert return string to text */
Simplify and standardize conversions between TEXT datums and ordinary C strings. This patch introduces four support functions cstring_to_text, cstring_to_text_with_len, text_to_cstring, and text_to_cstring_buffer, and two macros CStringGetTextDatum and TextDatumGetCString. A number of existing macros that provided variants on these themes were removed. Most of the places that need to make such conversions now require just one function or macro call, in place of the multiple notational layers that used to be needed. There are no longer any direct calls of textout or textin, and we got most of the places that were using handmade conversions via memcpy (there may be a few still lurking, though). This commit doesn't make any serious effort to eliminate transient memory leaks caused by detoasting toasted text objects before they reach text_to_cstring. We changed PG_GETARG_TEXT_P to PG_GETARG_TEXT_PP in a few places where it was easy, but much more could be done. Brendan Jurd and Tom Lane
2008-03-25 23:42:46 +01:00
PG_RETURN_TEXT_P(cstring_to_text(new_value));
> 2. This patch includes the same Table Function API fixes that I > submitted on July 9: > > http://archives.postgresql.org/pgsql-patches/2002-07/msg00056.php > > Please disregard that one *if* this one is applied. If this one is > rejected please go ahead with the July 9th patch. The July 9th Table Function API patch mentioned above is now in CVS, so here is an updated version of the guc patch which should apply cleanly against CVS tip. Joe Conway
2002-07-20 07:49:28 +02:00
}
Support SET FROM CURRENT in CREATE/ALTER FUNCTION, ALTER DATABASE, ALTER ROLE. (Actually, it works as a plain statement too, but I didn't document that because it seems a bit useless.) Unify VariableResetStmt with VariableSetStmt, and clean up some ancient cruft in the representation of same.
2007-09-03 20:46:30 +02:00
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
/*
* Common code for DefineCustomXXXVariable subroutines: allocate the
* new variable's config struct and fill in generic fields.
*/
static struct config_generic *
init_custom_variable(const char *name,
const char *short_desc,
const char *long_desc,
GucContext context,
Some infrastructure changes for the upcoming auto-explain contrib module: * Refactor explain.c slightly to export a convenient-to-use subroutine for printing EXPLAIN results. * Provide hooks for plugins to get control at ExecutorStart and ExecutorEnd as well as ExecutorRun. * Add some minimal support for tracking the total runtime of ExecutorRun. This code won't actually do anything unless a plugin prods it to. * Change the API of the DefineCustomXXXVariable functions to allow nonzero "flags" to be specified for a custom GUC variable. While at it, also make the "bootstrap" default value for custom GUCs be explicitly specified as a parameter to these functions. This is to eliminate confusion over where the default comes from, as has been expressed in the past by some users of the custom-variable facility. * Refactor GUC code a bit to ensure that a custom variable gets initialized to something valid (like its default value) even if the placeholder value was invalid.
2008-11-19 02:10:24 +01:00
int flags,
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
enum config_type type,
size_t sz)
{
struct config_generic *gen;
Allow loadable modules to create PGC_POSTMASTER GUC variables, but only when loaded via shared_preload_libraries. Needed for support of pg_stat_statements, or pretty much anything else that wants a GUC to control size of a shared memory allocation.
2009-01-03 21:03:08 +01:00
/*
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
* Only allow custom PGC_POSTMASTER variables to be created during shared
* library preload; any later than that, we can't ensure that the value
* doesn't change after startup. This is a fatal elog if it happens; just
* erroring out isn't safe because we don't know what the calling loadable
* module might already have hooked into.
Allow loadable modules to create PGC_POSTMASTER GUC variables, but only when loaded via shared_preload_libraries. Needed for support of pg_stat_statements, or pretty much anything else that wants a GUC to control size of a shared memory allocation.
2009-01-03 21:03:08 +01:00
*/
if (context == PGC_POSTMASTER &&
!process_shared_preload_libraries_in_progress)
elog(FATAL, "cannot create PGC_POSTMASTER variables after startup");
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
gen = (struct config_generic *) guc_malloc(ERROR, sz);
memset(gen, 0, sz);
gen->name = guc_strdup(ERROR, name);
gen->context = context;
gen->group = CUSTOM_OPTIONS;
gen->short_desc = short_desc;
gen->long_desc = long_desc;
Some infrastructure changes for the upcoming auto-explain contrib module: * Refactor explain.c slightly to export a convenient-to-use subroutine for printing EXPLAIN results. * Provide hooks for plugins to get control at ExecutorStart and ExecutorEnd as well as ExecutorRun. * Add some minimal support for tracking the total runtime of ExecutorRun. This code won't actually do anything unless a plugin prods it to. * Change the API of the DefineCustomXXXVariable functions to allow nonzero "flags" to be specified for a custom GUC variable. While at it, also make the "bootstrap" default value for custom GUCs be explicitly specified as a parameter to these functions. This is to eliminate confusion over where the default comes from, as has been expressed in the past by some users of the custom-variable facility. * Refactor GUC code a bit to ensure that a custom variable gets initialized to something valid (like its default value) even if the placeholder value was invalid.
2008-11-19 02:10:24 +01:00
gen->flags = flags;
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
gen->vartype = type;
return gen;
}
/*
* Common code for DefineCustomXXXVariable subroutines: insert the new
* variable into the GUC variable array, replacing any placeholder.
*/
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
static void
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
define_custom_variable(struct config_generic * variable)
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
{
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
const char *name = variable->name;
const char **nameAddr = &name;
const char *value;
struct config_string *pHolder;
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
struct config_generic **res;
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
Some infrastructure changes for the upcoming auto-explain contrib module: * Refactor explain.c slightly to export a convenient-to-use subroutine for printing EXPLAIN results. * Provide hooks for plugins to get control at ExecutorStart and ExecutorEnd as well as ExecutorRun. * Add some minimal support for tracking the total runtime of ExecutorRun. This code won't actually do anything unless a plugin prods it to. * Change the API of the DefineCustomXXXVariable functions to allow nonzero "flags" to be specified for a custom GUC variable. While at it, also make the "bootstrap" default value for custom GUCs be explicitly specified as a parameter to these functions. This is to eliminate confusion over where the default comes from, as has been expressed in the past by some users of the custom-variable facility. * Refactor GUC code a bit to ensure that a custom variable gets initialized to something valid (like its default value) even if the placeholder value was invalid.
2008-11-19 02:10:24 +01:00
/*
* See if there's a placeholder by the same name.
*/
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
res = (struct config_generic **) bsearch((void *) &nameAddr,
(void *) guc_variables,
num_guc_variables,
sizeof(struct config_generic *),
guc_var_compare);
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
if (res == NULL)
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
{
Some infrastructure changes for the upcoming auto-explain contrib module: * Refactor explain.c slightly to export a convenient-to-use subroutine for printing EXPLAIN results. * Provide hooks for plugins to get control at ExecutorStart and ExecutorEnd as well as ExecutorRun. * Add some minimal support for tracking the total runtime of ExecutorRun. This code won't actually do anything unless a plugin prods it to. * Change the API of the DefineCustomXXXVariable functions to allow nonzero "flags" to be specified for a custom GUC variable. While at it, also make the "bootstrap" default value for custom GUCs be explicitly specified as a parameter to these functions. This is to eliminate confusion over where the default comes from, as has been expressed in the past by some users of the custom-variable facility. * Refactor GUC code a bit to ensure that a custom variable gets initialized to something valid (like its default value) even if the placeholder value was invalid.
2008-11-19 02:10:24 +01:00
/*
* No placeholder to replace, so we can just add it ... but first,
* make sure it's initialized to its default value.
*/
InitializeOneGUCOption(variable);
Fix unchecked mallocs/strdups added by recent placeholder-config-vars patch. Thomas Hallgren
2004-07-06 01:14:14 +02:00
add_guc_variable(variable, ERROR);
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
return;
}
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
/*
* This better be a placeholder
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
*/
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
if (((*res)->flags & GUC_CUSTOM_PLACEHOLDER) == 0)
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
ereport(ERROR,
(errcode(ERRCODE_INTERNAL_ERROR),
errmsg("attempt to redefine parameter \"%s\"", name)));
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
Assert((*res)->vartype == PGC_STRING);
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
pHolder = (struct config_string *) (*res);
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
Some infrastructure changes for the upcoming auto-explain contrib module: * Refactor explain.c slightly to export a convenient-to-use subroutine for printing EXPLAIN results. * Provide hooks for plugins to get control at ExecutorStart and ExecutorEnd as well as ExecutorRun. * Add some minimal support for tracking the total runtime of ExecutorRun. This code won't actually do anything unless a plugin prods it to. * Change the API of the DefineCustomXXXVariable functions to allow nonzero "flags" to be specified for a custom GUC variable. While at it, also make the "bootstrap" default value for custom GUCs be explicitly specified as a parameter to these functions. This is to eliminate confusion over where the default comes from, as has been expressed in the past by some users of the custom-variable facility. * Refactor GUC code a bit to ensure that a custom variable gets initialized to something valid (like its default value) even if the placeholder value was invalid.
2008-11-19 02:10:24 +01:00
/*
* First, set the variable to its default value. We must do this even
* though we intend to immediately apply a new value, since it's possible
* that the new value is invalid.
*/
InitializeOneGUCOption(variable);
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
/*
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
* Replace the placeholder. We aren't changing the name, so no re-sorting
* is necessary
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
*/
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
*res = variable;
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
/*
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
* Assign the string value stored in the placeholder to the real variable.
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
*
* XXX this is not really good enough --- it should be a nontransactional
* assignment, since we don't want it to roll back if the current xact
Some infrastructure changes for the upcoming auto-explain contrib module: * Refactor explain.c slightly to export a convenient-to-use subroutine for printing EXPLAIN results. * Provide hooks for plugins to get control at ExecutorStart and ExecutorEnd as well as ExecutorRun. * Add some minimal support for tracking the total runtime of ExecutorRun. This code won't actually do anything unless a plugin prods it to. * Change the API of the DefineCustomXXXVariable functions to allow nonzero "flags" to be specified for a custom GUC variable. While at it, also make the "bootstrap" default value for custom GUCs be explicitly specified as a parameter to these functions. This is to eliminate confusion over where the default comes from, as has been expressed in the past by some users of the custom-variable facility. * Refactor GUC code a bit to ensure that a custom variable gets initialized to something valid (like its default value) even if the placeholder value was invalid.
2008-11-19 02:10:24 +01:00
* fails later. (Or do we?)
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
*/
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
value = *pHolder->variable;
if (value)
Fix define_custom_variable to preserve any sourcefile/sourceline information present in the placeholder. Noted while testing pg_stat_statements.
2009-01-05 14:23:33 +01:00
{
if (set_config_option(name, value,
Remember the source GucContext for each GUC parameter. We used to just remember the GucSource, but saving GucContext too provides a little more information --- notably, whether a SET was done by a superuser or regular user. This allows us to rip out the fairly dodgy code that define_custom_variable used to use to try to infer the context to re-install a pre-existing setting with. In particular, it now works for a superuser to SET a extension's SUSET custom variable before loading the associated extension, because GUC can remember whether the SET was done as a superuser or not. The plperl regression tests contain an example where this is useful.
2011-10-04 22:13:16 +02:00
pHolder->gen.scontext, pHolder->gen.source,
GUC_ACTION_SET, true) != 0)
Fix define_custom_variable to preserve any sourcefile/sourceline information present in the placeholder. Noted while testing pg_stat_statements.
2009-01-05 14:23:33 +01:00
{
/* Also copy over any saved source-location information */
if (pHolder->gen.sourcefile)
set_config_sourcefile(name, pHolder->gen.sourcefile,
pHolder->gen.sourceline);
}
}
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
/*
* Free up as much as we conveniently can of the placeholder structure
* (this neglects any stack items...)
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
*/
Nested transactions. There is still much left to do, especially on the performance front, but with feature freeze upon us I think it's time to drive a stake in the ground and say that this will be in 7.5. Alvaro Herrera, with some help from Tom Lane.
2004-07-01 02:52:04 +02:00
set_string_field(pHolder, pHolder->variable, NULL);
set_string_field(pHolder, &pHolder->reset_val, NULL);
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
free(pHolder);
}
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
void
Add SET ROLE. This is a partial commit of Stephen Frost's recent patch; I'm still working on the has_role function and information_schema changes.
2005-07-26 00:12:34 +02:00
DefineCustomBoolVariable(const char *name,
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
const char *short_desc,
const char *long_desc,
bool *valueAddr,
Some infrastructure changes for the upcoming auto-explain contrib module: * Refactor explain.c slightly to export a convenient-to-use subroutine for printing EXPLAIN results. * Provide hooks for plugins to get control at ExecutorStart and ExecutorEnd as well as ExecutorRun. * Add some minimal support for tracking the total runtime of ExecutorRun. This code won't actually do anything unless a plugin prods it to. * Change the API of the DefineCustomXXXVariable functions to allow nonzero "flags" to be specified for a custom GUC variable. While at it, also make the "bootstrap" default value for custom GUCs be explicitly specified as a parameter to these functions. This is to eliminate confusion over where the default comes from, as has been expressed in the past by some users of the custom-variable facility. * Refactor GUC code a bit to ensure that a custom variable gets initialized to something valid (like its default value) even if the placeholder value was invalid.
2008-11-19 02:10:24 +01:00
bool bootValue,
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
GucContext context,
Some infrastructure changes for the upcoming auto-explain contrib module: * Refactor explain.c slightly to export a convenient-to-use subroutine for printing EXPLAIN results. * Provide hooks for plugins to get control at ExecutorStart and ExecutorEnd as well as ExecutorRun. * Add some minimal support for tracking the total runtime of ExecutorRun. This code won't actually do anything unless a plugin prods it to. * Change the API of the DefineCustomXXXVariable functions to allow nonzero "flags" to be specified for a custom GUC variable. While at it, also make the "bootstrap" default value for custom GUCs be explicitly specified as a parameter to these functions. This is to eliminate confusion over where the default comes from, as has been expressed in the past by some users of the custom-variable facility. * Refactor GUC code a bit to ensure that a custom variable gets initialized to something valid (like its default value) even if the placeholder value was invalid.
2008-11-19 02:10:24 +01:00
int flags,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
GucBoolCheckHook check_hook,
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
GucBoolAssignHook assign_hook,
GucShowHook show_hook)
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
{
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
struct config_bool *var;
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
var = (struct config_bool *)
Some infrastructure changes for the upcoming auto-explain contrib module: * Refactor explain.c slightly to export a convenient-to-use subroutine for printing EXPLAIN results. * Provide hooks for plugins to get control at ExecutorStart and ExecutorEnd as well as ExecutorRun. * Add some minimal support for tracking the total runtime of ExecutorRun. This code won't actually do anything unless a plugin prods it to. * Change the API of the DefineCustomXXXVariable functions to allow nonzero "flags" to be specified for a custom GUC variable. While at it, also make the "bootstrap" default value for custom GUCs be explicitly specified as a parameter to these functions. This is to eliminate confusion over where the default comes from, as has been expressed in the past by some users of the custom-variable facility. * Refactor GUC code a bit to ensure that a custom variable gets initialized to something valid (like its default value) even if the placeholder value was invalid.
2008-11-19 02:10:24 +01:00
init_custom_variable(name, short_desc, long_desc, context, flags,
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
PGC_BOOL, sizeof(struct config_bool));
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
var->variable = valueAddr;
Some infrastructure changes for the upcoming auto-explain contrib module: * Refactor explain.c slightly to export a convenient-to-use subroutine for printing EXPLAIN results. * Provide hooks for plugins to get control at ExecutorStart and ExecutorEnd as well as ExecutorRun. * Add some minimal support for tracking the total runtime of ExecutorRun. This code won't actually do anything unless a plugin prods it to. * Change the API of the DefineCustomXXXVariable functions to allow nonzero "flags" to be specified for a custom GUC variable. While at it, also make the "bootstrap" default value for custom GUCs be explicitly specified as a parameter to these functions. This is to eliminate confusion over where the default comes from, as has been expressed in the past by some users of the custom-variable facility. * Refactor GUC code a bit to ensure that a custom variable gets initialized to something valid (like its default value) even if the placeholder value was invalid.
2008-11-19 02:10:24 +01:00
var->boot_val = bootValue;
var->reset_val = bootValue;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
var->check_hook = check_hook;
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
var->assign_hook = assign_hook;
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
var->show_hook = show_hook;
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
define_custom_variable(&var->gen);
}
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
void
Add SET ROLE. This is a partial commit of Stephen Frost's recent patch; I'm still working on the has_role function and information_schema changes.
2005-07-26 00:12:34 +02:00
DefineCustomIntVariable(const char *name,
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
const char *short_desc,
const char *long_desc,
int *valueAddr,
Some infrastructure changes for the upcoming auto-explain contrib module: * Refactor explain.c slightly to export a convenient-to-use subroutine for printing EXPLAIN results. * Provide hooks for plugins to get control at ExecutorStart and ExecutorEnd as well as ExecutorRun. * Add some minimal support for tracking the total runtime of ExecutorRun. This code won't actually do anything unless a plugin prods it to. * Change the API of the DefineCustomXXXVariable functions to allow nonzero "flags" to be specified for a custom GUC variable. While at it, also make the "bootstrap" default value for custom GUCs be explicitly specified as a parameter to these functions. This is to eliminate confusion over where the default comes from, as has been expressed in the past by some users of the custom-variable facility. * Refactor GUC code a bit to ensure that a custom variable gets initialized to something valid (like its default value) even if the placeholder value was invalid.
2008-11-19 02:10:24 +01:00
int bootValue,
Add missing min/max parameters to DefineCustomIntVariable() and DefineCustomRealVariable(). Thomas Hallgren
2005-03-25 17:17:28 +01:00
int minValue,
int maxValue,
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
GucContext context,
Some infrastructure changes for the upcoming auto-explain contrib module: * Refactor explain.c slightly to export a convenient-to-use subroutine for printing EXPLAIN results. * Provide hooks for plugins to get control at ExecutorStart and ExecutorEnd as well as ExecutorRun. * Add some minimal support for tracking the total runtime of ExecutorRun. This code won't actually do anything unless a plugin prods it to. * Change the API of the DefineCustomXXXVariable functions to allow nonzero "flags" to be specified for a custom GUC variable. While at it, also make the "bootstrap" default value for custom GUCs be explicitly specified as a parameter to these functions. This is to eliminate confusion over where the default comes from, as has been expressed in the past by some users of the custom-variable facility. * Refactor GUC code a bit to ensure that a custom variable gets initialized to something valid (like its default value) even if the placeholder value was invalid.
2008-11-19 02:10:24 +01:00
int flags,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
GucIntCheckHook check_hook,
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
GucIntAssignHook assign_hook,
GucShowHook show_hook)
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
{
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
struct config_int *var;
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
var = (struct config_int *)
Some infrastructure changes for the upcoming auto-explain contrib module: * Refactor explain.c slightly to export a convenient-to-use subroutine for printing EXPLAIN results. * Provide hooks for plugins to get control at ExecutorStart and ExecutorEnd as well as ExecutorRun. * Add some minimal support for tracking the total runtime of ExecutorRun. This code won't actually do anything unless a plugin prods it to. * Change the API of the DefineCustomXXXVariable functions to allow nonzero "flags" to be specified for a custom GUC variable. While at it, also make the "bootstrap" default value for custom GUCs be explicitly specified as a parameter to these functions. This is to eliminate confusion over where the default comes from, as has been expressed in the past by some users of the custom-variable facility. * Refactor GUC code a bit to ensure that a custom variable gets initialized to something valid (like its default value) even if the placeholder value was invalid.
2008-11-19 02:10:24 +01:00
init_custom_variable(name, short_desc, long_desc, context, flags,
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
PGC_INT, sizeof(struct config_int));
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
var->variable = valueAddr;
Some infrastructure changes for the upcoming auto-explain contrib module: * Refactor explain.c slightly to export a convenient-to-use subroutine for printing EXPLAIN results. * Provide hooks for plugins to get control at ExecutorStart and ExecutorEnd as well as ExecutorRun. * Add some minimal support for tracking the total runtime of ExecutorRun. This code won't actually do anything unless a plugin prods it to. * Change the API of the DefineCustomXXXVariable functions to allow nonzero "flags" to be specified for a custom GUC variable. While at it, also make the "bootstrap" default value for custom GUCs be explicitly specified as a parameter to these functions. This is to eliminate confusion over where the default comes from, as has been expressed in the past by some users of the custom-variable facility. * Refactor GUC code a bit to ensure that a custom variable gets initialized to something valid (like its default value) even if the placeholder value was invalid.
2008-11-19 02:10:24 +01:00
var->boot_val = bootValue;
var->reset_val = bootValue;
Add missing min/max parameters to DefineCustomIntVariable() and DefineCustomRealVariable(). Thomas Hallgren
2005-03-25 17:17:28 +01:00
var->min = minValue;
var->max = maxValue;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
var->check_hook = check_hook;
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
var->assign_hook = assign_hook;
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
var->show_hook = show_hook;
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
define_custom_variable(&var->gen);
}
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
void
Add SET ROLE. This is a partial commit of Stephen Frost's recent patch; I'm still working on the has_role function and information_schema changes.
2005-07-26 00:12:34 +02:00
DefineCustomRealVariable(const char *name,
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
const char *short_desc,
const char *long_desc,
double *valueAddr,
Some infrastructure changes for the upcoming auto-explain contrib module: * Refactor explain.c slightly to export a convenient-to-use subroutine for printing EXPLAIN results. * Provide hooks for plugins to get control at ExecutorStart and ExecutorEnd as well as ExecutorRun. * Add some minimal support for tracking the total runtime of ExecutorRun. This code won't actually do anything unless a plugin prods it to. * Change the API of the DefineCustomXXXVariable functions to allow nonzero "flags" to be specified for a custom GUC variable. While at it, also make the "bootstrap" default value for custom GUCs be explicitly specified as a parameter to these functions. This is to eliminate confusion over where the default comes from, as has been expressed in the past by some users of the custom-variable facility. * Refactor GUC code a bit to ensure that a custom variable gets initialized to something valid (like its default value) even if the placeholder value was invalid.
2008-11-19 02:10:24 +01:00
double bootValue,
Add missing min/max parameters to DefineCustomIntVariable() and DefineCustomRealVariable(). Thomas Hallgren
2005-03-25 17:17:28 +01:00
double minValue,
double maxValue,
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
GucContext context,
Some infrastructure changes for the upcoming auto-explain contrib module: * Refactor explain.c slightly to export a convenient-to-use subroutine for printing EXPLAIN results. * Provide hooks for plugins to get control at ExecutorStart and ExecutorEnd as well as ExecutorRun. * Add some minimal support for tracking the total runtime of ExecutorRun. This code won't actually do anything unless a plugin prods it to. * Change the API of the DefineCustomXXXVariable functions to allow nonzero "flags" to be specified for a custom GUC variable. While at it, also make the "bootstrap" default value for custom GUCs be explicitly specified as a parameter to these functions. This is to eliminate confusion over where the default comes from, as has been expressed in the past by some users of the custom-variable facility. * Refactor GUC code a bit to ensure that a custom variable gets initialized to something valid (like its default value) even if the placeholder value was invalid.
2008-11-19 02:10:24 +01:00
int flags,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
GucRealCheckHook check_hook,
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
GucRealAssignHook assign_hook,
GucShowHook show_hook)
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
{
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
struct config_real *var;
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
var = (struct config_real *)
Some infrastructure changes for the upcoming auto-explain contrib module: * Refactor explain.c slightly to export a convenient-to-use subroutine for printing EXPLAIN results. * Provide hooks for plugins to get control at ExecutorStart and ExecutorEnd as well as ExecutorRun. * Add some minimal support for tracking the total runtime of ExecutorRun. This code won't actually do anything unless a plugin prods it to. * Change the API of the DefineCustomXXXVariable functions to allow nonzero "flags" to be specified for a custom GUC variable. While at it, also make the "bootstrap" default value for custom GUCs be explicitly specified as a parameter to these functions. This is to eliminate confusion over where the default comes from, as has been expressed in the past by some users of the custom-variable facility. * Refactor GUC code a bit to ensure that a custom variable gets initialized to something valid (like its default value) even if the placeholder value was invalid.
2008-11-19 02:10:24 +01:00
init_custom_variable(name, short_desc, long_desc, context, flags,
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
PGC_REAL, sizeof(struct config_real));
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
var->variable = valueAddr;
Some infrastructure changes for the upcoming auto-explain contrib module: * Refactor explain.c slightly to export a convenient-to-use subroutine for printing EXPLAIN results. * Provide hooks for plugins to get control at ExecutorStart and ExecutorEnd as well as ExecutorRun. * Add some minimal support for tracking the total runtime of ExecutorRun. This code won't actually do anything unless a plugin prods it to. * Change the API of the DefineCustomXXXVariable functions to allow nonzero "flags" to be specified for a custom GUC variable. While at it, also make the "bootstrap" default value for custom GUCs be explicitly specified as a parameter to these functions. This is to eliminate confusion over where the default comes from, as has been expressed in the past by some users of the custom-variable facility. * Refactor GUC code a bit to ensure that a custom variable gets initialized to something valid (like its default value) even if the placeholder value was invalid.
2008-11-19 02:10:24 +01:00
var->boot_val = bootValue;
var->reset_val = bootValue;
Add missing min/max parameters to DefineCustomIntVariable() and DefineCustomRealVariable(). Thomas Hallgren
2005-03-25 17:17:28 +01:00
var->min = minValue;
var->max = maxValue;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
var->check_hook = check_hook;
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
var->assign_hook = assign_hook;
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
var->show_hook = show_hook;
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
define_custom_variable(&var->gen);
}
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
void
Add SET ROLE. This is a partial commit of Stephen Frost's recent patch; I'm still working on the has_role function and information_schema changes.
2005-07-26 00:12:34 +02:00
DefineCustomStringVariable(const char *name,
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
const char *short_desc,
const char *long_desc,
char **valueAddr,
Some infrastructure changes for the upcoming auto-explain contrib module: * Refactor explain.c slightly to export a convenient-to-use subroutine for printing EXPLAIN results. * Provide hooks for plugins to get control at ExecutorStart and ExecutorEnd as well as ExecutorRun. * Add some minimal support for tracking the total runtime of ExecutorRun. This code won't actually do anything unless a plugin prods it to. * Change the API of the DefineCustomXXXVariable functions to allow nonzero "flags" to be specified for a custom GUC variable. While at it, also make the "bootstrap" default value for custom GUCs be explicitly specified as a parameter to these functions. This is to eliminate confusion over where the default comes from, as has been expressed in the past by some users of the custom-variable facility. * Refactor GUC code a bit to ensure that a custom variable gets initialized to something valid (like its default value) even if the placeholder value was invalid.
2008-11-19 02:10:24 +01:00
const char *bootValue,
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
GucContext context,
Some infrastructure changes for the upcoming auto-explain contrib module: * Refactor explain.c slightly to export a convenient-to-use subroutine for printing EXPLAIN results. * Provide hooks for plugins to get control at ExecutorStart and ExecutorEnd as well as ExecutorRun. * Add some minimal support for tracking the total runtime of ExecutorRun. This code won't actually do anything unless a plugin prods it to. * Change the API of the DefineCustomXXXVariable functions to allow nonzero "flags" to be specified for a custom GUC variable. While at it, also make the "bootstrap" default value for custom GUCs be explicitly specified as a parameter to these functions. This is to eliminate confusion over where the default comes from, as has been expressed in the past by some users of the custom-variable facility. * Refactor GUC code a bit to ensure that a custom variable gets initialized to something valid (like its default value) even if the placeholder value was invalid.
2008-11-19 02:10:24 +01:00
int flags,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
GucStringCheckHook check_hook,
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
GucStringAssignHook assign_hook,
GucShowHook show_hook)
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
{
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
struct config_string *var;
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
var = (struct config_string *)
Some infrastructure changes for the upcoming auto-explain contrib module: * Refactor explain.c slightly to export a convenient-to-use subroutine for printing EXPLAIN results. * Provide hooks for plugins to get control at ExecutorStart and ExecutorEnd as well as ExecutorRun. * Add some minimal support for tracking the total runtime of ExecutorRun. This code won't actually do anything unless a plugin prods it to. * Change the API of the DefineCustomXXXVariable functions to allow nonzero "flags" to be specified for a custom GUC variable. While at it, also make the "bootstrap" default value for custom GUCs be explicitly specified as a parameter to these functions. This is to eliminate confusion over where the default comes from, as has been expressed in the past by some users of the custom-variable facility. * Refactor GUC code a bit to ensure that a custom variable gets initialized to something valid (like its default value) even if the placeholder value was invalid.
2008-11-19 02:10:24 +01:00
init_custom_variable(name, short_desc, long_desc, context, flags,
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
PGC_STRING, sizeof(struct config_string));
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
var->variable = valueAddr;
Some infrastructure changes for the upcoming auto-explain contrib module: * Refactor explain.c slightly to export a convenient-to-use subroutine for printing EXPLAIN results. * Provide hooks for plugins to get control at ExecutorStart and ExecutorEnd as well as ExecutorRun. * Add some minimal support for tracking the total runtime of ExecutorRun. This code won't actually do anything unless a plugin prods it to. * Change the API of the DefineCustomXXXVariable functions to allow nonzero "flags" to be specified for a custom GUC variable. While at it, also make the "bootstrap" default value for custom GUCs be explicitly specified as a parameter to these functions. This is to eliminate confusion over where the default comes from, as has been expressed in the past by some users of the custom-variable facility. * Refactor GUC code a bit to ensure that a custom variable gets initialized to something valid (like its default value) even if the placeholder value was invalid.
2008-11-19 02:10:24 +01:00
var->boot_val = bootValue;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
var->check_hook = check_hook;
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
var->assign_hook = assign_hook;
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
var->show_hook = show_hook;
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
define_custom_variable(&var->gen);
}
Some cleanups of enum-guc code, per comments from Tom.
2008-03-16 17:42:44 +01:00
void
DefineCustomEnumVariable(const char *name,
const char *short_desc,
const char *long_desc,
int *valueAddr,
Some infrastructure changes for the upcoming auto-explain contrib module: * Refactor explain.c slightly to export a convenient-to-use subroutine for printing EXPLAIN results. * Provide hooks for plugins to get control at ExecutorStart and ExecutorEnd as well as ExecutorRun. * Add some minimal support for tracking the total runtime of ExecutorRun. This code won't actually do anything unless a plugin prods it to. * Change the API of the DefineCustomXXXVariable functions to allow nonzero "flags" to be specified for a custom GUC variable. While at it, also make the "bootstrap" default value for custom GUCs be explicitly specified as a parameter to these functions. This is to eliminate confusion over where the default comes from, as has been expressed in the past by some users of the custom-variable facility. * Refactor GUC code a bit to ensure that a custom variable gets initialized to something valid (like its default value) even if the placeholder value was invalid.
2008-11-19 02:10:24 +01:00
int bootValue,
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
const struct config_enum_entry * options,
Some cleanups of enum-guc code, per comments from Tom.
2008-03-16 17:42:44 +01:00
GucContext context,
Some infrastructure changes for the upcoming auto-explain contrib module: * Refactor explain.c slightly to export a convenient-to-use subroutine for printing EXPLAIN results. * Provide hooks for plugins to get control at ExecutorStart and ExecutorEnd as well as ExecutorRun. * Add some minimal support for tracking the total runtime of ExecutorRun. This code won't actually do anything unless a plugin prods it to. * Change the API of the DefineCustomXXXVariable functions to allow nonzero "flags" to be specified for a custom GUC variable. While at it, also make the "bootstrap" default value for custom GUCs be explicitly specified as a parameter to these functions. This is to eliminate confusion over where the default comes from, as has been expressed in the past by some users of the custom-variable facility. * Refactor GUC code a bit to ensure that a custom variable gets initialized to something valid (like its default value) even if the placeholder value was invalid.
2008-11-19 02:10:24 +01:00
int flags,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
GucEnumCheckHook check_hook,
Some cleanups of enum-guc code, per comments from Tom.
2008-03-16 17:42:44 +01:00
GucEnumAssignHook assign_hook,
GucShowHook show_hook)
{
struct config_enum *var;
var = (struct config_enum *)
Some infrastructure changes for the upcoming auto-explain contrib module: * Refactor explain.c slightly to export a convenient-to-use subroutine for printing EXPLAIN results. * Provide hooks for plugins to get control at ExecutorStart and ExecutorEnd as well as ExecutorRun. * Add some minimal support for tracking the total runtime of ExecutorRun. This code won't actually do anything unless a plugin prods it to. * Change the API of the DefineCustomXXXVariable functions to allow nonzero "flags" to be specified for a custom GUC variable. While at it, also make the "bootstrap" default value for custom GUCs be explicitly specified as a parameter to these functions. This is to eliminate confusion over where the default comes from, as has been expressed in the past by some users of the custom-variable facility. * Refactor GUC code a bit to ensure that a custom variable gets initialized to something valid (like its default value) even if the placeholder value was invalid.
2008-11-19 02:10:24 +01:00
init_custom_variable(name, short_desc, long_desc, context, flags,
Some cleanups of enum-guc code, per comments from Tom.
2008-03-16 17:42:44 +01:00
PGC_ENUM, sizeof(struct config_enum));
var->variable = valueAddr;
Some infrastructure changes for the upcoming auto-explain contrib module: * Refactor explain.c slightly to export a convenient-to-use subroutine for printing EXPLAIN results. * Provide hooks for plugins to get control at ExecutorStart and ExecutorEnd as well as ExecutorRun. * Add some minimal support for tracking the total runtime of ExecutorRun. This code won't actually do anything unless a plugin prods it to. * Change the API of the DefineCustomXXXVariable functions to allow nonzero "flags" to be specified for a custom GUC variable. While at it, also make the "bootstrap" default value for custom GUCs be explicitly specified as a parameter to these functions. This is to eliminate confusion over where the default comes from, as has been expressed in the past by some users of the custom-variable facility. * Refactor GUC code a bit to ensure that a custom variable gets initialized to something valid (like its default value) even if the placeholder value was invalid.
2008-11-19 02:10:24 +01:00
var->boot_val = bootValue;
var->reset_val = bootValue;
Some cleanups of enum-guc code, per comments from Tom.
2008-03-16 17:42:44 +01:00
var->options = options;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
var->check_hook = check_hook;
Some cleanups of enum-guc code, per comments from Tom.
2008-03-16 17:42:44 +01:00
var->assign_hook = assign_hook;
var->show_hook = show_hook;
define_custom_variable(&var->gen);
}
Cosmetic improvements/code cleanup: - replace some function signatures of the form "some_type foo()" with "some_type foo(void)" - replace a few instances of a literal 0 being used as a NULL pointer; there are more instances of this in the code, but I just fixed a few - in src/backend/utils/mb/wstrncmp.c, replace K&R style function declarations with ANSI style, remove use of 'register' keyword - remove an "extern" modifier that was applied to a function definition (rather than a declaration)
2004-10-11 01:37:45 +02:00
void
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
EmitWarningsOnPlaceholders(const char *className)
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
{
Change a couple of ill-advised uses of INFO elog level to WARNINGs; in particular this allows EmitWarningsOnPlaceholders messages to show up in the postmaster log by default. Update elog.h comment to make it clearer what INFO is for, and fix one example in the SGML docs that was misusing it. Per my gripe of yesterday.
2009-01-06 17:39:52 +01:00
int classLen = strlen(className);
int i;
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
Change a couple of ill-advised uses of INFO elog level to WARNINGs; in particular this allows EmitWarningsOnPlaceholders messages to show up in the postmaster log by default. Update elog.h comment to make it clearer what INFO is for, and fix one example in the SGML docs that was misusing it. Per my gripe of yesterday.
2009-01-06 17:39:52 +01:00
for (i = 0; i < num_guc_variables; i++)
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
{
Change a couple of ill-advised uses of INFO elog level to WARNINGs; in particular this allows EmitWarningsOnPlaceholders messages to show up in the postmaster log by default. Update elog.h comment to make it clearer what INFO is for, and fix one example in the SGML docs that was misusing it. Per my gripe of yesterday.
2009-01-06 17:39:52 +01:00
struct config_generic *var = guc_variables[i];
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
if ((var->flags & GUC_CUSTOM_PLACEHOLDER) != 0 &&
Change a couple of ill-advised uses of INFO elog level to WARNINGs; in particular this allows EmitWarningsOnPlaceholders messages to show up in the postmaster log by default. Update elog.h comment to make it clearer what INFO is for, and fix one example in the SGML docs that was misusing it. Per my gripe of yesterday.
2009-01-06 17:39:52 +01:00
strncmp(className, var->name, classLen) == 0 &&
var->name[classLen] == GUC_QUALIFIER_SEPARATOR)
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
{
Change a couple of ill-advised uses of INFO elog level to WARNINGs; in particular this allows EmitWarningsOnPlaceholders messages to show up in the postmaster log by default. Update elog.h comment to make it clearer what INFO is for, and fix one example in the SGML docs that was misusing it. Per my gripe of yesterday.
2009-01-06 17:39:52 +01:00
ereport(WARNING,
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
(errcode(ERRCODE_UNDEFINED_OBJECT),
Change a couple of ill-advised uses of INFO elog level to WARNINGs; in particular this allows EmitWarningsOnPlaceholders messages to show up in the postmaster log by default. Update elog.h comment to make it clearer what INFO is for, and fix one example in the SGML docs that was misusing it. Per my gripe of yesterday.
2009-01-06 17:39:52 +01:00
errmsg("unrecognized configuration parameter \"%s\"",
var->name)));
The patch adresses the TODO list item "Allow external interfaces to extend the GUC variable set". Plugin modules like the pl<lang> modules needs a way to declare configuration parameters. The postmaster has no knowledge of such modules when it reads the postgresql.conf file. Rather than allowing totally unknown configuration parameters, the concept of a variable "class" is introduced. Variables that belongs to a declared classes will create a placeholder value of string type and will not generate an error. When a module is loaded, it will declare variables for such a class and make those variables "consume" any placeholders that has been defined. Finally, the module will generate warnings for unrecognized placeholders defined for its class. More detail: The design is outlined after the suggestions made by Tom Lane and Joe Conway in this thread: http://archives.postgresql.org/pgsql-hackers/2004-02/msg00229.php A new string variable 'custom_variable_classes' is introduced. This variable is a comma separated string of identifiers. Each identifier denots a 'class' that will allow its members to be added without error. This variable must be defined in postmaster.conf. The lexer (guc_file.l) is changed so that it can accept a qualified name in the form <ID>.<ID> as the name of a variable. I also changed so that the 'custom_variable_classes', if found, is added first of all variables in order to remove the order of declaration issue. The guc_variables table is made more dynamic. It is originally created with 20% slack and can grow dynamically. A capacity is introduced to avoid resizing every time a new variable is added. guc_variables and num_guc_variables becomes static (hidden). The GucInfoMain now uses the new function get_guc_variables() and GetNumConfigOptions instead or using the guc_variables directly. The find_option() function, when passed a missing name, will check if the name is qualified. If the name is qualified and if the qualifier denotes a class included in the 'custom_variable_classes', a placeholder variable will be created. Such a placeholder will not participate in a list operation but will otherwise function as a normal string variable. Define<type>GucVariable() functions will be added, one for each variable type. They are inteded to be used by add-on modules like the pl<lang> mappings. Example: extern void DefineCustomBoolVariable( const char* name, const char* short_desc, const char* long_desc, bool* valueAddr, GucContext context, GucBoolAssignHook assign_hook, GucShowHook show_hook); (I created typedefs for the assign-hook and show-hook functions). A call to these functions will define a new GUC-variable. If a placeholder exists it will be replaced but it's value will be used in place of the default value. The valueAddr is assumed ot point at a default value when the define function is called. The only constraint that is imposed on a Custom variable is that its name is qualified. Finally, a function: void EmittWarningsOnPlacholders(const char* className) was added. This function should be called when a module has completed its variable definitions. At that time, no placeholders should remain for the class that the module uses. If they do, elog(INFO, ...) messages will be issued to inform the user that unrecognized variables are present. Thomas Hallgren
2004-05-26 17:07:41 +02:00
}
}
}
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
/*
* SHOW command
*/
void
Restructure command destination handling so that we pass around DestReceiver pointers instead of just CommandDest values. The DestReceiver is made at the point where the destination is selected, rather than deep inside the executor. This cleans up the original kluge implementation of tstoreReceiver.c, and makes it easy to support retrieving results from utility statements inside portals. Thus, you can now do fun things like Bind and Execute a FETCH or EXPLAIN command, and it'll all work as expected (e.g., you can Describe the portal, or use Execute's count parameter to suspend the output partway through). Implementation involves stuffing the utility command's output into a Tuplestore, which would be kind of annoying for huge output sets, but should be quite acceptable for typical uses of utility commands.
2003-05-06 22:26:28 +02:00
GetPGVariable(const char *name, DestReceiver *dest)
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
{
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
if (guc_name_compare(name, "all") == 0)
Restructure command destination handling so that we pass around DestReceiver pointers instead of just CommandDest values. The DestReceiver is made at the point where the destination is selected, rather than deep inside the executor. This cleans up the original kluge implementation of tstoreReceiver.c, and makes it easy to support retrieving results from utility statements inside portals. Thus, you can now do fun things like Bind and Execute a FETCH or EXPLAIN command, and it'll all work as expected (e.g., you can Describe the portal, or use Execute's count parameter to suspend the output partway through). Implementation involves stuffing the utility command's output into a Tuplestore, which would be kind of annoying for huge output sets, but should be quite acceptable for typical uses of utility commands.
2003-05-06 22:26:28 +02:00
ShowAllGUCConfig(dest);
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
else
Restructure command destination handling so that we pass around DestReceiver pointers instead of just CommandDest values. The DestReceiver is made at the point where the destination is selected, rather than deep inside the executor. This cleans up the original kluge implementation of tstoreReceiver.c, and makes it easy to support retrieving results from utility statements inside portals. Thus, you can now do fun things like Bind and Execute a FETCH or EXPLAIN command, and it'll all work as expected (e.g., you can Describe the portal, or use Execute's count parameter to suspend the output partway through). Implementation involves stuffing the utility command's output into a Tuplestore, which would be kind of annoying for huge output sets, but should be quite acceptable for typical uses of utility commands.
2003-05-06 22:26:28 +02:00
ShowGUCConfigOption(name, dest);
}
TupleDesc
GetPGVariableResultDesc(const char *name)
{
TupleDesc tupdesc;
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
if (guc_name_compare(name, "all") == 0)
Restructure command destination handling so that we pass around DestReceiver pointers instead of just CommandDest values. The DestReceiver is made at the point where the destination is selected, rather than deep inside the executor. This cleans up the original kluge implementation of tstoreReceiver.c, and makes it easy to support retrieving results from utility statements inside portals. Thus, you can now do fun things like Bind and Execute a FETCH or EXPLAIN command, and it'll all work as expected (e.g., you can Describe the portal, or use Execute's count parameter to suspend the output partway through). Implementation involves stuffing the utility command's output into a Tuplestore, which would be kind of annoying for huge output sets, but should be quite acceptable for typical uses of utility commands.
2003-05-06 22:26:28 +02:00
{
Have SHOW ALL include variable descriptions. Matthias Schmidt
2005-06-14 22:42:53 +02:00
/* need a tuple descriptor representing three TEXT columns */
tupdesc = CreateTemplateTupleDesc(3, false);
Restructure command destination handling so that we pass around DestReceiver pointers instead of just CommandDest values. The DestReceiver is made at the point where the destination is selected, rather than deep inside the executor. This cleans up the original kluge implementation of tstoreReceiver.c, and makes it easy to support retrieving results from utility statements inside portals. Thus, you can now do fun things like Bind and Execute a FETCH or EXPLAIN command, and it'll all work as expected (e.g., you can Describe the portal, or use Execute's count parameter to suspend the output partway through). Implementation involves stuffing the utility command's output into a Tuplestore, which would be kind of annoying for huge output sets, but should be quite acceptable for typical uses of utility commands.
2003-05-06 22:26:28 +02:00
TupleDescInitEntry(tupdesc, (AttrNumber) 1, "name",
Replace TupleTableSlot convention for whole-row variables and function results with tuples as ordinary varlena Datums. This commit does not in itself do much for us, except eliminate the horrid memory leak associated with evaluation of whole-row variables. However, it lays the groundwork for allowing composite types as table columns, and perhaps some other useful features as well. Per my proposal of a few days ago.
2004-04-01 23:28:47 +02:00
TEXTOID, -1, 0);
Restructure command destination handling so that we pass around DestReceiver pointers instead of just CommandDest values. The DestReceiver is made at the point where the destination is selected, rather than deep inside the executor. This cleans up the original kluge implementation of tstoreReceiver.c, and makes it easy to support retrieving results from utility statements inside portals. Thus, you can now do fun things like Bind and Execute a FETCH or EXPLAIN command, and it'll all work as expected (e.g., you can Describe the portal, or use Execute's count parameter to suspend the output partway through). Implementation involves stuffing the utility command's output into a Tuplestore, which would be kind of annoying for huge output sets, but should be quite acceptable for typical uses of utility commands.
2003-05-06 22:26:28 +02:00
TupleDescInitEntry(tupdesc, (AttrNumber) 2, "setting",
Replace TupleTableSlot convention for whole-row variables and function results with tuples as ordinary varlena Datums. This commit does not in itself do much for us, except eliminate the horrid memory leak associated with evaluation of whole-row variables. However, it lays the groundwork for allowing composite types as table columns, and perhaps some other useful features as well. Per my proposal of a few days ago.
2004-04-01 23:28:47 +02:00
TEXTOID, -1, 0);
Have SHOW ALL include variable descriptions. Matthias Schmidt
2005-06-14 22:42:53 +02:00
TupleDescInitEntry(tupdesc, (AttrNumber) 3, "description",
TEXTOID, -1, 0);
Restructure command destination handling so that we pass around DestReceiver pointers instead of just CommandDest values. The DestReceiver is made at the point where the destination is selected, rather than deep inside the executor. This cleans up the original kluge implementation of tstoreReceiver.c, and makes it easy to support retrieving results from utility statements inside portals. Thus, you can now do fun things like Bind and Execute a FETCH or EXPLAIN command, and it'll all work as expected (e.g., you can Describe the portal, or use Execute's count parameter to suspend the output partway through). Implementation involves stuffing the utility command's output into a Tuplestore, which would be kind of annoying for huge output sets, but should be quite acceptable for typical uses of utility commands.
2003-05-06 22:26:28 +02:00
}
else
{
const char *varname;
/* Get the canonical spelling of name */
(void) GetConfigOptionByName(name, &varname);
/* need a tuple descriptor representing a single TEXT column */
tupdesc = CreateTemplateTupleDesc(1, false);
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
TupleDescInitEntry(tupdesc, (AttrNumber) 1, varname,
TEXTOID, -1, 0);
Restructure command destination handling so that we pass around DestReceiver pointers instead of just CommandDest values. The DestReceiver is made at the point where the destination is selected, rather than deep inside the executor. This cleans up the original kluge implementation of tstoreReceiver.c, and makes it easy to support retrieving results from utility statements inside portals. Thus, you can now do fun things like Bind and Execute a FETCH or EXPLAIN command, and it'll all work as expected (e.g., you can Describe the portal, or use Execute's count parameter to suspend the output partway through). Implementation involves stuffing the utility command's output into a Tuplestore, which would be kind of annoying for huge output sets, but should be quite acceptable for typical uses of utility commands.
2003-05-06 22:26:28 +02:00
}
return tupdesc;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
}
/*
* SHOW command
*/
Add a GUC_SUPERUSER_ONLY flag to mark GUC variables that should not be examinable by non-superusers, and use it to protect the recently-added GUC variables for data directory and config files. For now I have only flagged those variables that could be used to deduce something about the server's filesystem layout, but possibly we should also mark vars related to logging settings and other admin-only information?
2004-10-22 21:48:19 +02:00
static void
Restructure command destination handling so that we pass around DestReceiver pointers instead of just CommandDest values. The DestReceiver is made at the point where the destination is selected, rather than deep inside the executor. This cleans up the original kluge implementation of tstoreReceiver.c, and makes it easy to support retrieving results from utility statements inside portals. Thus, you can now do fun things like Bind and Execute a FETCH or EXPLAIN command, and it'll all work as expected (e.g., you can Describe the portal, or use Execute's count parameter to suspend the output partway through). Implementation involves stuffing the utility command's output into a Tuplestore, which would be kind of annoying for huge output sets, but should be quite acceptable for typical uses of utility commands.
2003-05-06 22:26:28 +02:00
ShowGUCConfigOption(const char *name, DestReceiver *dest)
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
{
> 2. This patch includes the same Table Function API fixes that I > submitted on July 9: > > http://archives.postgresql.org/pgsql-patches/2002-07/msg00056.php > > Please disregard that one *if* this one is applied. If this one is > rejected please go ahead with the July 9th patch. The July 9th Table Function API patch mentioned above is now in CVS, so here is an updated version of the guc patch which should apply cleanly against CVS tip. Joe Conway
2002-07-20 07:49:28 +02:00
TupOutputState *tstate;
pgindent run.
2002-09-04 22:31:48 +02:00
TupleDesc tupdesc;
const char *varname;
char *value;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
Code review for SHOW output changes; fix horology expected files for new SHOW output format.
2002-07-20 17:12:56 +02:00
/* Get the value and canonical spelling of name */
value = GetConfigOptionByName(name, &varname);
> 2. This patch includes the same Table Function API fixes that I > submitted on July 9: > > http://archives.postgresql.org/pgsql-patches/2002-07/msg00056.php > > Please disregard that one *if* this one is applied. If this one is > rejected please go ahead with the July 9th patch. The July 9th Table Function API patch mentioned above is now in CVS, so here is an updated version of the guc patch which should apply cleanly against CVS tip. Joe Conway
2002-07-20 07:49:28 +02:00
/* need a tuple descriptor representing a single TEXT column */
Code review for HeapTupleHeader changes. Add version number to page headers (overlaying low byte of page size) and add HEAP_HASOID bit to t_infomask, per earlier discussion. Simplify scheme for overlaying fields in tuple header (no need for cmax to live in more than one place). Don't try to clear infomask status bits in tqual.c --- not safe to do it there. Don't try to force output table of a SELECT INTO to have OIDs, either. Get rid of unnecessarily complex three-state scheme for TupleDesc.tdhasoids, which has already caused one recent failure. Improve documentation.
2002-09-02 03:05:06 +02:00
tupdesc = CreateTemplateTupleDesc(1, false);
Rewriter and planner should use only resno, not resname, to identify target columns in INSERT and UPDATE targetlists. Don't rely on resname to be accurate in ruleutils, either. This fixes bug reported by Donald Fraser, in which renaming a column referenced in a rule did not work very well.
2003-08-12 01:04:50 +02:00
TupleDescInitEntry(tupdesc, (AttrNumber) 1, varname,
Replace TupleTableSlot convention for whole-row variables and function results with tuples as ordinary varlena Datums. This commit does not in itself do much for us, except eliminate the horrid memory leak associated with evaluation of whole-row variables. However, it lays the groundwork for allowing composite types as table columns, and perhaps some other useful features as well. Per my proposal of a few days ago.
2004-04-01 23:28:47 +02:00
TEXTOID, -1, 0);
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
> 2. This patch includes the same Table Function API fixes that I > submitted on July 9: > > http://archives.postgresql.org/pgsql-patches/2002-07/msg00056.php > > Please disregard that one *if* this one is applied. If this one is > rejected please go ahead with the July 9th patch. The July 9th Table Function API patch mentioned above is now in CVS, so here is an updated version of the guc patch which should apply cleanly against CVS tip. Joe Conway
2002-07-20 07:49:28 +02:00
/* prepare for projection of tuples */
tstate = begin_tup_output_tupdesc(dest, tupdesc);
/* Send it */
Code review for standalone composite types, query-specified composite types, SRFs. Not happy with memory management yet, but I'll commit these other changes.
2002-08-29 02:17:06 +02:00
do_text_output_oneline(tstate, value);
> 2. This patch includes the same Table Function API fixes that I > submitted on July 9: > > http://archives.postgresql.org/pgsql-patches/2002-07/msg00056.php > > Please disregard that one *if* this one is applied. If this one is > rejected please go ahead with the July 9th patch. The July 9th Table Function API patch mentioned above is now in CVS, so here is an updated version of the guc patch which should apply cleanly against CVS tip. Joe Conway
2002-07-20 07:49:28 +02:00
end_tup_output(tstate);
reset all: command line and .conf options change defaults on RESET ALL those are restored. show all: GUC + non-GUC. SHOW ALL, RESET ALL Marko Kreen
2001-06-07 06:50:57 +02:00
}
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
/*
* SHOW ALL command
*/
Add a GUC_SUPERUSER_ONLY flag to mark GUC variables that should not be examinable by non-superusers, and use it to protect the recently-added GUC variables for data directory and config files. For now I have only flagged those variables that could be used to deduce something about the server's filesystem layout, but possibly we should also mark vars related to logging settings and other admin-only information?
2004-10-22 21:48:19 +02:00
static void
Restructure command destination handling so that we pass around DestReceiver pointers instead of just CommandDest values. The DestReceiver is made at the point where the destination is selected, rather than deep inside the executor. This cleans up the original kluge implementation of tstoreReceiver.c, and makes it easy to support retrieving results from utility statements inside portals. Thus, you can now do fun things like Bind and Execute a FETCH or EXPLAIN command, and it'll all work as expected (e.g., you can Describe the portal, or use Execute's count parameter to suspend the output partway through). Implementation involves stuffing the utility command's output into a Tuplestore, which would be kind of annoying for huge output sets, but should be quite acceptable for typical uses of utility commands.
2003-05-06 22:26:28 +02:00
ShowAllGUCConfig(DestReceiver *dest)
reset all: command line and .conf options change defaults on RESET ALL those are restored. show all: GUC + non-GUC. SHOW ALL, RESET ALL Marko Kreen
2001-06-07 06:50:57 +02:00
{
Add a GUC_SUPERUSER_ONLY flag to mark GUC variables that should not be examinable by non-superusers, and use it to protect the recently-added GUC variables for data directory and config files. For now I have only flagged those variables that could be used to deduce something about the server's filesystem layout, but possibly we should also mark vars related to logging settings and other admin-only information?
2004-10-22 21:48:19 +02:00
bool am_superuser = superuser();
reset all: command line and .conf options change defaults on RESET ALL those are restored. show all: GUC + non-GUC. SHOW ALL, RESET ALL Marko Kreen
2001-06-07 06:50:57 +02:00
int i;
> 2. This patch includes the same Table Function API fixes that I > submitted on July 9: > > http://archives.postgresql.org/pgsql-patches/2002-07/msg00056.php > > Please disregard that one *if* this one is applied. If this one is > rejected please go ahead with the July 9th patch. The July 9th Table Function API patch mentioned above is now in CVS, so here is an updated version of the guc patch which should apply cleanly against CVS tip. Joe Conway
2002-07-20 07:49:28 +02:00
TupOutputState *tstate;
pgindent run.
2002-09-04 22:31:48 +02:00
TupleDesc tupdesc;
pgindent run for 9.0
2010-02-26 03:01:40 +01:00
Datum values[3];
bool isnull[3] = {false, false, false};
> 2. This patch includes the same Table Function API fixes that I > submitted on July 9: > > http://archives.postgresql.org/pgsql-patches/2002-07/msg00056.php > > Please disregard that one *if* this one is applied. If this one is > rejected please go ahead with the July 9th patch. The July 9th Table Function API patch mentioned above is now in CVS, so here is an updated version of the guc patch which should apply cleanly against CVS tip. Joe Conway
2002-07-20 07:49:28 +02:00
Have SHOW ALL include variable descriptions. Matthias Schmidt
2005-06-14 22:42:53 +02:00
/* need a tuple descriptor representing three TEXT columns */
tupdesc = CreateTemplateTupleDesc(3, false);
> 2. This patch includes the same Table Function API fixes that I > submitted on July 9: > > http://archives.postgresql.org/pgsql-patches/2002-07/msg00056.php > > Please disregard that one *if* this one is applied. If this one is > rejected please go ahead with the July 9th patch. The July 9th Table Function API patch mentioned above is now in CVS, so here is an updated version of the guc patch which should apply cleanly against CVS tip. Joe Conway
2002-07-20 07:49:28 +02:00
TupleDescInitEntry(tupdesc, (AttrNumber) 1, "name",
Replace TupleTableSlot convention for whole-row variables and function results with tuples as ordinary varlena Datums. This commit does not in itself do much for us, except eliminate the horrid memory leak associated with evaluation of whole-row variables. However, it lays the groundwork for allowing composite types as table columns, and perhaps some other useful features as well. Per my proposal of a few days ago.
2004-04-01 23:28:47 +02:00
TEXTOID, -1, 0);
> 2. This patch includes the same Table Function API fixes that I > submitted on July 9: > > http://archives.postgresql.org/pgsql-patches/2002-07/msg00056.php > > Please disregard that one *if* this one is applied. If this one is > rejected please go ahead with the July 9th patch. The July 9th Table Function API patch mentioned above is now in CVS, so here is an updated version of the guc patch which should apply cleanly against CVS tip. Joe Conway
2002-07-20 07:49:28 +02:00
TupleDescInitEntry(tupdesc, (AttrNumber) 2, "setting",
Replace TupleTableSlot convention for whole-row variables and function results with tuples as ordinary varlena Datums. This commit does not in itself do much for us, except eliminate the horrid memory leak associated with evaluation of whole-row variables. However, it lays the groundwork for allowing composite types as table columns, and perhaps some other useful features as well. Per my proposal of a few days ago.
2004-04-01 23:28:47 +02:00
TEXTOID, -1, 0);
Have SHOW ALL include variable descriptions. Matthias Schmidt
2005-06-14 22:42:53 +02:00
TupleDescInitEntry(tupdesc, (AttrNumber) 3, "description",
TEXTOID, -1, 0);
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
> 2. This patch includes the same Table Function API fixes that I > submitted on July 9: > > http://archives.postgresql.org/pgsql-patches/2002-07/msg00056.php > > Please disregard that one *if* this one is applied. If this one is > rejected please go ahead with the July 9th patch. The July 9th Table Function API patch mentioned above is now in CVS, so here is an updated version of the guc patch which should apply cleanly against CVS tip. Joe Conway
2002-07-20 07:49:28 +02:00
/* prepare for projection of tuples */
tstate = begin_tup_output_tupdesc(dest, tupdesc);
reset all: command line and .conf options change defaults on RESET ALL those are restored. show all: GUC + non-GUC. SHOW ALL, RESET ALL Marko Kreen
2001-06-07 06:50:57 +02:00
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
for (i = 0; i < num_guc_variables; i++)
{
Code review for SHOW output changes; fix horology expected files for new SHOW output format.
2002-07-20 17:12:56 +02:00
struct config_generic *conf = guc_variables[i];
pgindent run for 9.0
2010-02-26 03:01:40 +01:00
char *setting;
Code review for SHOW output changes; fix horology expected files for new SHOW output format.
2002-07-20 17:12:56 +02:00
Add a GUC_SUPERUSER_ONLY flag to mark GUC variables that should not be examinable by non-superusers, and use it to protect the recently-added GUC variables for data directory and config files. For now I have only flagged those variables that could be used to deduce something about the server's filesystem layout, but possibly we should also mark vars related to logging settings and other admin-only information?
2004-10-22 21:48:19 +02:00
if ((conf->flags & GUC_NO_SHOW_ALL) ||
((conf->flags & GUC_SUPERUSER_ONLY) && !am_superuser))
Code review for SHOW output changes; fix horology expected files for new SHOW output format.
2002-07-20 17:12:56 +02:00
continue;
> 2. This patch includes the same Table Function API fixes that I > submitted on July 9: > > http://archives.postgresql.org/pgsql-patches/2002-07/msg00056.php > > Please disregard that one *if* this one is applied. If this one is > rejected please go ahead with the July 9th patch. The July 9th Table Function API patch mentioned above is now in CVS, so here is an updated version of the guc patch which should apply cleanly against CVS tip. Joe Conway
2002-07-20 07:49:28 +02:00
/* assign to the values array */
Change do_tup_output() to take Datum/isnull arrays instead of a char * array, so it doesn't go through BuildTupleFromCStrings. This is more or less a wash for current uses, but will avoid inefficiency for planned changes to EXPLAIN. Robert Haas
2009-07-22 19:00:23 +02:00
values[0] = PointerGetDatum(cstring_to_text(conf->name));
setting = _ShowOption(conf, true);
if (setting)
{
values[1] = PointerGetDatum(cstring_to_text(setting));
isnull[1] = false;
}
else
{
values[1] = PointerGetDatum(NULL);
isnull[1] = true;
}
values[2] = PointerGetDatum(cstring_to_text(conf->short_desc));
> 2. This patch includes the same Table Function API fixes that I > submitted on July 9: > > http://archives.postgresql.org/pgsql-patches/2002-07/msg00056.php > > Please disregard that one *if* this one is applied. If this one is > rejected please go ahead with the July 9th patch. The July 9th Table Function API patch mentioned above is now in CVS, so here is an updated version of the guc patch which should apply cleanly against CVS tip. Joe Conway
2002-07-20 07:49:28 +02:00
/* send it to dest */
Change do_tup_output() to take Datum/isnull arrays instead of a char * array, so it doesn't go through BuildTupleFromCStrings. This is more or less a wash for current uses, but will avoid inefficiency for planned changes to EXPLAIN. Robert Haas
2009-07-22 19:00:23 +02:00
do_tup_output(tstate, values, isnull);
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
Revert (again) GUC patch to return commented fields to their default values, due to concern about the patch.
2006-08-14 04:27:27 +02:00
/* clean up */
Change do_tup_output() to take Datum/isnull arrays instead of a char * array, so it doesn't go through BuildTupleFromCStrings. This is more or less a wash for current uses, but will avoid inefficiency for planned changes to EXPLAIN. Robert Haas
2009-07-22 19:00:23 +02:00
pfree(DatumGetPointer(values[0]));
if (setting)
{
pfree(setting);
pfree(DatumGetPointer(values[1]));
}
pfree(DatumGetPointer(values[2]));
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
}
> 2. This patch includes the same Table Function API fixes that I > submitted on July 9: > > http://archives.postgresql.org/pgsql-patches/2002-07/msg00056.php > > Please disregard that one *if* this one is applied. If this one is > rejected please go ahead with the July 9th patch. The July 9th Table Function API patch mentioned above is now in CVS, so here is an updated version of the guc patch which should apply cleanly against CVS tip. Joe Conway
2002-07-20 07:49:28 +02:00
end_tup_output(tstate);
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
}
reset all: command line and .conf options change defaults on RESET ALL those are restored. show all: GUC + non-GUC. SHOW ALL, RESET ALL Marko Kreen
2001-06-07 06:50:57 +02:00
> 2. This patch includes the same Table Function API fixes that I > submitted on July 9: > > http://archives.postgresql.org/pgsql-patches/2002-07/msg00056.php > > Please disregard that one *if* this one is applied. If this one is > rejected please go ahead with the July 9th patch. The July 9th Table Function API patch mentioned above is now in CVS, so here is an updated version of the guc patch which should apply cleanly against CVS tip. Joe Conway
2002-07-20 07:49:28 +02:00
/*
Code review for SHOW output changes; fix horology expected files for new SHOW output format.
2002-07-20 17:12:56 +02:00
* Return GUC variable value by name; optionally return canonical
* form of name. Return value is palloc'd.
> 2. This patch includes the same Table Function API fixes that I > submitted on July 9: > > http://archives.postgresql.org/pgsql-patches/2002-07/msg00056.php > > Please disregard that one *if* this one is applied. If this one is > rejected please go ahead with the July 9th patch. The July 9th Table Function API patch mentioned above is now in CVS, so here is an updated version of the guc patch which should apply cleanly against CVS tip. Joe Conway
2002-07-20 07:49:28 +02:00
*/
char *
Code review for SHOW output changes; fix horology expected files for new SHOW output format.
2002-07-20 17:12:56 +02:00
GetConfigOptionByName(const char *name, const char **varname)
> 2. This patch includes the same Table Function API fixes that I > submitted on July 9: > > http://archives.postgresql.org/pgsql-patches/2002-07/msg00056.php > > Please disregard that one *if* this one is applied. If this one is > rejected please go ahead with the July 9th patch. The July 9th Table Function API patch mentioned above is now in CVS, so here is an updated version of the guc patch which should apply cleanly against CVS tip. Joe Conway
2002-07-20 07:49:28 +02:00
{
struct config_generic *record;
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
record = find_option(name, false, ERROR);
> 2. This patch includes the same Table Function API fixes that I > submitted on July 9: > > http://archives.postgresql.org/pgsql-patches/2002-07/msg00056.php > > Please disregard that one *if* this one is applied. If this one is > rejected please go ahead with the July 9th patch. The July 9th Table Function API patch mentioned above is now in CVS, so here is an updated version of the guc patch which should apply cleanly against CVS tip. Joe Conway
2002-07-20 07:49:28 +02:00
if (record == NULL)
Error message editing in backend/utils (except /adt).
2003-07-25 22:18:01 +02:00
ereport(ERROR,
(errcode(ERRCODE_UNDEFINED_OBJECT),
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
errmsg("unrecognized configuration parameter \"%s\"", name)));
Add a GUC_SUPERUSER_ONLY flag to mark GUC variables that should not be examinable by non-superusers, and use it to protect the recently-added GUC variables for data directory and config files. For now I have only flagged those variables that could be used to deduce something about the server's filesystem layout, but possibly we should also mark vars related to logging settings and other admin-only information?
2004-10-22 21:48:19 +02:00
if ((record->flags & GUC_SUPERUSER_ONLY) && !superuser())
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
errmsg("must be superuser to examine \"%s\"", name)));
> 2. This patch includes the same Table Function API fixes that I > submitted on July 9: > > http://archives.postgresql.org/pgsql-patches/2002-07/msg00056.php > > Please disregard that one *if* this one is applied. If this one is > rejected please go ahead with the July 9th patch. The July 9th Table Function API patch mentioned above is now in CVS, so here is an updated version of the guc patch which should apply cleanly against CVS tip. Joe Conway
2002-07-20 07:49:28 +02:00
Code review for SHOW output changes; fix horology expected files for new SHOW output format.
2002-07-20 17:12:56 +02:00
if (varname)
*varname = record->name;
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
return _ShowOption(record, true);
> 2. This patch includes the same Table Function API fixes that I > submitted on July 9: > > http://archives.postgresql.org/pgsql-patches/2002-07/msg00056.php > > Please disregard that one *if* this one is applied. If this one is > rejected please go ahead with the July 9th patch. The July 9th Table Function API patch mentioned above is now in CVS, so here is an updated version of the guc patch which should apply cleanly against CVS tip. Joe Conway
2002-07-20 07:49:28 +02:00
}
/*
Code review for SHOW output changes; fix horology expected files for new SHOW output format.
2002-07-20 17:12:56 +02:00
* Return GUC variable value by variable number; optionally return canonical
* form of name. Return value is palloc'd.
> 2. This patch includes the same Table Function API fixes that I > submitted on July 9: > > http://archives.postgresql.org/pgsql-patches/2002-07/msg00056.php > > Please disregard that one *if* this one is applied. If this one is > rejected please go ahead with the July 9th patch. The July 9th Table Function API patch mentioned above is now in CVS, so here is an updated version of the guc patch which should apply cleanly against CVS tip. Joe Conway
2002-07-20 07:49:28 +02:00
*/
> Joe Conway <mail@joeconway.com> writes: >>ISTM that "source" is worth knowing. > > Hm, possibly. Any other opinions? This version has the seven fields I proposed, including "source". Here's an example that shows why I think it's valuable: regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | on context | user vartype | bool source | default min_val | max_val | regression=# update pg_settings set setting = 'off' where name = 'enable_seqscan'; -[ RECORD 1 ]--- set_config | off regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | session min_val | max_val | regression=# alter user postgres set enable_seqscan to 'off'; ALTER USER (log out and then back in again) regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | user min_val | max_val | In the first case, enable_seqscan is set to its default value. After setting it to off, it is obvious that the value has been changed for the session only. In the third case, you can see that the value has been set specifically for the user. Joe Conway
2003-07-27 06:35:54 +02:00
void
GetConfigOptionByNum(int varnum, const char **values, bool *noshow)
> 2. This patch includes the same Table Function API fixes that I > submitted on July 9: > > http://archives.postgresql.org/pgsql-patches/2002-07/msg00056.php > > Please disregard that one *if* this one is applied. If this one is > rejected please go ahead with the July 9th patch. The July 9th Table Function API patch mentioned above is now in CVS, so here is an updated version of the guc patch which should apply cleanly against CVS tip. Joe Conway
2002-07-20 07:49:28 +02:00
{
pgindent run.
2003-08-04 02:43:34 +02:00
char buffer[256];
struct config_generic *conf;
Here are two patches. The guc_and_tablefunc patch addresses the two changes mentioned above, and also adds a new function to the tablefunc API. The tablefunc API change adds the following function: * Oid foidGetTypeId(Oid foid) - Get a function's typeid given the * function Oid. Use this together with TypeGetTupleDesc() to get a * TupleDesc which is derived from the function's declared return type. In the next post I'll send the contrib/tablefunc patch, which illustrates the usage of this new function. Also attached is a doc patch for this change. The doc patch also adds a function that I failed to document previously. Joe Conway
2002-07-30 18:20:03 +02:00
/* check requested variable number valid */
Assert((varnum >= 0) && (varnum < num_guc_variables));
conf = guc_variables[varnum];
> 2. This patch includes the same Table Function API fixes that I > submitted on July 9: > > http://archives.postgresql.org/pgsql-patches/2002-07/msg00056.php > > Please disregard that one *if* this one is applied. If this one is > rejected please go ahead with the July 9th patch. The July 9th Table Function API patch mentioned above is now in CVS, so here is an updated version of the guc patch which should apply cleanly against CVS tip. Joe Conway
2002-07-20 07:49:28 +02:00
Here are two patches. The guc_and_tablefunc patch addresses the two changes mentioned above, and also adds a new function to the tablefunc API. The tablefunc API change adds the following function: * Oid foidGetTypeId(Oid foid) - Get a function's typeid given the * function Oid. Use this together with TypeGetTupleDesc() to get a * TupleDesc which is derived from the function's declared return type. In the next post I'll send the contrib/tablefunc patch, which illustrates the usage of this new function. Also attached is a doc patch for this change. The doc patch also adds a function that I failed to document previously. Joe Conway
2002-07-30 18:20:03 +02:00
if (noshow)
Add a GUC_SUPERUSER_ONLY flag to mark GUC variables that should not be examinable by non-superusers, and use it to protect the recently-added GUC variables for data directory and config files. For now I have only flagged those variables that could be used to deduce something about the server's filesystem layout, but possibly we should also mark vars related to logging settings and other admin-only information?
2004-10-22 21:48:19 +02:00
{
if ((conf->flags & GUC_NO_SHOW_ALL) ||
((conf->flags & GUC_SUPERUSER_ONLY) && !superuser()))
*noshow = true;
else
*noshow = false;
}
Here are two patches. The guc_and_tablefunc patch addresses the two changes mentioned above, and also adds a new function to the tablefunc API. The tablefunc API change adds the following function: * Oid foidGetTypeId(Oid foid) - Get a function's typeid given the * function Oid. Use this together with TypeGetTupleDesc() to get a * TupleDesc which is derived from the function's declared return type. In the next post I'll send the contrib/tablefunc patch, which illustrates the usage of this new function. Also attached is a doc patch for this change. The doc patch also adds a function that I failed to document previously. Joe Conway
2002-07-30 18:20:03 +02:00
> Joe Conway <mail@joeconway.com> writes: >>ISTM that "source" is worth knowing. > > Hm, possibly. Any other opinions? This version has the seven fields I proposed, including "source". Here's an example that shows why I think it's valuable: regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | on context | user vartype | bool source | default min_val | max_val | regression=# update pg_settings set setting = 'off' where name = 'enable_seqscan'; -[ RECORD 1 ]--- set_config | off regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | session min_val | max_val | regression=# alter user postgres set enable_seqscan to 'off'; ALTER USER (log out and then back in again) regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | user min_val | max_val | In the first case, enable_seqscan is set to its default value. After setting it to off, it is obvious that the value has been changed for the session only. In the third case, you can see that the value has been set specifically for the user. Joe Conway
2003-07-27 06:35:54 +02:00
/* first get the generic attributes */
/* name */
values[0] = conf->name;
/* setting : use _ShowOption in order to avoid duplicating the logic */
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
values[1] = _ShowOption(conf, false);
/* unit */
if (conf->vartype == PGC_INT)
{
The attached patch changes units of the some default values in postgresql.conf. - shared_buffers = 32000kB => 32MB - temp_buffers = 8000kB => 8MB - wal_buffers = 8 => 64kB The code of initdb was a bit modified to write MB-unit values. Values greater than 8000kB are rounded out to MB. GUC_UNIT_XBLOCKS is added for wal_buffers. It is like GUC_UNIT_BLOCKS, but uses XLOG_BLCKSZ instead of BLCKSZ. Also, I cleaned up the test of GUC_UNIT_* flags in preparation to add more unit flags in less bits. ITAGAKI Takahiro
2006-10-03 23:11:55 +02:00
static char buf[8];
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
The attached patch changes units of the some default values in postgresql.conf. - shared_buffers = 32000kB => 32MB - temp_buffers = 8000kB => 8MB - wal_buffers = 8 => 64kB The code of initdb was a bit modified to write MB-unit values. Values greater than 8000kB are rounded out to MB. GUC_UNIT_XBLOCKS is added for wal_buffers. It is like GUC_UNIT_BLOCKS, but uses XLOG_BLCKSZ instead of BLCKSZ. Also, I cleaned up the test of GUC_UNIT_* flags in preparation to add more unit flags in less bits. ITAGAKI Takahiro
2006-10-03 23:11:55 +02:00
switch (conf->flags & (GUC_UNIT_MEMORY | GUC_UNIT_TIME))
{
case GUC_UNIT_KB:
values[2] = "kB";
break;
case GUC_UNIT_BLOCKS:
pgindent run for 8.2.
2006-10-04 02:30:14 +02:00
snprintf(buf, sizeof(buf), "%dkB", BLCKSZ / 1024);
The attached patch changes units of the some default values in postgresql.conf. - shared_buffers = 32000kB => 32MB - temp_buffers = 8000kB => 8MB - wal_buffers = 8 => 64kB The code of initdb was a bit modified to write MB-unit values. Values greater than 8000kB are rounded out to MB. GUC_UNIT_XBLOCKS is added for wal_buffers. It is like GUC_UNIT_BLOCKS, but uses XLOG_BLCKSZ instead of BLCKSZ. Also, I cleaned up the test of GUC_UNIT_* flags in preparation to add more unit flags in less bits. ITAGAKI Takahiro
2006-10-03 23:11:55 +02:00
values[2] = buf;
break;
case GUC_UNIT_XBLOCKS:
pgindent run for 8.2.
2006-10-04 02:30:14 +02:00
snprintf(buf, sizeof(buf), "%dkB", XLOG_BLCKSZ / 1024);
The attached patch changes units of the some default values in postgresql.conf. - shared_buffers = 32000kB => 32MB - temp_buffers = 8000kB => 8MB - wal_buffers = 8 => 64kB The code of initdb was a bit modified to write MB-unit values. Values greater than 8000kB are rounded out to MB. GUC_UNIT_XBLOCKS is added for wal_buffers. It is like GUC_UNIT_BLOCKS, but uses XLOG_BLCKSZ instead of BLCKSZ. Also, I cleaned up the test of GUC_UNIT_* flags in preparation to add more unit flags in less bits. ITAGAKI Takahiro
2006-10-03 23:11:55 +02:00
values[2] = buf;
break;
case GUC_UNIT_MS:
values[2] = "ms";
break;
case GUC_UNIT_S:
values[2] = "s";
break;
case GUC_UNIT_MIN:
values[2] = "min";
break;
default:
values[2] = "";
break;
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
}
}
else
values[2] = NULL;
> Joe Conway <mail@joeconway.com> writes: >>ISTM that "source" is worth knowing. > > Hm, possibly. Any other opinions? This version has the seven fields I proposed, including "source". Here's an example that shows why I think it's valuable: regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | on context | user vartype | bool source | default min_val | max_val | regression=# update pg_settings set setting = 'off' where name = 'enable_seqscan'; -[ RECORD 1 ]--- set_config | off regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | session min_val | max_val | regression=# alter user postgres set enable_seqscan to 'off'; ALTER USER (log out and then back in again) regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | user min_val | max_val | In the first case, enable_seqscan is set to its default value. After setting it to off, it is obvious that the value has been changed for the session only. In the third case, you can see that the value has been set specifically for the user. Joe Conway
2003-07-27 06:35:54 +02:00
Added new group of read-only GUC variables to allow simple access to certain compile-time options (FUNC_MAX_ARGS, INDEX_MAX_KEYS, NAMEDATALEN, BLCKSZ, HAVE_INT64_TIMESTAMP). Also added "category", "short_desc", and "extra_desc" to the pg_settings view. Per recent discussion here: http://archives.postgresql.org/pgsql-patches/2003-11/msg00363.php
2003-12-03 19:52:00 +01:00
/* group */
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
values[3] = config_group_names[conf->group];
Added new group of read-only GUC variables to allow simple access to certain compile-time options (FUNC_MAX_ARGS, INDEX_MAX_KEYS, NAMEDATALEN, BLCKSZ, HAVE_INT64_TIMESTAMP). Also added "category", "short_desc", and "extra_desc" to the pg_settings view. Per recent discussion here: http://archives.postgresql.org/pgsql-patches/2003-11/msg00363.php
2003-12-03 19:52:00 +01:00
/* short_desc */
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
values[4] = conf->short_desc;
Added new group of read-only GUC variables to allow simple access to certain compile-time options (FUNC_MAX_ARGS, INDEX_MAX_KEYS, NAMEDATALEN, BLCKSZ, HAVE_INT64_TIMESTAMP). Also added "category", "short_desc", and "extra_desc" to the pg_settings view. Per recent discussion here: http://archives.postgresql.org/pgsql-patches/2003-11/msg00363.php
2003-12-03 19:52:00 +01:00
/* extra_desc */
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
values[5] = conf->long_desc;
Added new group of read-only GUC variables to allow simple access to certain compile-time options (FUNC_MAX_ARGS, INDEX_MAX_KEYS, NAMEDATALEN, BLCKSZ, HAVE_INT64_TIMESTAMP). Also added "category", "short_desc", and "extra_desc" to the pg_settings view. Per recent discussion here: http://archives.postgresql.org/pgsql-patches/2003-11/msg00363.php
2003-12-03 19:52:00 +01:00
> Joe Conway <mail@joeconway.com> writes: >>ISTM that "source" is worth knowing. > > Hm, possibly. Any other opinions? This version has the seven fields I proposed, including "source". Here's an example that shows why I think it's valuable: regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | on context | user vartype | bool source | default min_val | max_val | regression=# update pg_settings set setting = 'off' where name = 'enable_seqscan'; -[ RECORD 1 ]--- set_config | off regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | session min_val | max_val | regression=# alter user postgres set enable_seqscan to 'off'; ALTER USER (log out and then back in again) regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | user min_val | max_val | In the first case, enable_seqscan is set to its default value. After setting it to off, it is obvious that the value has been changed for the session only. In the third case, you can see that the value has been set specifically for the user. Joe Conway
2003-07-27 06:35:54 +02:00
/* context */
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
values[6] = GucContext_Names[conf->context];
> Joe Conway <mail@joeconway.com> writes: >>ISTM that "source" is worth knowing. > > Hm, possibly. Any other opinions? This version has the seven fields I proposed, including "source". Here's an example that shows why I think it's valuable: regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | on context | user vartype | bool source | default min_val | max_val | regression=# update pg_settings set setting = 'off' where name = 'enable_seqscan'; -[ RECORD 1 ]--- set_config | off regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | session min_val | max_val | regression=# alter user postgres set enable_seqscan to 'off'; ALTER USER (log out and then back in again) regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | user min_val | max_val | In the first case, enable_seqscan is set to its default value. After setting it to off, it is obvious that the value has been changed for the session only. In the third case, you can see that the value has been set specifically for the user. Joe Conway
2003-07-27 06:35:54 +02:00
/* vartype */
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
values[7] = config_type_names[conf->vartype];
> Joe Conway <mail@joeconway.com> writes: >>ISTM that "source" is worth knowing. > > Hm, possibly. Any other opinions? This version has the seven fields I proposed, including "source". Here's an example that shows why I think it's valuable: regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | on context | user vartype | bool source | default min_val | max_val | regression=# update pg_settings set setting = 'off' where name = 'enable_seqscan'; -[ RECORD 1 ]--- set_config | off regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | session min_val | max_val | regression=# alter user postgres set enable_seqscan to 'off'; ALTER USER (log out and then back in again) regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | user min_val | max_val | In the first case, enable_seqscan is set to its default value. After setting it to off, it is obvious that the value has been changed for the session only. In the third case, you can see that the value has been set specifically for the user. Joe Conway
2003-07-27 06:35:54 +02:00
/* source */
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
values[8] = GucSource_Names[conf->source];
> Joe Conway <mail@joeconway.com> writes: >>ISTM that "source" is worth knowing. > > Hm, possibly. Any other opinions? This version has the seven fields I proposed, including "source". Here's an example that shows why I think it's valuable: regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | on context | user vartype | bool source | default min_val | max_val | regression=# update pg_settings set setting = 'off' where name = 'enable_seqscan'; -[ RECORD 1 ]--- set_config | off regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | session min_val | max_val | regression=# alter user postgres set enable_seqscan to 'off'; ALTER USER (log out and then back in again) regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | user min_val | max_val | In the first case, enable_seqscan is set to its default value. After setting it to off, it is obvious that the value has been changed for the session only. In the third case, you can see that the value has been set specifically for the user. Joe Conway
2003-07-27 06:35:54 +02:00
/* now get the type specifc attributes */
switch (conf->vartype)
{
case PGC_BOOL:
{
Add columns boot_val and reset_val to the pg_settings view, to expose the value a parameter has at server start and will have after RESET, respectively. Greg Smith, with some modifications by me.
2008-10-06 15:05:40 +02:00
struct config_bool *lconf = (struct config_bool *) conf;
> Joe Conway <mail@joeconway.com> writes: >>ISTM that "source" is worth knowing. > > Hm, possibly. Any other opinions? This version has the seven fields I proposed, including "source". Here's an example that shows why I think it's valuable: regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | on context | user vartype | bool source | default min_val | max_val | regression=# update pg_settings set setting = 'off' where name = 'enable_seqscan'; -[ RECORD 1 ]--- set_config | off regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | session min_val | max_val | regression=# alter user postgres set enable_seqscan to 'off'; ALTER USER (log out and then back in again) regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | user min_val | max_val | In the first case, enable_seqscan is set to its default value. After setting it to off, it is obvious that the value has been changed for the session only. In the third case, you can see that the value has been set specifically for the user. Joe Conway
2003-07-27 06:35:54 +02:00
/* min_val */
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
values[9] = NULL;
> Joe Conway <mail@joeconway.com> writes: >>ISTM that "source" is worth knowing. > > Hm, possibly. Any other opinions? This version has the seven fields I proposed, including "source". Here's an example that shows why I think it's valuable: regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | on context | user vartype | bool source | default min_val | max_val | regression=# update pg_settings set setting = 'off' where name = 'enable_seqscan'; -[ RECORD 1 ]--- set_config | off regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | session min_val | max_val | regression=# alter user postgres set enable_seqscan to 'off'; ALTER USER (log out and then back in again) regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | user min_val | max_val | In the first case, enable_seqscan is set to its default value. After setting it to off, it is obvious that the value has been changed for the session only. In the third case, you can see that the value has been set specifically for the user. Joe Conway
2003-07-27 06:35:54 +02:00
/* max_val */
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
values[10] = NULL;
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
/* enumvals */
values[11] = NULL;
Add columns boot_val and reset_val to the pg_settings view, to expose the value a parameter has at server start and will have after RESET, respectively. Greg Smith, with some modifications by me.
2008-10-06 15:05:40 +02:00
/* boot_val */
values[12] = pstrdup(lconf->boot_val ? "on" : "off");
/* reset_val */
values[13] = pstrdup(lconf->reset_val ? "on" : "off");
> Joe Conway <mail@joeconway.com> writes: >>ISTM that "source" is worth knowing. > > Hm, possibly. Any other opinions? This version has the seven fields I proposed, including "source". Here's an example that shows why I think it's valuable: regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | on context | user vartype | bool source | default min_val | max_val | regression=# update pg_settings set setting = 'off' where name = 'enable_seqscan'; -[ RECORD 1 ]--- set_config | off regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | session min_val | max_val | regression=# alter user postgres set enable_seqscan to 'off'; ALTER USER (log out and then back in again) regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | user min_val | max_val | In the first case, enable_seqscan is set to its default value. After setting it to off, it is obvious that the value has been changed for the session only. In the third case, you can see that the value has been set specifically for the user. Joe Conway
2003-07-27 06:35:54 +02:00
}
break;
case PGC_INT:
{
struct config_int *lconf = (struct config_int *) conf;
/* min_val */
snprintf(buffer, sizeof(buffer), "%d", lconf->min);
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
values[9] = pstrdup(buffer);
> Joe Conway <mail@joeconway.com> writes: >>ISTM that "source" is worth knowing. > > Hm, possibly. Any other opinions? This version has the seven fields I proposed, including "source". Here's an example that shows why I think it's valuable: regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | on context | user vartype | bool source | default min_val | max_val | regression=# update pg_settings set setting = 'off' where name = 'enable_seqscan'; -[ RECORD 1 ]--- set_config | off regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | session min_val | max_val | regression=# alter user postgres set enable_seqscan to 'off'; ALTER USER (log out and then back in again) regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | user min_val | max_val | In the first case, enable_seqscan is set to its default value. After setting it to off, it is obvious that the value has been changed for the session only. In the third case, you can see that the value has been set specifically for the user. Joe Conway
2003-07-27 06:35:54 +02:00
/* max_val */
snprintf(buffer, sizeof(buffer), "%d", lconf->max);
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
values[10] = pstrdup(buffer);
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
/* enumvals */
values[11] = NULL;
Add columns boot_val and reset_val to the pg_settings view, to expose the value a parameter has at server start and will have after RESET, respectively. Greg Smith, with some modifications by me.
2008-10-06 15:05:40 +02:00
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
/* boot_val */
snprintf(buffer, sizeof(buffer), "%d", lconf->boot_val);
values[12] = pstrdup(buffer);
Add columns boot_val and reset_val to the pg_settings view, to expose the value a parameter has at server start and will have after RESET, respectively. Greg Smith, with some modifications by me.
2008-10-06 15:05:40 +02:00
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
/* reset_val */
snprintf(buffer, sizeof(buffer), "%d", lconf->reset_val);
values[13] = pstrdup(buffer);
> Joe Conway <mail@joeconway.com> writes: >>ISTM that "source" is worth knowing. > > Hm, possibly. Any other opinions? This version has the seven fields I proposed, including "source". Here's an example that shows why I think it's valuable: regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | on context | user vartype | bool source | default min_val | max_val | regression=# update pg_settings set setting = 'off' where name = 'enable_seqscan'; -[ RECORD 1 ]--- set_config | off regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | session min_val | max_val | regression=# alter user postgres set enable_seqscan to 'off'; ALTER USER (log out and then back in again) regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | user min_val | max_val | In the first case, enable_seqscan is set to its default value. After setting it to off, it is obvious that the value has been changed for the session only. In the third case, you can see that the value has been set specifically for the user. Joe Conway
2003-07-27 06:35:54 +02:00
}
break;
case PGC_REAL:
{
struct config_real *lconf = (struct config_real *) conf;
/* min_val */
snprintf(buffer, sizeof(buffer), "%g", lconf->min);
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
values[9] = pstrdup(buffer);
> Joe Conway <mail@joeconway.com> writes: >>ISTM that "source" is worth knowing. > > Hm, possibly. Any other opinions? This version has the seven fields I proposed, including "source". Here's an example that shows why I think it's valuable: regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | on context | user vartype | bool source | default min_val | max_val | regression=# update pg_settings set setting = 'off' where name = 'enable_seqscan'; -[ RECORD 1 ]--- set_config | off regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | session min_val | max_val | regression=# alter user postgres set enable_seqscan to 'off'; ALTER USER (log out and then back in again) regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | user min_val | max_val | In the first case, enable_seqscan is set to its default value. After setting it to off, it is obvious that the value has been changed for the session only. In the third case, you can see that the value has been set specifically for the user. Joe Conway
2003-07-27 06:35:54 +02:00
/* max_val */
snprintf(buffer, sizeof(buffer), "%g", lconf->max);
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
values[10] = pstrdup(buffer);
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
/* enumvals */
values[11] = NULL;
Add columns boot_val and reset_val to the pg_settings view, to expose the value a parameter has at server start and will have after RESET, respectively. Greg Smith, with some modifications by me.
2008-10-06 15:05:40 +02:00
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
/* boot_val */
snprintf(buffer, sizeof(buffer), "%g", lconf->boot_val);
values[12] = pstrdup(buffer);
Add columns boot_val and reset_val to the pg_settings view, to expose the value a parameter has at server start and will have after RESET, respectively. Greg Smith, with some modifications by me.
2008-10-06 15:05:40 +02:00
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
/* reset_val */
snprintf(buffer, sizeof(buffer), "%g", lconf->reset_val);
values[13] = pstrdup(buffer);
> Joe Conway <mail@joeconway.com> writes: >>ISTM that "source" is worth knowing. > > Hm, possibly. Any other opinions? This version has the seven fields I proposed, including "source". Here's an example that shows why I think it's valuable: regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | on context | user vartype | bool source | default min_val | max_val | regression=# update pg_settings set setting = 'off' where name = 'enable_seqscan'; -[ RECORD 1 ]--- set_config | off regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | session min_val | max_val | regression=# alter user postgres set enable_seqscan to 'off'; ALTER USER (log out and then back in again) regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | user min_val | max_val | In the first case, enable_seqscan is set to its default value. After setting it to off, it is obvious that the value has been changed for the session only. In the third case, you can see that the value has been set specifically for the user. Joe Conway
2003-07-27 06:35:54 +02:00
}
break;
case PGC_STRING:
{
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
struct config_string *lconf = (struct config_string *) conf;
Add columns boot_val and reset_val to the pg_settings view, to expose the value a parameter has at server start and will have after RESET, respectively. Greg Smith, with some modifications by me.
2008-10-06 15:05:40 +02:00
> Joe Conway <mail@joeconway.com> writes: >>ISTM that "source" is worth knowing. > > Hm, possibly. Any other opinions? This version has the seven fields I proposed, including "source". Here's an example that shows why I think it's valuable: regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | on context | user vartype | bool source | default min_val | max_val | regression=# update pg_settings set setting = 'off' where name = 'enable_seqscan'; -[ RECORD 1 ]--- set_config | off regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | session min_val | max_val | regression=# alter user postgres set enable_seqscan to 'off'; ALTER USER (log out and then back in again) regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | user min_val | max_val | In the first case, enable_seqscan is set to its default value. After setting it to off, it is obvious that the value has been changed for the session only. In the third case, you can see that the value has been set specifically for the user. Joe Conway
2003-07-27 06:35:54 +02:00
/* min_val */
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
values[9] = NULL;
> Joe Conway <mail@joeconway.com> writes: >>ISTM that "source" is worth knowing. > > Hm, possibly. Any other opinions? This version has the seven fields I proposed, including "source". Here's an example that shows why I think it's valuable: regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | on context | user vartype | bool source | default min_val | max_val | regression=# update pg_settings set setting = 'off' where name = 'enable_seqscan'; -[ RECORD 1 ]--- set_config | off regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | session min_val | max_val | regression=# alter user postgres set enable_seqscan to 'off'; ALTER USER (log out and then back in again) regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | user min_val | max_val | In the first case, enable_seqscan is set to its default value. After setting it to off, it is obvious that the value has been changed for the session only. In the third case, you can see that the value has been set specifically for the user. Joe Conway
2003-07-27 06:35:54 +02:00
/* max_val */
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
values[10] = NULL;
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
/* enumvals */
values[11] = NULL;
Add columns boot_val and reset_val to the pg_settings view, to expose the value a parameter has at server start and will have after RESET, respectively. Greg Smith, with some modifications by me.
2008-10-06 15:05:40 +02:00
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
/* boot_val */
if (lconf->boot_val == NULL)
values[12] = NULL;
Add columns boot_val and reset_val to the pg_settings view, to expose the value a parameter has at server start and will have after RESET, respectively. Greg Smith, with some modifications by me.
2008-10-06 15:05:40 +02:00
else
values[12] = pstrdup(lconf->boot_val);
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
/* reset_val */
if (lconf->reset_val == NULL)
values[13] = NULL;
Add columns boot_val and reset_val to the pg_settings view, to expose the value a parameter has at server start and will have after RESET, respectively. Greg Smith, with some modifications by me.
2008-10-06 15:05:40 +02:00
else
values[13] = pstrdup(lconf->reset_val);
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
}
break;
case PGC_ENUM:
{
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
struct config_enum *lconf = (struct config_enum *) conf;
Add columns boot_val and reset_val to the pg_settings view, to expose the value a parameter has at server start and will have after RESET, respectively. Greg Smith, with some modifications by me.
2008-10-06 15:05:40 +02:00
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
/* min_val */
values[9] = NULL;
/* max_val */
values[10] = NULL;
/* enumvals */
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
/*
* NOTE! enumvals with double quotes in them are not
* supported!
*/
Minor code embellishments.
2008-12-02 03:00:32 +01:00
values[11] = config_enum_get_options((struct config_enum *) conf,
"{\"", "\"}", "\",\"");
Add columns boot_val and reset_val to the pg_settings view, to expose the value a parameter has at server start and will have after RESET, respectively. Greg Smith, with some modifications by me.
2008-10-06 15:05:40 +02:00
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
/* boot_val */
Minor code embellishments.
2008-12-02 03:00:32 +01:00
values[12] = pstrdup(config_enum_lookup_by_value(lconf,
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
lconf->boot_val));
Add columns boot_val and reset_val to the pg_settings view, to expose the value a parameter has at server start and will have after RESET, respectively. Greg Smith, with some modifications by me.
2008-10-06 15:05:40 +02:00
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
/* reset_val */
Minor code embellishments.
2008-12-02 03:00:32 +01:00
values[13] = pstrdup(config_enum_lookup_by_value(lconf,
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
lconf->reset_val));
> Joe Conway <mail@joeconway.com> writes: >>ISTM that "source" is worth knowing. > > Hm, possibly. Any other opinions? This version has the seven fields I proposed, including "source". Here's an example that shows why I think it's valuable: regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | on context | user vartype | bool source | default min_val | max_val | regression=# update pg_settings set setting = 'off' where name = 'enable_seqscan'; -[ RECORD 1 ]--- set_config | off regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | session min_val | max_val | regression=# alter user postgres set enable_seqscan to 'off'; ALTER USER (log out and then back in again) regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | user min_val | max_val | In the first case, enable_seqscan is set to its default value. After setting it to off, it is obvious that the value has been changed for the session only. In the third case, you can see that the value has been set specifically for the user. Joe Conway
2003-07-27 06:35:54 +02:00
}
break;
default:
{
/*
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
* should never get here, but in case we do, set 'em to NULL
> Joe Conway <mail@joeconway.com> writes: >>ISTM that "source" is worth knowing. > > Hm, possibly. Any other opinions? This version has the seven fields I proposed, including "source". Here's an example that shows why I think it's valuable: regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | on context | user vartype | bool source | default min_val | max_val | regression=# update pg_settings set setting = 'off' where name = 'enable_seqscan'; -[ RECORD 1 ]--- set_config | off regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | session min_val | max_val | regression=# alter user postgres set enable_seqscan to 'off'; ALTER USER (log out and then back in again) regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | user min_val | max_val | In the first case, enable_seqscan is set to its default value. After setting it to off, it is obvious that the value has been changed for the session only. In the third case, you can see that the value has been set specifically for the user. Joe Conway
2003-07-27 06:35:54 +02:00
*/
/* min_val */
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
values[9] = NULL;
> Joe Conway <mail@joeconway.com> writes: >>ISTM that "source" is worth knowing. > > Hm, possibly. Any other opinions? This version has the seven fields I proposed, including "source". Here's an example that shows why I think it's valuable: regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | on context | user vartype | bool source | default min_val | max_val | regression=# update pg_settings set setting = 'off' where name = 'enable_seqscan'; -[ RECORD 1 ]--- set_config | off regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | session min_val | max_val | regression=# alter user postgres set enable_seqscan to 'off'; ALTER USER (log out and then back in again) regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | user min_val | max_val | In the first case, enable_seqscan is set to its default value. After setting it to off, it is obvious that the value has been changed for the session only. In the third case, you can see that the value has been set specifically for the user. Joe Conway
2003-07-27 06:35:54 +02:00
/* max_val */
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
values[10] = NULL;
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
/* enumvals */
values[11] = NULL;
Add columns boot_val and reset_val to the pg_settings view, to expose the value a parameter has at server start and will have after RESET, respectively. Greg Smith, with some modifications by me.
2008-10-06 15:05:40 +02:00
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
/* boot_val */
values[12] = NULL;
Add columns boot_val and reset_val to the pg_settings view, to expose the value a parameter has at server start and will have after RESET, respectively. Greg Smith, with some modifications by me.
2008-10-06 15:05:40 +02:00
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
/* reset_val */
values[13] = NULL;
> Joe Conway <mail@joeconway.com> writes: >>ISTM that "source" is worth knowing. > > Hm, possibly. Any other opinions? This version has the seven fields I proposed, including "source". Here's an example that shows why I think it's valuable: regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | on context | user vartype | bool source | default min_val | max_val | regression=# update pg_settings set setting = 'off' where name = 'enable_seqscan'; -[ RECORD 1 ]--- set_config | off regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | session min_val | max_val | regression=# alter user postgres set enable_seqscan to 'off'; ALTER USER (log out and then back in again) regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | user min_val | max_val | In the first case, enable_seqscan is set to its default value. After setting it to off, it is obvious that the value has been changed for the session only. In the third case, you can see that the value has been set specifically for the user. Joe Conway
2003-07-27 06:35:54 +02:00
}
break;
}
Add "source file" and "source line" information to each GUC variable. initdb forced due to changes in the pg_settings view. Magnus Hagander and Alvaro Herrera.
2008-09-10 20:09:20 +02:00
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
/*
* If the setting came from a config file, set the source location. For
* security reasons, we don't show source file/line number for
Only show source file and line numbers to superusers, for consistent security level with other parts of the system. Per gripe from Tom
2008-09-23 23:12:03 +02:00
* non-superusers.
*/
if (conf->source == PGC_S_FILE && superuser())
Add "source file" and "source line" information to each GUC variable. initdb forced due to changes in the pg_settings view. Magnus Hagander and Alvaro Herrera.
2008-09-10 20:09:20 +02:00
{
Add columns boot_val and reset_val to the pg_settings view, to expose the value a parameter has at server start and will have after RESET, respectively. Greg Smith, with some modifications by me.
2008-10-06 15:05:40 +02:00
values[14] = conf->sourcefile;
Add "source file" and "source line" information to each GUC variable. initdb forced due to changes in the pg_settings view. Magnus Hagander and Alvaro Herrera.
2008-09-10 20:09:20 +02:00
snprintf(buffer, sizeof(buffer), "%d", conf->sourceline);
Add columns boot_val and reset_val to the pg_settings view, to expose the value a parameter has at server start and will have after RESET, respectively. Greg Smith, with some modifications by me.
2008-10-06 15:05:40 +02:00
values[15] = pstrdup(buffer);
Add "source file" and "source line" information to each GUC variable. initdb forced due to changes in the pg_settings view. Magnus Hagander and Alvaro Herrera.
2008-09-10 20:09:20 +02:00
}
else
{
Add columns boot_val and reset_val to the pg_settings view, to expose the value a parameter has at server start and will have after RESET, respectively. Greg Smith, with some modifications by me.
2008-10-06 15:05:40 +02:00
values[14] = NULL;
values[15] = NULL;
Add "source file" and "source line" information to each GUC variable. initdb forced due to changes in the pg_settings view. Magnus Hagander and Alvaro Herrera.
2008-09-10 20:09:20 +02:00
}
> 2. This patch includes the same Table Function API fixes that I > submitted on July 9: > > http://archives.postgresql.org/pgsql-patches/2002-07/msg00056.php > > Please disregard that one *if* this one is applied. If this one is > rejected please go ahead with the July 9th patch. The July 9th Table Function API patch mentioned above is now in CVS, so here is an updated version of the guc patch which should apply cleanly against CVS tip. Joe Conway
2002-07-20 07:49:28 +02:00
}
/*
* Return the total number of GUC variables
*/
int
GetNumConfigOptions(void)
{
return num_guc_variables;
}
/*
* show_config_by_name - equiv to SHOW X command but implemented as
* a function.
*/
Datum
show_config_by_name(PG_FUNCTION_ARGS)
{
pgindent run.
2002-09-04 22:31:48 +02:00
char *varname;
char *varval;
> 2. This patch includes the same Table Function API fixes that I > submitted on July 9: > > http://archives.postgresql.org/pgsql-patches/2002-07/msg00056.php > > Please disregard that one *if* this one is applied. If this one is > rejected please go ahead with the July 9th patch. The July 9th Table Function API patch mentioned above is now in CVS, so here is an updated version of the guc patch which should apply cleanly against CVS tip. Joe Conway
2002-07-20 07:49:28 +02:00
/* Get the GUC variable name */
Simplify and standardize conversions between TEXT datums and ordinary C strings. This patch introduces four support functions cstring_to_text, cstring_to_text_with_len, text_to_cstring, and text_to_cstring_buffer, and two macros CStringGetTextDatum and TextDatumGetCString. A number of existing macros that provided variants on these themes were removed. Most of the places that need to make such conversions now require just one function or macro call, in place of the multiple notational layers that used to be needed. There are no longer any direct calls of textout or textin, and we got most of the places that were using handmade conversions via memcpy (there may be a few still lurking, though). This commit doesn't make any serious effort to eliminate transient memory leaks caused by detoasting toasted text objects before they reach text_to_cstring. We changed PG_GETARG_TEXT_P to PG_GETARG_TEXT_PP in a few places where it was easy, but much more could be done. Brendan Jurd and Tom Lane
2008-03-25 23:42:46 +01:00
varname = TextDatumGetCString(PG_GETARG_DATUM(0));
> 2. This patch includes the same Table Function API fixes that I > submitted on July 9: > > http://archives.postgresql.org/pgsql-patches/2002-07/msg00056.php > > Please disregard that one *if* this one is applied. If this one is > rejected please go ahead with the July 9th patch. The July 9th Table Function API patch mentioned above is now in CVS, so here is an updated version of the guc patch which should apply cleanly against CVS tip. Joe Conway
2002-07-20 07:49:28 +02:00
/* Get the value */
Code review for SHOW output changes; fix horology expected files for new SHOW output format.
2002-07-20 17:12:56 +02:00
varval = GetConfigOptionByName(varname, NULL);
> 2. This patch includes the same Table Function API fixes that I > submitted on July 9: > > http://archives.postgresql.org/pgsql-patches/2002-07/msg00056.php > > Please disregard that one *if* this one is applied. If this one is > rejected please go ahead with the July 9th patch. The July 9th Table Function API patch mentioned above is now in CVS, so here is an updated version of the guc patch which should apply cleanly against CVS tip. Joe Conway
2002-07-20 07:49:28 +02:00
/* Convert to text */
Simplify and standardize conversions between TEXT datums and ordinary C strings. This patch introduces four support functions cstring_to_text, cstring_to_text_with_len, text_to_cstring, and text_to_cstring_buffer, and two macros CStringGetTextDatum and TextDatumGetCString. A number of existing macros that provided variants on these themes were removed. Most of the places that need to make such conversions now require just one function or macro call, in place of the multiple notational layers that used to be needed. There are no longer any direct calls of textout or textin, and we got most of the places that were using handmade conversions via memcpy (there may be a few still lurking, though). This commit doesn't make any serious effort to eliminate transient memory leaks caused by detoasting toasted text objects before they reach text_to_cstring. We changed PG_GETARG_TEXT_P to PG_GETARG_TEXT_PP in a few places where it was easy, but much more could be done. Brendan Jurd and Tom Lane
2008-03-25 23:42:46 +01:00
PG_RETURN_TEXT_P(cstring_to_text(varval));
> 2. This patch includes the same Table Function API fixes that I > submitted on July 9: > > http://archives.postgresql.org/pgsql-patches/2002-07/msg00056.php > > Please disregard that one *if* this one is applied. If this one is > rejected please go ahead with the July 9th patch. The July 9th Table Function API patch mentioned above is now in CVS, so here is an updated version of the guc patch which should apply cleanly against CVS tip. Joe Conway
2002-07-20 07:49:28 +02:00
}
As discussed on several occasions previously, the new anonymous composite type capability makes it possible to create a system view based on a table function in a way that is hopefully palatable to everyone. The attached patch takes advantage of this, moving show_all_settings() from contrib/tablefunc into the backend (renamed all_settings(). It is defined as a builtin returning type RECORD. During initdb a system view is created to expose the same information presently available through SHOW ALL. For example: test=# select * from pg_settings where name like '%debug%'; name | setting -----------------------+--------- debug_assertions | on debug_pretty_print | off debug_print_parse | off debug_print_plan | off debug_print_query | off debug_print_rewritten | off wal_debug | 0 (7 rows) Additionally during initdb two rules are created which make it possible to change settings by updating the system view -- a "virtual table" as Tom put it. Here's an example: Joe Conway
2002-08-15 04:51:27 +02:00
/*
* show_all_settings - equiv to SHOW ALL command but implemented as
* a Table Function.
*/
Add columns boot_val and reset_val to the pg_settings view, to expose the value a parameter has at server start and will have after RESET, respectively. Greg Smith, with some modifications by me.
2008-10-06 15:05:40 +02:00
#define NUM_PG_SETTINGS_ATTS 16
> Joe Conway <mail@joeconway.com> writes: >>ISTM that "source" is worth knowing. > > Hm, possibly. Any other opinions? This version has the seven fields I proposed, including "source". Here's an example that shows why I think it's valuable: regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | on context | user vartype | bool source | default min_val | max_val | regression=# update pg_settings set setting = 'off' where name = 'enable_seqscan'; -[ RECORD 1 ]--- set_config | off regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | session min_val | max_val | regression=# alter user postgres set enable_seqscan to 'off'; ALTER USER (log out and then back in again) regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | user min_val | max_val | In the first case, enable_seqscan is set to its default value. After setting it to off, it is obvious that the value has been changed for the session only. In the third case, you can see that the value has been set specifically for the user. Joe Conway
2003-07-27 06:35:54 +02:00
As discussed on several occasions previously, the new anonymous composite type capability makes it possible to create a system view based on a table function in a way that is hopefully palatable to everyone. The attached patch takes advantage of this, moving show_all_settings() from contrib/tablefunc into the backend (renamed all_settings(). It is defined as a builtin returning type RECORD. During initdb a system view is created to expose the same information presently available through SHOW ALL. For example: test=# select * from pg_settings where name like '%debug%'; name | setting -----------------------+--------- debug_assertions | on debug_pretty_print | off debug_print_parse | off debug_print_plan | off debug_print_query | off debug_print_rewritten | off wal_debug | 0 (7 rows) Additionally during initdb two rules are created which make it possible to change settings by updating the system view -- a "virtual table" as Tom put it. Here's an example: Joe Conway
2002-08-15 04:51:27 +02:00
Datum
show_all_settings(PG_FUNCTION_ARGS)
{
pgindent run.
2002-09-04 22:31:48 +02:00
FuncCallContext *funcctx;
TupleDesc tupdesc;
int call_cntr;
int max_calls;
AttInMetadata *attinmeta;
MemoryContext oldcontext;
As discussed on several occasions previously, the new anonymous composite type capability makes it possible to create a system view based on a table function in a way that is hopefully palatable to everyone. The attached patch takes advantage of this, moving show_all_settings() from contrib/tablefunc into the backend (renamed all_settings(). It is defined as a builtin returning type RECORD. During initdb a system view is created to expose the same information presently available through SHOW ALL. For example: test=# select * from pg_settings where name like '%debug%'; name | setting -----------------------+--------- debug_assertions | on debug_pretty_print | off debug_print_parse | off debug_print_plan | off debug_print_query | off debug_print_rewritten | off wal_debug | 0 (7 rows) Additionally during initdb two rules are created which make it possible to change settings by updating the system view -- a "virtual table" as Tom put it. Here's an example: Joe Conway
2002-08-15 04:51:27 +02:00
/* stuff done only on the first call of the function */
pgindent run.
2002-09-04 22:31:48 +02:00
if (SRF_IS_FIRSTCALL())
{
As discussed on several occasions previously, the new anonymous composite type capability makes it possible to create a system view based on a table function in a way that is hopefully palatable to everyone. The attached patch takes advantage of this, moving show_all_settings() from contrib/tablefunc into the backend (renamed all_settings(). It is defined as a builtin returning type RECORD. During initdb a system view is created to expose the same information presently available through SHOW ALL. For example: test=# select * from pg_settings where name like '%debug%'; name | setting -----------------------+--------- debug_assertions | on debug_pretty_print | off debug_print_parse | off debug_print_plan | off debug_print_query | off debug_print_rewritten | off wal_debug | 0 (7 rows) Additionally during initdb two rules are created which make it possible to change settings by updating the system view -- a "virtual table" as Tom put it. Here's an example: Joe Conway
2002-08-15 04:51:27 +02:00
/* create a function context for cross-call persistence */
pgindent run.
2002-09-04 22:31:48 +02:00
funcctx = SRF_FIRSTCALL_INIT();
As discussed on several occasions previously, the new anonymous composite type capability makes it possible to create a system view based on a table function in a way that is hopefully palatable to everyone. The attached patch takes advantage of this, moving show_all_settings() from contrib/tablefunc into the backend (renamed all_settings(). It is defined as a builtin returning type RECORD. During initdb a system view is created to expose the same information presently available through SHOW ALL. For example: test=# select * from pg_settings where name like '%debug%'; name | setting -----------------------+--------- debug_assertions | on debug_pretty_print | off debug_print_parse | off debug_print_plan | off debug_print_query | off debug_print_rewritten | off wal_debug | 0 (7 rows) Additionally during initdb two rules are created which make it possible to change settings by updating the system view -- a "virtual table" as Tom put it. Here's an example: Joe Conway
2002-08-15 04:51:27 +02:00
pgindent run.
2002-09-04 22:31:48 +02:00
/*
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
* switch to memory context appropriate for multiple function calls
pgindent run.
2002-09-04 22:31:48 +02:00
*/
Adjust nodeFunctionscan.c to reset transient memory context between calls to the table function, thus preventing memory leakage accumulation across calls. This means that SRFs need to be careful to distinguish permanent and local storage; adjust code and documentation accordingly. Patch by Joe Conway, very minor tweaks by Tom Lane.
2002-08-29 19:14:33 +02:00
oldcontext = MemoryContextSwitchTo(funcctx->multi_call_memory_ctx);
> Joe Conway <mail@joeconway.com> writes: >>ISTM that "source" is worth knowing. > > Hm, possibly. Any other opinions? This version has the seven fields I proposed, including "source". Here's an example that shows why I think it's valuable: regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | on context | user vartype | bool source | default min_val | max_val | regression=# update pg_settings set setting = 'off' where name = 'enable_seqscan'; -[ RECORD 1 ]--- set_config | off regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | session min_val | max_val | regression=# alter user postgres set enable_seqscan to 'off'; ALTER USER (log out and then back in again) regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | user min_val | max_val | In the first case, enable_seqscan is set to its default value. After setting it to off, it is obvious that the value has been changed for the session only. In the third case, you can see that the value has been set specifically for the user. Joe Conway
2003-07-27 06:35:54 +02:00
/*
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
* need a tuple descriptor representing NUM_PG_SETTINGS_ATTS columns
* of the appropriate types
> Joe Conway <mail@joeconway.com> writes: >>ISTM that "source" is worth knowing. > > Hm, possibly. Any other opinions? This version has the seven fields I proposed, including "source". Here's an example that shows why I think it's valuable: regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | on context | user vartype | bool source | default min_val | max_val | regression=# update pg_settings set setting = 'off' where name = 'enable_seqscan'; -[ RECORD 1 ]--- set_config | off regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | session min_val | max_val | regression=# alter user postgres set enable_seqscan to 'off'; ALTER USER (log out and then back in again) regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | user min_val | max_val | In the first case, enable_seqscan is set to its default value. After setting it to off, it is obvious that the value has been changed for the session only. In the third case, you can see that the value has been set specifically for the user. Joe Conway
2003-07-27 06:35:54 +02:00
*/
tupdesc = CreateTemplateTupleDesc(NUM_PG_SETTINGS_ATTS, false);
As discussed on several occasions previously, the new anonymous composite type capability makes it possible to create a system view based on a table function in a way that is hopefully palatable to everyone. The attached patch takes advantage of this, moving show_all_settings() from contrib/tablefunc into the backend (renamed all_settings(). It is defined as a builtin returning type RECORD. During initdb a system view is created to expose the same information presently available through SHOW ALL. For example: test=# select * from pg_settings where name like '%debug%'; name | setting -----------------------+--------- debug_assertions | on debug_pretty_print | off debug_print_parse | off debug_print_plan | off debug_print_query | off debug_print_rewritten | off wal_debug | 0 (7 rows) Additionally during initdb two rules are created which make it possible to change settings by updating the system view -- a "virtual table" as Tom put it. Here's an example: Joe Conway
2002-08-15 04:51:27 +02:00
TupleDescInitEntry(tupdesc, (AttrNumber) 1, "name",
Replace TupleTableSlot convention for whole-row variables and function results with tuples as ordinary varlena Datums. This commit does not in itself do much for us, except eliminate the horrid memory leak associated with evaluation of whole-row variables. However, it lays the groundwork for allowing composite types as table columns, and perhaps some other useful features as well. Per my proposal of a few days ago.
2004-04-01 23:28:47 +02:00
TEXTOID, -1, 0);
As discussed on several occasions previously, the new anonymous composite type capability makes it possible to create a system view based on a table function in a way that is hopefully palatable to everyone. The attached patch takes advantage of this, moving show_all_settings() from contrib/tablefunc into the backend (renamed all_settings(). It is defined as a builtin returning type RECORD. During initdb a system view is created to expose the same information presently available through SHOW ALL. For example: test=# select * from pg_settings where name like '%debug%'; name | setting -----------------------+--------- debug_assertions | on debug_pretty_print | off debug_print_parse | off debug_print_plan | off debug_print_query | off debug_print_rewritten | off wal_debug | 0 (7 rows) Additionally during initdb two rules are created which make it possible to change settings by updating the system view -- a "virtual table" as Tom put it. Here's an example: Joe Conway
2002-08-15 04:51:27 +02:00
TupleDescInitEntry(tupdesc, (AttrNumber) 2, "setting",
Replace TupleTableSlot convention for whole-row variables and function results with tuples as ordinary varlena Datums. This commit does not in itself do much for us, except eliminate the horrid memory leak associated with evaluation of whole-row variables. However, it lays the groundwork for allowing composite types as table columns, and perhaps some other useful features as well. Per my proposal of a few days ago.
2004-04-01 23:28:47 +02:00
TEXTOID, -1, 0);
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
TupleDescInitEntry(tupdesc, (AttrNumber) 3, "unit",
TEXTOID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber) 4, "category",
Replace TupleTableSlot convention for whole-row variables and function results with tuples as ordinary varlena Datums. This commit does not in itself do much for us, except eliminate the horrid memory leak associated with evaluation of whole-row variables. However, it lays the groundwork for allowing composite types as table columns, and perhaps some other useful features as well. Per my proposal of a few days ago.
2004-04-01 23:28:47 +02:00
TEXTOID, -1, 0);
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
TupleDescInitEntry(tupdesc, (AttrNumber) 5, "short_desc",
Replace TupleTableSlot convention for whole-row variables and function results with tuples as ordinary varlena Datums. This commit does not in itself do much for us, except eliminate the horrid memory leak associated with evaluation of whole-row variables. However, it lays the groundwork for allowing composite types as table columns, and perhaps some other useful features as well. Per my proposal of a few days ago.
2004-04-01 23:28:47 +02:00
TEXTOID, -1, 0);
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
TupleDescInitEntry(tupdesc, (AttrNumber) 6, "extra_desc",
Replace TupleTableSlot convention for whole-row variables and function results with tuples as ordinary varlena Datums. This commit does not in itself do much for us, except eliminate the horrid memory leak associated with evaluation of whole-row variables. However, it lays the groundwork for allowing composite types as table columns, and perhaps some other useful features as well. Per my proposal of a few days ago.
2004-04-01 23:28:47 +02:00
TEXTOID, -1, 0);
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
TupleDescInitEntry(tupdesc, (AttrNumber) 7, "context",
Replace TupleTableSlot convention for whole-row variables and function results with tuples as ordinary varlena Datums. This commit does not in itself do much for us, except eliminate the horrid memory leak associated with evaluation of whole-row variables. However, it lays the groundwork for allowing composite types as table columns, and perhaps some other useful features as well. Per my proposal of a few days ago.
2004-04-01 23:28:47 +02:00
TEXTOID, -1, 0);
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
TupleDescInitEntry(tupdesc, (AttrNumber) 8, "vartype",
Replace TupleTableSlot convention for whole-row variables and function results with tuples as ordinary varlena Datums. This commit does not in itself do much for us, except eliminate the horrid memory leak associated with evaluation of whole-row variables. However, it lays the groundwork for allowing composite types as table columns, and perhaps some other useful features as well. Per my proposal of a few days ago.
2004-04-01 23:28:47 +02:00
TEXTOID, -1, 0);
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
TupleDescInitEntry(tupdesc, (AttrNumber) 9, "source",
Replace TupleTableSlot convention for whole-row variables and function results with tuples as ordinary varlena Datums. This commit does not in itself do much for us, except eliminate the horrid memory leak associated with evaluation of whole-row variables. However, it lays the groundwork for allowing composite types as table columns, and perhaps some other useful features as well. Per my proposal of a few days ago.
2004-04-01 23:28:47 +02:00
TEXTOID, -1, 0);
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
TupleDescInitEntry(tupdesc, (AttrNumber) 10, "min_val",
Replace TupleTableSlot convention for whole-row variables and function results with tuples as ordinary varlena Datums. This commit does not in itself do much for us, except eliminate the horrid memory leak associated with evaluation of whole-row variables. However, it lays the groundwork for allowing composite types as table columns, and perhaps some other useful features as well. Per my proposal of a few days ago.
2004-04-01 23:28:47 +02:00
TEXTOID, -1, 0);
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
TupleDescInitEntry(tupdesc, (AttrNumber) 11, "max_val",
Replace TupleTableSlot convention for whole-row variables and function results with tuples as ordinary varlena Datums. This commit does not in itself do much for us, except eliminate the horrid memory leak associated with evaluation of whole-row variables. However, it lays the groundwork for allowing composite types as table columns, and perhaps some other useful features as well. Per my proposal of a few days ago.
2004-04-01 23:28:47 +02:00
TEXTOID, -1, 0);
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
TupleDescInitEntry(tupdesc, (AttrNumber) 12, "enumvals",
Make the enumvals column of pg_settings be text[] instead of just a comma separated string.
2008-11-21 19:49:24 +01:00
TEXTARRAYOID, -1, 0);
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
TupleDescInitEntry(tupdesc, (AttrNumber) 13, "boot_val",
TEXTOID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber) 14, "reset_val",
TEXTOID, -1, 0);
Add columns boot_val and reset_val to the pg_settings view, to expose the value a parameter has at server start and will have after RESET, respectively. Greg Smith, with some modifications by me.
2008-10-06 15:05:40 +02:00
TupleDescInitEntry(tupdesc, (AttrNumber) 15, "sourcefile",
Add "source file" and "source line" information to each GUC variable. initdb forced due to changes in the pg_settings view. Magnus Hagander and Alvaro Herrera.
2008-09-10 20:09:20 +02:00
TEXTOID, -1, 0);
Add columns boot_val and reset_val to the pg_settings view, to expose the value a parameter has at server start and will have after RESET, respectively. Greg Smith, with some modifications by me.
2008-10-06 15:05:40 +02:00
TupleDescInitEntry(tupdesc, (AttrNumber) 16, "sourceline",
Add "source file" and "source line" information to each GUC variable. initdb forced due to changes in the pg_settings view. Magnus Hagander and Alvaro Herrera.
2008-09-10 20:09:20 +02:00
INT4OID, -1, 0);
As discussed on several occasions previously, the new anonymous composite type capability makes it possible to create a system view based on a table function in a way that is hopefully palatable to everyone. The attached patch takes advantage of this, moving show_all_settings() from contrib/tablefunc into the backend (renamed all_settings(). It is defined as a builtin returning type RECORD. During initdb a system view is created to expose the same information presently available through SHOW ALL. For example: test=# select * from pg_settings where name like '%debug%'; name | setting -----------------------+--------- debug_assertions | on debug_pretty_print | off debug_print_parse | off debug_print_plan | off debug_print_query | off debug_print_rewritten | off wal_debug | 0 (7 rows) Additionally during initdb two rules are created which make it possible to change settings by updating the system view -- a "virtual table" as Tom put it. Here's an example: Joe Conway
2002-08-15 04:51:27 +02:00
/*
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
* Generate attribute metadata needed later to produce tuples from raw
* C strings
As discussed on several occasions previously, the new anonymous composite type capability makes it possible to create a system view based on a table function in a way that is hopefully palatable to everyone. The attached patch takes advantage of this, moving show_all_settings() from contrib/tablefunc into the backend (renamed all_settings(). It is defined as a builtin returning type RECORD. During initdb a system view is created to expose the same information presently available through SHOW ALL. For example: test=# select * from pg_settings where name like '%debug%'; name | setting -----------------------+--------- debug_assertions | on debug_pretty_print | off debug_print_parse | off debug_print_plan | off debug_print_query | off debug_print_rewritten | off wal_debug | 0 (7 rows) Additionally during initdb two rules are created which make it possible to change settings by updating the system view -- a "virtual table" as Tom put it. Here's an example: Joe Conway
2002-08-15 04:51:27 +02:00
*/
attinmeta = TupleDescGetAttInMetadata(tupdesc);
funcctx->attinmeta = attinmeta;
/* total number of tuples to be returned */
funcctx->max_calls = GetNumConfigOptions();
Adjust nodeFunctionscan.c to reset transient memory context between calls to the table function, thus preventing memory leakage accumulation across calls. This means that SRFs need to be careful to distinguish permanent and local storage; adjust code and documentation accordingly. Patch by Joe Conway, very minor tweaks by Tom Lane.
2002-08-29 19:14:33 +02:00
MemoryContextSwitchTo(oldcontext);
pgindent run.
2002-09-04 22:31:48 +02:00
}
As discussed on several occasions previously, the new anonymous composite type capability makes it possible to create a system view based on a table function in a way that is hopefully palatable to everyone. The attached patch takes advantage of this, moving show_all_settings() from contrib/tablefunc into the backend (renamed all_settings(). It is defined as a builtin returning type RECORD. During initdb a system view is created to expose the same information presently available through SHOW ALL. For example: test=# select * from pg_settings where name like '%debug%'; name | setting -----------------------+--------- debug_assertions | on debug_pretty_print | off debug_print_parse | off debug_print_plan | off debug_print_query | off debug_print_rewritten | off wal_debug | 0 (7 rows) Additionally during initdb two rules are created which make it possible to change settings by updating the system view -- a "virtual table" as Tom put it. Here's an example: Joe Conway
2002-08-15 04:51:27 +02:00
/* stuff done on every call of the function */
pgindent run.
2002-09-04 22:31:48 +02:00
funcctx = SRF_PERCALL_SETUP();
As discussed on several occasions previously, the new anonymous composite type capability makes it possible to create a system view based on a table function in a way that is hopefully palatable to everyone. The attached patch takes advantage of this, moving show_all_settings() from contrib/tablefunc into the backend (renamed all_settings(). It is defined as a builtin returning type RECORD. During initdb a system view is created to expose the same information presently available through SHOW ALL. For example: test=# select * from pg_settings where name like '%debug%'; name | setting -----------------------+--------- debug_assertions | on debug_pretty_print | off debug_print_parse | off debug_print_plan | off debug_print_query | off debug_print_rewritten | off wal_debug | 0 (7 rows) Additionally during initdb two rules are created which make it possible to change settings by updating the system view -- a "virtual table" as Tom put it. Here's an example: Joe Conway
2002-08-15 04:51:27 +02:00
call_cntr = funcctx->call_cntr;
max_calls = funcctx->max_calls;
attinmeta = funcctx->attinmeta;
pgindent run.
2002-09-04 22:31:48 +02:00
if (call_cntr < max_calls) /* do when there is more left to send */
{
> Joe Conway <mail@joeconway.com> writes: >>ISTM that "source" is worth knowing. > > Hm, possibly. Any other opinions? This version has the seven fields I proposed, including "source". Here's an example that shows why I think it's valuable: regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | on context | user vartype | bool source | default min_val | max_val | regression=# update pg_settings set setting = 'off' where name = 'enable_seqscan'; -[ RECORD 1 ]--- set_config | off regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | session min_val | max_val | regression=# alter user postgres set enable_seqscan to 'off'; ALTER USER (log out and then back in again) regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | user min_val | max_val | In the first case, enable_seqscan is set to its default value. After setting it to off, it is obvious that the value has been changed for the session only. In the third case, you can see that the value has been set specifically for the user. Joe Conway
2003-07-27 06:35:54 +02:00
char *values[NUM_PG_SETTINGS_ATTS];
As discussed on several occasions previously, the new anonymous composite type capability makes it possible to create a system view based on a table function in a way that is hopefully palatable to everyone. The attached patch takes advantage of this, moving show_all_settings() from contrib/tablefunc into the backend (renamed all_settings(). It is defined as a builtin returning type RECORD. During initdb a system view is created to expose the same information presently available through SHOW ALL. For example: test=# select * from pg_settings where name like '%debug%'; name | setting -----------------------+--------- debug_assertions | on debug_pretty_print | off debug_print_parse | off debug_print_plan | off debug_print_query | off debug_print_rewritten | off wal_debug | 0 (7 rows) Additionally during initdb two rules are created which make it possible to change settings by updating the system view -- a "virtual table" as Tom put it. Here's an example: Joe Conway
2002-08-15 04:51:27 +02:00
bool noshow;
HeapTuple tuple;
Datum result;
/*
* Get the next visible GUC variable name and value
*/
do
{
> Joe Conway <mail@joeconway.com> writes: >>ISTM that "source" is worth knowing. > > Hm, possibly. Any other opinions? This version has the seven fields I proposed, including "source". Here's an example that shows why I think it's valuable: regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | on context | user vartype | bool source | default min_val | max_val | regression=# update pg_settings set setting = 'off' where name = 'enable_seqscan'; -[ RECORD 1 ]--- set_config | off regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | session min_val | max_val | regression=# alter user postgres set enable_seqscan to 'off'; ALTER USER (log out and then back in again) regression=# \x Expanded display is on. regression=# select * from pg_settings where name = 'enable_seqscan'; -[ RECORD 1 ]----------- name | enable_seqscan setting | off context | user vartype | bool source | user min_val | max_val | In the first case, enable_seqscan is set to its default value. After setting it to off, it is obvious that the value has been changed for the session only. In the third case, you can see that the value has been set specifically for the user. Joe Conway
2003-07-27 06:35:54 +02:00
GetConfigOptionByNum(call_cntr, (const char **) values, &noshow);
As discussed on several occasions previously, the new anonymous composite type capability makes it possible to create a system view based on a table function in a way that is hopefully palatable to everyone. The attached patch takes advantage of this, moving show_all_settings() from contrib/tablefunc into the backend (renamed all_settings(). It is defined as a builtin returning type RECORD. During initdb a system view is created to expose the same information presently available through SHOW ALL. For example: test=# select * from pg_settings where name like '%debug%'; name | setting -----------------------+--------- debug_assertions | on debug_pretty_print | off debug_print_parse | off debug_print_plan | off debug_print_query | off debug_print_rewritten | off wal_debug | 0 (7 rows) Additionally during initdb two rules are created which make it possible to change settings by updating the system view -- a "virtual table" as Tom put it. Here's an example: Joe Conway
2002-08-15 04:51:27 +02:00
if (noshow)
{
/* bump the counter and get the next config setting */
call_cntr = ++funcctx->call_cntr;
/* make sure we haven't gone too far now */
if (call_cntr >= max_calls)
pgindent run.
2002-09-04 22:31:48 +02:00
SRF_RETURN_DONE(funcctx);
As discussed on several occasions previously, the new anonymous composite type capability makes it possible to create a system view based on a table function in a way that is hopefully palatable to everyone. The attached patch takes advantage of this, moving show_all_settings() from contrib/tablefunc into the backend (renamed all_settings(). It is defined as a builtin returning type RECORD. During initdb a system view is created to expose the same information presently available through SHOW ALL. For example: test=# select * from pg_settings where name like '%debug%'; name | setting -----------------------+--------- debug_assertions | on debug_pretty_print | off debug_print_parse | off debug_print_plan | off debug_print_query | off debug_print_rewritten | off wal_debug | 0 (7 rows) Additionally during initdb two rules are created which make it possible to change settings by updating the system view -- a "virtual table" as Tom put it. Here's an example: Joe Conway
2002-08-15 04:51:27 +02:00
}
} while (noshow);
/* build a tuple */
tuple = BuildTupleFromCStrings(attinmeta, values);
/* make the tuple into a datum */
Replace TupleTableSlot convention for whole-row variables and function results with tuples as ordinary varlena Datums. This commit does not in itself do much for us, except eliminate the horrid memory leak associated with evaluation of whole-row variables. However, it lays the groundwork for allowing composite types as table columns, and perhaps some other useful features as well. Per my proposal of a few days ago.
2004-04-01 23:28:47 +02:00
result = HeapTupleGetDatum(tuple);
As discussed on several occasions previously, the new anonymous composite type capability makes it possible to create a system view based on a table function in a way that is hopefully palatable to everyone. The attached patch takes advantage of this, moving show_all_settings() from contrib/tablefunc into the backend (renamed all_settings(). It is defined as a builtin returning type RECORD. During initdb a system view is created to expose the same information presently available through SHOW ALL. For example: test=# select * from pg_settings where name like '%debug%'; name | setting -----------------------+--------- debug_assertions | on debug_pretty_print | off debug_print_parse | off debug_print_plan | off debug_print_query | off debug_print_rewritten | off wal_debug | 0 (7 rows) Additionally during initdb two rules are created which make it possible to change settings by updating the system view -- a "virtual table" as Tom put it. Here's an example: Joe Conway
2002-08-15 04:51:27 +02:00
pgindent run.
2002-09-04 22:31:48 +02:00
SRF_RETURN_NEXT(funcctx, result);
}
else
In the continuing saga of FE/BE protocol revisions, add reporting of initial values and runtime changes in selected parameters. This gets rid of the need for an initial 'select pg_client_encoding()' query in libpq, bringing us back to one message transmitted in each direction for a standard connection startup. To allow server version to be sent using the same GUC mechanism that handles other parameters, invent the concept of a never-settable GUC parameter: you can 'show server_version' but it's not settable by any GUC input source. Create 'lc_collate' and 'lc_ctype' never-settable parameters so that people can find out these settings without need for pg_controldata. (These side ideas were all discussed some time ago in pgsql-hackers, but not yet implemented.)
2003-04-25 21:45:10 +02:00
{
/* do when there is no more left */
pgindent run.
2002-09-04 22:31:48 +02:00
SRF_RETURN_DONE(funcctx);
In the continuing saga of FE/BE protocol revisions, add reporting of initial values and runtime changes in selected parameters. This gets rid of the need for an initial 'select pg_client_encoding()' query in libpq, bringing us back to one message transmitted in each direction for a standard connection startup. To allow server version to be sent using the same GUC mechanism that handles other parameters, invent the concept of a never-settable GUC parameter: you can 'show server_version' but it's not settable by any GUC input source. Create 'lc_collate' and 'lc_ctype' never-settable parameters so that people can find out these settings without need for pg_controldata. (These side ideas were all discussed some time ago in pgsql-hackers, but not yet implemented.)
2003-04-25 21:45:10 +02:00
}
As discussed on several occasions previously, the new anonymous composite type capability makes it possible to create a system view based on a table function in a way that is hopefully palatable to everyone. The attached patch takes advantage of this, moving show_all_settings() from contrib/tablefunc into the backend (renamed all_settings(). It is defined as a builtin returning type RECORD. During initdb a system view is created to expose the same information presently available through SHOW ALL. For example: test=# select * from pg_settings where name like '%debug%'; name | setting -----------------------+--------- debug_assertions | on debug_pretty_print | off debug_print_parse | off debug_print_plan | off debug_print_query | off debug_print_rewritten | off wal_debug | 0 (7 rows) Additionally during initdb two rules are created which make it possible to change settings by updating the system view -- a "virtual table" as Tom put it. Here's an example: Joe Conway
2002-08-15 04:51:27 +02:00
}
> 2. This patch includes the same Table Function API fixes that I > submitted on July 9: > > http://archives.postgresql.org/pgsql-patches/2002-07/msg00056.php > > Please disregard that one *if* this one is applied. If this one is > rejected please go ahead with the July 9th patch. The July 9th Table Function API patch mentioned above is now in CVS, so here is an updated version of the guc patch which should apply cleanly against CVS tip. Joe Conway
2002-07-20 07:49:28 +02:00
static char *
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
_ShowOption(struct config_generic * record, bool use_units)
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
{
char buffer[256];
const char *val;
reset all: command line and .conf options change defaults on RESET ALL those are restored. show all: GUC + non-GUC. SHOW ALL, RESET ALL Marko Kreen
2001-06-07 06:50:57 +02:00
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
switch (record->vartype)
{
case PGC_BOOL:
{
struct config_bool *conf = (struct config_bool *) record;
reset all: command line and .conf options change defaults on RESET ALL those are restored. show all: GUC + non-GUC. SHOW ALL, RESET ALL Marko Kreen
2001-06-07 06:50:57 +02:00
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
if (conf->show_hook)
val = (*conf->show_hook) ();
else
val = *conf->variable ? "on" : "off";
}
break;
case PGC_INT:
{
struct config_int *conf = (struct config_int *) record;
if (conf->show_hook)
val = (*conf->show_hook) ();
else
{
Prevent integer overflows during units conversion when displaying a GUC variable that has units. Per report from Stefan Kaltenbrunner. Backport to 8.2. I also backported my patch of 2007-06-21 that prevented comparable overflows on the input side, since that now seems to have enough field track record to be back-patched safely. That patch included addition of hints listing the available unit names, which I did not bother to strip out of it --- this will make a little more work for the translators, but they can copy the translation from 8.3, and anyway an untranslated hint is better than no hint.
2008-07-06 21:48:45 +02:00
/*
* Use int64 arithmetic to avoid overflows in units
Remove all the special-case code for INT64_IS_BUSTED, per decision that we're not going to support that anymore. I did keep the 64-bit-CRC-with-32-bit-arithmetic code, since it has a performance excuse to live. It's a bit moot since that's all ifdef'd out, of course.
2010-01-07 05:53:35 +01:00
* conversion.
Prevent integer overflows during units conversion when displaying a GUC variable that has units. Per report from Stefan Kaltenbrunner. Backport to 8.2. I also backported my patch of 2007-06-21 that prevented comparable overflows on the input side, since that now seems to have enough field track record to be back-patched safely. That patch included addition of hints listing the available unit names, which I did not bother to strip out of it --- this will make a little more work for the translators, but they can copy the translation from 8.3, and anyway an untranslated hint is better than no hint.
2008-07-06 21:48:45 +02:00
*/
int64 result = *conf->variable;
const char *unit;
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
Prevent integer overflows during units conversion when displaying a GUC variable that has units. Per report from Stefan Kaltenbrunner. Backport to 8.2. I also backported my patch of 2007-06-21 that prevented comparable overflows on the input side, since that now seems to have enough field track record to be back-patched safely. That patch included addition of hints listing the available unit names, which I did not bother to strip out of it --- this will make a little more work for the translators, but they can copy the translation from 8.3, and anyway an untranslated hint is better than no hint.
2008-07-06 21:48:45 +02:00
if (use_units && result > 0 &&
(record->flags & GUC_UNIT_MEMORY))
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
{
The attached patch changes units of the some default values in postgresql.conf. - shared_buffers = 32000kB => 32MB - temp_buffers = 8000kB => 8MB - wal_buffers = 8 => 64kB The code of initdb was a bit modified to write MB-unit values. Values greater than 8000kB are rounded out to MB. GUC_UNIT_XBLOCKS is added for wal_buffers. It is like GUC_UNIT_BLOCKS, but uses XLOG_BLCKSZ instead of BLCKSZ. Also, I cleaned up the test of GUC_UNIT_* flags in preparation to add more unit flags in less bits. ITAGAKI Takahiro
2006-10-03 23:11:55 +02:00
switch (record->flags & GUC_UNIT_MEMORY)
{
case GUC_UNIT_BLOCKS:
pgindent run for 8.2.
2006-10-04 02:30:14 +02:00
result *= BLCKSZ / 1024;
The attached patch changes units of the some default values in postgresql.conf. - shared_buffers = 32000kB => 32MB - temp_buffers = 8000kB => 8MB - wal_buffers = 8 => 64kB The code of initdb was a bit modified to write MB-unit values. Values greater than 8000kB are rounded out to MB. GUC_UNIT_XBLOCKS is added for wal_buffers. It is like GUC_UNIT_BLOCKS, but uses XLOG_BLCKSZ instead of BLCKSZ. Also, I cleaned up the test of GUC_UNIT_* flags in preparation to add more unit flags in less bits. ITAGAKI Takahiro
2006-10-03 23:11:55 +02:00
break;
case GUC_UNIT_XBLOCKS:
pgindent run for 8.2.
2006-10-04 02:30:14 +02:00
result *= XLOG_BLCKSZ / 1024;
The attached patch changes units of the some default values in postgresql.conf. - shared_buffers = 32000kB => 32MB - temp_buffers = 8000kB => 8MB - wal_buffers = 8 => 64kB The code of initdb was a bit modified to write MB-unit values. Values greater than 8000kB are rounded out to MB. GUC_UNIT_XBLOCKS is added for wal_buffers. It is like GUC_UNIT_BLOCKS, but uses XLOG_BLCKSZ instead of BLCKSZ. Also, I cleaned up the test of GUC_UNIT_* flags in preparation to add more unit flags in less bits. ITAGAKI Takahiro
2006-10-03 23:11:55 +02:00
break;
}
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
if (result % KB_PER_GB == 0)
{
result /= KB_PER_GB;
Prevent integer overflows during units conversion when displaying a GUC variable that has units. Per report from Stefan Kaltenbrunner. Backport to 8.2. I also backported my patch of 2007-06-21 that prevented comparable overflows on the input side, since that now seems to have enough field track record to be back-patched safely. That patch included addition of hints listing the available unit names, which I did not bother to strip out of it --- this will make a little more work for the translators, but they can copy the translation from 8.3, and anyway an untranslated hint is better than no hint.
2008-07-06 21:48:45 +02:00
unit = "GB";
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
}
else if (result % KB_PER_MB == 0)
{
result /= KB_PER_MB;
Prevent integer overflows during units conversion when displaying a GUC variable that has units. Per report from Stefan Kaltenbrunner. Backport to 8.2. I also backported my patch of 2007-06-21 that prevented comparable overflows on the input side, since that now seems to have enough field track record to be back-patched safely. That patch included addition of hints listing the available unit names, which I did not bother to strip out of it --- this will make a little more work for the translators, but they can copy the translation from 8.3, and anyway an untranslated hint is better than no hint.
2008-07-06 21:48:45 +02:00
unit = "MB";
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
}
else
{
Prevent integer overflows during units conversion when displaying a GUC variable that has units. Per report from Stefan Kaltenbrunner. Backport to 8.2. I also backported my patch of 2007-06-21 that prevented comparable overflows on the input side, since that now seems to have enough field track record to be back-patched safely. That patch included addition of hints listing the available unit names, which I did not bother to strip out of it --- this will make a little more work for the translators, but they can copy the translation from 8.3, and anyway an untranslated hint is better than no hint.
2008-07-06 21:48:45 +02:00
unit = "kB";
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
}
}
Prevent integer overflows during units conversion when displaying a GUC variable that has units. Per report from Stefan Kaltenbrunner. Backport to 8.2. I also backported my patch of 2007-06-21 that prevented comparable overflows on the input side, since that now seems to have enough field track record to be back-patched safely. That patch included addition of hints listing the available unit names, which I did not bother to strip out of it --- this will make a little more work for the translators, but they can copy the translation from 8.3, and anyway an untranslated hint is better than no hint.
2008-07-06 21:48:45 +02:00
else if (use_units && result > 0 &&
(record->flags & GUC_UNIT_TIME))
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
{
The attached patch changes units of the some default values in postgresql.conf. - shared_buffers = 32000kB => 32MB - temp_buffers = 8000kB => 8MB - wal_buffers = 8 => 64kB The code of initdb was a bit modified to write MB-unit values. Values greater than 8000kB are rounded out to MB. GUC_UNIT_XBLOCKS is added for wal_buffers. It is like GUC_UNIT_BLOCKS, but uses XLOG_BLCKSZ instead of BLCKSZ. Also, I cleaned up the test of GUC_UNIT_* flags in preparation to add more unit flags in less bits. ITAGAKI Takahiro
2006-10-03 23:11:55 +02:00
switch (record->flags & GUC_UNIT_TIME)
{
case GUC_UNIT_S:
result *= MS_PER_S;
break;
case GUC_UNIT_MIN:
result *= MS_PER_MIN;
break;
}
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
if (result % MS_PER_D == 0)
{
result /= MS_PER_D;
Prevent integer overflows during units conversion when displaying a GUC variable that has units. Per report from Stefan Kaltenbrunner. Backport to 8.2. I also backported my patch of 2007-06-21 that prevented comparable overflows on the input side, since that now seems to have enough field track record to be back-patched safely. That patch included addition of hints listing the available unit names, which I did not bother to strip out of it --- this will make a little more work for the translators, but they can copy the translation from 8.3, and anyway an untranslated hint is better than no hint.
2008-07-06 21:48:45 +02:00
unit = "d";
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
}
else if (result % MS_PER_H == 0)
{
result /= MS_PER_H;
Prevent integer overflows during units conversion when displaying a GUC variable that has units. Per report from Stefan Kaltenbrunner. Backport to 8.2. I also backported my patch of 2007-06-21 that prevented comparable overflows on the input side, since that now seems to have enough field track record to be back-patched safely. That patch included addition of hints listing the available unit names, which I did not bother to strip out of it --- this will make a little more work for the translators, but they can copy the translation from 8.3, and anyway an untranslated hint is better than no hint.
2008-07-06 21:48:45 +02:00
unit = "h";
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
}
else if (result % MS_PER_MIN == 0)
{
result /= MS_PER_MIN;
Prevent integer overflows during units conversion when displaying a GUC variable that has units. Per report from Stefan Kaltenbrunner. Backport to 8.2. I also backported my patch of 2007-06-21 that prevented comparable overflows on the input side, since that now seems to have enough field track record to be back-patched safely. That patch included addition of hints listing the available unit names, which I did not bother to strip out of it --- this will make a little more work for the translators, but they can copy the translation from 8.3, and anyway an untranslated hint is better than no hint.
2008-07-06 21:48:45 +02:00
unit = "min";
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
}
else if (result % MS_PER_S == 0)
{
result /= MS_PER_S;
Prevent integer overflows during units conversion when displaying a GUC variable that has units. Per report from Stefan Kaltenbrunner. Backport to 8.2. I also backported my patch of 2007-06-21 that prevented comparable overflows on the input side, since that now seems to have enough field track record to be back-patched safely. That patch included addition of hints listing the available unit names, which I did not bother to strip out of it --- this will make a little more work for the translators, but they can copy the translation from 8.3, and anyway an untranslated hint is better than no hint.
2008-07-06 21:48:45 +02:00
unit = "s";
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
}
else
{
Prevent integer overflows during units conversion when displaying a GUC variable that has units. Per report from Stefan Kaltenbrunner. Backport to 8.2. I also backported my patch of 2007-06-21 that prevented comparable overflows on the input side, since that now seems to have enough field track record to be back-patched safely. That patch included addition of hints listing the available unit names, which I did not bother to strip out of it --- this will make a little more work for the translators, but they can copy the translation from 8.3, and anyway an untranslated hint is better than no hint.
2008-07-06 21:48:45 +02:00
unit = "ms";
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
}
}
else
Prevent integer overflows during units conversion when displaying a GUC variable that has units. Per report from Stefan Kaltenbrunner. Backport to 8.2. I also backported my patch of 2007-06-21 that prevented comparable overflows on the input side, since that now seems to have enough field track record to be back-patched safely. That patch included addition of hints listing the available unit names, which I did not bother to strip out of it --- this will make a little more work for the translators, but they can copy the translation from 8.3, and anyway an untranslated hint is better than no hint.
2008-07-06 21:48:45 +02:00
unit = "";
Allow units to be specified with configuration settings.
2006-07-27 10:30:41 +02:00
Prevent integer overflows during units conversion when displaying a GUC variable that has units. Per report from Stefan Kaltenbrunner. Backport to 8.2. I also backported my patch of 2007-06-21 that prevented comparable overflows on the input side, since that now seems to have enough field track record to be back-patched safely. That patch included addition of hints listing the available unit names, which I did not bother to strip out of it --- this will make a little more work for the translators, but they can copy the translation from 8.3, and anyway an untranslated hint is better than no hint.
2008-07-06 21:48:45 +02:00
snprintf(buffer, sizeof(buffer), INT64_FORMAT "%s",
result, unit);
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
val = buffer;
}
}
break;
case PGC_REAL:
{
struct config_real *conf = (struct config_real *) record;
if (conf->show_hook)
val = (*conf->show_hook) ();
else
{
snprintf(buffer, sizeof(buffer), "%g",
*conf->variable);
val = buffer;
}
}
break;
case PGC_STRING:
{
struct config_string *conf = (struct config_string *) record;
reset all: command line and .conf options change defaults on RESET ALL those are restored. show all: GUC + non-GUC. SHOW ALL, RESET ALL Marko Kreen
2001-06-07 06:50:57 +02:00
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
if (conf->show_hook)
val = (*conf->show_hook) ();
else if (*conf->variable && **conf->variable)
val = *conf->variable;
else
When a GUC string variable is not set, print the empty string (in SHOW etc.), not "unset". An "unset" state doesn't really exist; all variables behave like an empty string value if the string being pointed to has not been initialized.
2006-07-26 13:39:47 +02:00
val = "";
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
}
break;
reset all: command line and .conf options change defaults on RESET ALL those are restored. show all: GUC + non-GUC. SHOW ALL, RESET ALL Marko Kreen
2001-06-07 06:50:57 +02:00
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
case PGC_ENUM:
{
struct config_enum *conf = (struct config_enum *) record;
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
if (conf->show_hook)
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
val = (*conf->show_hook) ();
else
Fix postgres --describe-config for guc enums, breakage noted by Alvaro. While at it, rename option lookup functions to make names clearer, per discussion with Tom.
2008-03-17 18:45:09 +01:00
val = config_enum_lookup_by_value(conf, *conf->variable);
Implement enum type for guc parameters, and convert a couple of existing variables to it. More need to be converted, but I wanted to get this in before it conflicts with too much... Other than just centralising the text-to-int conversion for parameters, this allows the pg_settings view to contain a list of available options and allows an error hint to show what values are allowed.
2008-03-10 13:55:13 +01:00
}
break;
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
default:
/* just to keep compiler quiet */
val = "???";
break;
}
reset all: command line and .conf options change defaults on RESET ALL those are restored. show all: GUC + non-GUC. SHOW ALL, RESET ALL Marko Kreen
2001-06-07 06:50:57 +02:00
> 2. This patch includes the same Table Function API fixes that I > submitted on July 9: > > http://archives.postgresql.org/pgsql-patches/2002-07/msg00056.php > > Please disregard that one *if* this one is applied. If this one is > rejected please go ahead with the July 9th patch. The July 9th Table Function API patch mentioned above is now in CVS, so here is an updated version of the guc patch which should apply cleanly against CVS tip. Joe Conway
2002-07-20 07:49:28 +02:00
return pstrdup(val);
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
}
Move PGPORT envar handling to ResetAllOptions(). Improve long options parsing to not clobber the optarg string -- so that we can bring SetOptsFile() up to speed.
2000-07-03 22:46:10 +02:00
SECOND ATTEMPT Dump/read non-default GUC values for use by exec'ed backend, for Win32.
2003-05-03 00:02:47 +02:00
#ifdef EXEC_BACKEND
Arrange for the postmaster (and standalone backends, initdb, etc) to chdir into PGDATA and subsequently use relative paths instead of absolute paths to access all files under PGDATA. This seems to give a small performance improvement, and it should make the system more robust against naive DBAs doing things like moving a database directory that has a live postmaster in it. Per recent discussion.
2005-07-04 06:51:52 +02:00
SECOND ATTEMPT Dump/read non-default GUC values for use by exec'ed backend, for Win32.
2003-05-03 00:02:47 +02:00
/*
write_nondefault_variables must take care to write custom_variable_classes first; otherwise backends reading the file might reject values of custom variables. Per experimentation with auto_explain.
2009-01-02 03:02:10 +01:00
* These routines dump out all non-default GUC options into a binary
SECOND ATTEMPT Dump/read non-default GUC values for use by exec'ed backend, for Win32.
2003-05-03 00:02:47 +02:00
* file that is read by all exec'ed backends. The format is:
*
* variable name, string, null terminated
* variable value, string, null terminated
* variable source, integer
Remember the source GucContext for each GUC parameter. We used to just remember the GucSource, but saving GucContext too provides a little more information --- notably, whether a SET was done by a superuser or regular user. This allows us to rip out the fairly dodgy code that define_custom_variable used to use to try to infer the context to re-install a pre-existing setting with. In particular, it now works for a superuser to SET a extension's SUSET custom variable before loading the associated extension, because GUC can remember whether the SET was done as a superuser or not. The plperl regression tests contain an example where this is useful.
2011-10-04 22:13:16 +02:00
* variable scontext, integer
SECOND ATTEMPT Dump/read non-default GUC values for use by exec'ed backend, for Win32.
2003-05-03 00:02:47 +02:00
*/
write_nondefault_variables must take care to write custom_variable_classes first; otherwise backends reading the file might reject values of custom variables. Per experimentation with auto_explain.
2009-01-02 03:02:10 +01:00
static void
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
write_one_nondefault_variable(FILE *fp, struct config_generic * gconf)
write_nondefault_variables must take care to write custom_variable_classes first; otherwise backends reading the file might reject values of custom variables. Per experimentation with auto_explain.
2009-01-02 03:02:10 +01:00
{
if (gconf->source == PGC_S_DEFAULT)
return;
fprintf(fp, "%s", gconf->name);
fputc(0, fp);
switch (gconf->vartype)
{
case PGC_BOOL:
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
{
struct config_bool *conf = (struct config_bool *) gconf;
write_nondefault_variables must take care to write custom_variable_classes first; otherwise backends reading the file might reject values of custom variables. Per experimentation with auto_explain.
2009-01-02 03:02:10 +01:00
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
if (*conf->variable)
fprintf(fp, "true");
else
fprintf(fp, "false");
}
break;
write_nondefault_variables must take care to write custom_variable_classes first; otherwise backends reading the file might reject values of custom variables. Per experimentation with auto_explain.
2009-01-02 03:02:10 +01:00
case PGC_INT:
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
{
struct config_int *conf = (struct config_int *) gconf;
write_nondefault_variables must take care to write custom_variable_classes first; otherwise backends reading the file might reject values of custom variables. Per experimentation with auto_explain.
2009-01-02 03:02:10 +01:00
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
fprintf(fp, "%d", *conf->variable);
}
break;
write_nondefault_variables must take care to write custom_variable_classes first; otherwise backends reading the file might reject values of custom variables. Per experimentation with auto_explain.
2009-01-02 03:02:10 +01:00
case PGC_REAL:
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
{
struct config_real *conf = (struct config_real *) gconf;
write_nondefault_variables must take care to write custom_variable_classes first; otherwise backends reading the file might reject values of custom variables. Per experimentation with auto_explain.
2009-01-02 03:02:10 +01:00
Remember the source GucContext for each GUC parameter. We used to just remember the GucSource, but saving GucContext too provides a little more information --- notably, whether a SET was done by a superuser or regular user. This allows us to rip out the fairly dodgy code that define_custom_variable used to use to try to infer the context to re-install a pre-existing setting with. In particular, it now works for a superuser to SET a extension's SUSET custom variable before loading the associated extension, because GUC can remember whether the SET was done as a superuser or not. The plperl regression tests contain an example where this is useful.
2011-10-04 22:13:16 +02:00
fprintf(fp, "%.17g", *conf->variable);
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
}
break;
write_nondefault_variables must take care to write custom_variable_classes first; otherwise backends reading the file might reject values of custom variables. Per experimentation with auto_explain.
2009-01-02 03:02:10 +01:00
case PGC_STRING:
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
{
struct config_string *conf = (struct config_string *) gconf;
write_nondefault_variables must take care to write custom_variable_classes first; otherwise backends reading the file might reject values of custom variables. Per experimentation with auto_explain.
2009-01-02 03:02:10 +01:00
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
fprintf(fp, "%s", *conf->variable);
}
break;
write_nondefault_variables must take care to write custom_variable_classes first; otherwise backends reading the file might reject values of custom variables. Per experimentation with auto_explain.
2009-01-02 03:02:10 +01:00
case PGC_ENUM:
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
{
struct config_enum *conf = (struct config_enum *) gconf;
fprintf(fp, "%s",
config_enum_lookup_by_value(conf, *conf->variable));
}
break;
write_nondefault_variables must take care to write custom_variable_classes first; otherwise backends reading the file might reject values of custom variables. Per experimentation with auto_explain.
2009-01-02 03:02:10 +01:00
}
fputc(0, fp);
Remember the source GucContext for each GUC parameter. We used to just remember the GucSource, but saving GucContext too provides a little more information --- notably, whether a SET was done by a superuser or regular user. This allows us to rip out the fairly dodgy code that define_custom_variable used to use to try to infer the context to re-install a pre-existing setting with. In particular, it now works for a superuser to SET a extension's SUSET custom variable before loading the associated extension, because GUC can remember whether the SET was done as a superuser or not. The plperl regression tests contain an example where this is useful.
2011-10-04 22:13:16 +02:00
fwrite(&gconf->source, 1, sizeof(gconf->source), fp);
fwrite(&gconf->scontext, 1, sizeof(gconf->scontext), fp);
write_nondefault_variables must take care to write custom_variable_classes first; otherwise backends reading the file might reject values of custom variables. Per experimentation with auto_explain.
2009-01-02 03:02:10 +01:00
}
SECOND ATTEMPT Dump/read non-default GUC values for use by exec'ed backend, for Win32.
2003-05-03 00:02:47 +02:00
void
write_nondefault_variables(GucContext context)
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
int elevel;
FILE *fp;
write_nondefault_variables must take care to write custom_variable_classes first; otherwise backends reading the file might reject values of custom variables. Per experimentation with auto_explain.
2009-01-02 03:02:10 +01:00
int i;
SECOND ATTEMPT Dump/read non-default GUC values for use by exec'ed backend, for Win32.
2003-05-03 00:02:47 +02:00
Assert(context == PGC_POSTMASTER || context == PGC_SIGHUP);
Code review for recent changes in guc-file.l. Avoid multiple frees, use of already-freed strings, other silliness. Also fix reporting of config file syntax errors so that it actually works reasonably well (eg, points at the correct line). Use palloc instead of malloc for temporary storage to reduce code clutter.
2004-09-01 00:43:58 +02:00
elevel = (context == PGC_SIGHUP) ? LOG : ERROR;
SECOND ATTEMPT Dump/read non-default GUC values for use by exec'ed backend, for Win32.
2003-05-03 00:02:47 +02:00
/*
* Open file
*/
Arrange for the postmaster (and standalone backends, initdb, etc) to chdir into PGDATA and subsequently use relative paths instead of absolute paths to access all files under PGDATA. This seems to give a small performance improvement, and it should make the system more robust against naive DBAs doing things like moving a database directory that has a live postmaster in it. Per recent discussion.
2005-07-04 06:51:52 +02:00
fp = AllocateFile(CONFIG_EXEC_PARAMS_NEW, "w");
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
if (!fp)
{
Error message editing in backend/utils (except /adt).
2003-07-25 22:18:01 +02:00
ereport(elevel,
(errcode_for_file_access(),
Arrange for the postmaster (and standalone backends, initdb, etc) to chdir into PGDATA and subsequently use relative paths instead of absolute paths to access all files under PGDATA. This seems to give a small performance improvement, and it should make the system more robust against naive DBAs doing things like moving a database directory that has a live postmaster in it. Per recent discussion.
2005-07-04 06:51:52 +02:00
errmsg("could not write to file \"%s\": %m",
CONFIG_EXEC_PARAMS_NEW)));
SECOND ATTEMPT Dump/read non-default GUC values for use by exec'ed backend, for Win32.
2003-05-03 00:02:47 +02:00
return;
}
for (i = 0; i < num_guc_variables; i++)
{
Remove the custom_variable_classes parameter. This variable provides only marginal error-prevention capability (since it can only check the prefix of a qualified GUC name), and the consensus is that that isn't worth the amount of hassle that maintaining the setting creates for DBAs. So, let's just remove it. With this commit, the system will silently accept a value for any qualified GUC name at all, whether it has anything to do with any known extension or not. (Unqualified names still have to match known built-in settings, though; and you will get a WARNING at extension load time if there's an unrecognized setting with that extension's prefix.) There's still some discussion ongoing about whether to tighten that up and if so how; but if we do come up with a solution, it's not likely to look anything like custom_variable_classes.
2011-10-04 18:36:18 +02:00
write_one_nondefault_variable(fp, guc_variables[i]);
SECOND ATTEMPT Dump/read non-default GUC values for use by exec'ed backend, for Win32.
2003-05-03 00:02:47 +02:00
}
Ensure that close() and fclose() are checked for errors, at least in cases involving writes. Per recent discussion about the possibility of close-time failures on some filesystems. There is a TODO item for this, too.
2004-01-26 23:35:32 +01:00
if (FreeFile(fp))
{
ereport(elevel,
(errcode_for_file_access(),
Arrange for the postmaster (and standalone backends, initdb, etc) to chdir into PGDATA and subsequently use relative paths instead of absolute paths to access all files under PGDATA. This seems to give a small performance improvement, and it should make the system more robust against naive DBAs doing things like moving a database directory that has a live postmaster in it. Per recent discussion.
2005-07-04 06:51:52 +02:00
errmsg("could not write to file \"%s\": %m",
CONFIG_EXEC_PARAMS_NEW)));
Ensure that close() and fclose() are checked for errors, at least in cases involving writes. Per recent discussion about the possibility of close-time failures on some filesystems. There is a TODO item for this, too.
2004-01-26 23:35:32 +01:00
return;
}
Adjust rename on Win32 to only link to temp name while holding lock, then release locks and loop over renaming to active file name.
2004-02-02 01:17:23 +01:00
/*
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
* Put new file in place. This could delay on Win32, but we don't hold
* any exclusive locks.
Adjust rename on Win32 to only link to temp name while holding lock, then release locks and loop over renaming to active file name.
2004-02-02 01:17:23 +01:00
*/
Arrange for the postmaster (and standalone backends, initdb, etc) to chdir into PGDATA and subsequently use relative paths instead of absolute paths to access all files under PGDATA. This seems to give a small performance improvement, and it should make the system more robust against naive DBAs doing things like moving a database directory that has a live postmaster in it. Per recent discussion.
2005-07-04 06:51:52 +02:00
rename(CONFIG_EXEC_PARAMS_NEW, CONFIG_EXEC_PARAMS);
SECOND ATTEMPT Dump/read non-default GUC values for use by exec'ed backend, for Win32.
2003-05-03 00:02:47 +02:00
}
/*
* Read string, including null byte from file
*
* Return NULL on EOF and nothing read
*/
static char *
read_string_with_null(FILE *fp)
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
int i = 0,
ch,
maxlen = 256;
char *str = NULL;
SECOND ATTEMPT Dump/read non-default GUC values for use by exec'ed backend, for Win32.
2003-05-03 00:02:47 +02:00
do
{
if ((ch = fgetc(fp)) == EOF)
{
if (i == 0)
return NULL;
else
Error message editing in backend/utils (except /adt).
2003-07-25 22:18:01 +02:00
elog(FATAL, "invalid format of exec config params file");
SECOND ATTEMPT Dump/read non-default GUC values for use by exec'ed backend, for Win32.
2003-05-03 00:02:47 +02:00
}
if (i == 0)
Fix unchecked mallocs/strdups added by recent placeholder-config-vars patch. Thomas Hallgren
2004-07-06 01:14:14 +02:00
str = guc_malloc(FATAL, maxlen);
SECOND ATTEMPT Dump/read non-default GUC values for use by exec'ed backend, for Win32.
2003-05-03 00:02:47 +02:00
else if (i == maxlen)
Fix unchecked mallocs/strdups added by recent placeholder-config-vars patch. Thomas Hallgren
2004-07-06 01:14:14 +02:00
str = guc_realloc(FATAL, str, maxlen *= 2);
SECOND ATTEMPT Dump/read non-default GUC values for use by exec'ed backend, for Win32.
2003-05-03 00:02:47 +02:00
str[i++] = ch;
} while (ch != 0);
return str;
}
/*
* This routine loads a previous postmaster dump of its non-default
* settings.
*/
void
read_nondefault_variables(void)
{
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
FILE *fp;
char *varname,
*varvalue;
Remember the source GucContext for each GUC parameter. We used to just remember the GucSource, but saving GucContext too provides a little more information --- notably, whether a SET was done by a superuser or regular user. This allows us to rip out the fairly dodgy code that define_custom_variable used to use to try to infer the context to re-install a pre-existing setting with. In particular, it now works for a superuser to SET a extension's SUSET custom variable before loading the associated extension, because GUC can remember whether the SET was done as a superuser or not. The plperl regression tests contain an example where this is useful.
2011-10-04 22:13:16 +02:00
GucSource varsource;
GucContext varscontext;
SECOND ATTEMPT Dump/read non-default GUC values for use by exec'ed backend, for Win32.
2003-05-03 00:02:47 +02:00
/*
* Open file
*/
Arrange for the postmaster (and standalone backends, initdb, etc) to chdir into PGDATA and subsequently use relative paths instead of absolute paths to access all files under PGDATA. This seems to give a small performance improvement, and it should make the system more robust against naive DBAs doing things like moving a database directory that has a live postmaster in it. Per recent discussion.
2005-07-04 06:51:52 +02:00
fp = AllocateFile(CONFIG_EXEC_PARAMS, "r");
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
if (!fp)
{
/* File not found is fine */
if (errno != ENOENT)
Error message editing in backend/utils (except /adt).
2003-07-25 22:18:01 +02:00
ereport(FATAL,
(errcode_for_file_access(),
Arrange for the postmaster (and standalone backends, initdb, etc) to chdir into PGDATA and subsequently use relative paths instead of absolute paths to access all files under PGDATA. This seems to give a small performance improvement, and it should make the system more robust against naive DBAs doing things like moving a database directory that has a live postmaster in it. Per recent discussion.
2005-07-04 06:51:52 +02:00
errmsg("could not read from file \"%s\": %m",
CONFIG_EXEC_PARAMS)));
SECOND ATTEMPT Dump/read non-default GUC values for use by exec'ed backend, for Win32.
2003-05-03 00:02:47 +02:00
return;
Add --help-config facility to dump information about GUC parameters without needing a running backend. Reorder postgresql.conf.sample to match new layout of runtime.sgml. This commit re-adds work lost in Wednesday's crash.
2003-07-04 18:41:22 +02:00
}
SECOND ATTEMPT Dump/read non-default GUC values for use by exec'ed backend, for Win32.
2003-05-03 00:02:47 +02:00
Error message editing in backend/utils (except /adt).
2003-07-25 22:18:01 +02:00
for (;;)
SECOND ATTEMPT Dump/read non-default GUC values for use by exec'ed backend, for Win32.
2003-05-03 00:02:47 +02:00
{
>> It certainly doesn't. There still was a bug with the locale stuff, >> though - the GUC variable was not set in the child >processes. So "show >> lc_collate" would *always* return "C", for example. attached >patch fixes >> this. > >Hm. Why were these vars not propagated by the regular >mechanism for GUC >variables (write_nondefault_variables or whatever it's called)? If the >problem is that it's not accepting PGC_INTERNAL values, then we need to >fix it there not here, because otherwise we'll have to pass all the >PGC_INTERNAL variables through the backend_variables file, which seems >like a recipe for more of the same sort of bug. Good point :-( I think the problem is not only that it specifically does not deal with PGC_INTERNAL variables. The problem is in the fact that write_nondefault_variables is called *before* the locale is read (because the locale is read from pg_control and not from any of the "usual" ways to read it). Attached patch is another stab at fixing it. It makes postmaster dump a new copy of the file once it has started the database (before it accepts any connections), which is when it will know about these parameters. Also updates the reading code to set the context to the one where the variable was originally set (PGC_POSTMASTER won't work for PGC_INTERNAL, and the other way around). We still pass lc_collate through the special file, because set_config_option on lc_collate will speficially *not* call setlocale(), and we need that call. But we no longer call set_config_option from there. Magnus Hagander
2004-06-11 05:54:54 +02:00
struct config_generic *record;
SECOND ATTEMPT Dump/read non-default GUC values for use by exec'ed backend, for Win32.
2003-05-03 00:02:47 +02:00
if ((varname = read_string_with_null(fp)) == NULL)
break;
Code review for GUC revert-values-if-removed-from-postgresql.conf patch; and in passing, fix some bogosities dating from the custom_variable_classes patch. Fix guc-file.l to correctly check changes in custom_variable_classes that are attempted concurrently with additions/removals of custom variables, and don't allow the new setting to be applied in advance of checking it. Clean up messy and undocumented situation for string variables with NULL boot_val. Fix DefineCustomVariable functions to initialize boot_val correctly. Prevent find_option from inserting bogus placeholders for custom variables that are simply inquired about rather than being set.
2007-09-10 02:57:22 +02:00
if ((record = find_option(varname, true, FATAL)) == NULL)
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
elog(FATAL, "failed to locate variable %s in exec config params file", varname);
SECOND ATTEMPT Dump/read non-default GUC values for use by exec'ed backend, for Win32.
2003-05-03 00:02:47 +02:00
if ((varvalue = read_string_with_null(fp)) == NULL)
Error message editing in backend/utils (except /adt).
2003-07-25 22:18:01 +02:00
elog(FATAL, "invalid format of exec config params file");
Remember the source GucContext for each GUC parameter. We used to just remember the GucSource, but saving GucContext too provides a little more information --- notably, whether a SET was done by a superuser or regular user. This allows us to rip out the fairly dodgy code that define_custom_variable used to use to try to infer the context to re-install a pre-existing setting with. In particular, it now works for a superuser to SET a extension's SUSET custom variable before loading the associated extension, because GUC can remember whether the SET was done as a superuser or not. The plperl regression tests contain an example where this is useful.
2011-10-04 22:13:16 +02:00
if (fread(&varsource, 1, sizeof(varsource), fp) != sizeof(varsource))
elog(FATAL, "invalid format of exec config params file");
if (fread(&varscontext, 1, sizeof(varscontext), fp) != sizeof(varscontext))
Error message editing in backend/utils (except /adt).
2003-07-25 22:18:01 +02:00
elog(FATAL, "invalid format of exec config params file");
SECOND ATTEMPT Dump/read non-default GUC values for use by exec'ed backend, for Win32.
2003-05-03 00:02:47 +02:00
Remember the source GucContext for each GUC parameter. We used to just remember the GucSource, but saving GucContext too provides a little more information --- notably, whether a SET was done by a superuser or regular user. This allows us to rip out the fairly dodgy code that define_custom_variable used to use to try to infer the context to re-install a pre-existing setting with. In particular, it now works for a superuser to SET a extension's SUSET custom variable before loading the associated extension, because GUC can remember whether the SET was done as a superuser or not. The plperl regression tests contain an example where this is useful.
2011-10-04 22:13:16 +02:00
(void) set_config_option(varname, varvalue,
varscontext, varsource,
GUC_ACTION_SET, true);
SECOND ATTEMPT Dump/read non-default GUC values for use by exec'ed backend, for Win32.
2003-05-03 00:02:47 +02:00
free(varname);
free(varvalue);
}
FreeFile(fp);
}
Standard pgindent run for 8.1.
2005-10-15 04:49:52 +02:00
#endif /* EXEC_BACKEND */
SECOND ATTEMPT Dump/read non-default GUC values for use by exec'ed backend, for Win32.
2003-05-03 00:02:47 +02:00
Move PGPORT envar handling to ResetAllOptions(). Improve long options parsing to not clobber the optarg string -- so that we can bring SetOptsFile() up to speed.
2000-07-03 22:46:10 +02:00
/*
* A little "long argument" simulation, although not quite GNU
* compliant. Takes a string of the form "some-option=some value" and
* returns name = "some_option" and value = "some value" in malloc'ed
* storage. Note that '-' is converted to '_' in the option name. If
* there is no '=' in the input string then value will be NULL.
*/
void
pgindent run. Make it all clean.
2001-03-22 05:01:46 +01:00
ParseLongOption(const char *string, char **name, char **value)
Move PGPORT envar handling to ResetAllOptions(). Improve long options parsing to not clobber the optarg string -- so that we can bring SetOptsFile() up to speed.
2000-07-03 22:46:10 +02:00
{
pgindent run. Make it all clean.
2001-03-22 05:01:46 +01:00
size_t equal_pos;
char *cp;
Move PGPORT envar handling to ResetAllOptions(). Improve long options parsing to not clobber the optarg string -- so that we can bring SetOptsFile() up to speed.
2000-07-03 22:46:10 +02:00
AssertArg(string);
AssertArg(name);
AssertArg(value);
equal_pos = strcspn(string, "=");
if (string[equal_pos] == '=')
{
Fix unchecked mallocs/strdups added by recent placeholder-config-vars patch. Thomas Hallgren
2004-07-06 01:14:14 +02:00
*name = guc_malloc(FATAL, equal_pos + 1);
Replace some strncpy() by strlcpy().
2007-02-07 01:52:35 +01:00
strlcpy(*name, string, equal_pos + 1);
Move PGPORT envar handling to ResetAllOptions(). Improve long options parsing to not clobber the optarg string -- so that we can bring SetOptsFile() up to speed.
2000-07-03 22:46:10 +02:00
Fix unchecked mallocs/strdups added by recent placeholder-config-vars patch. Thomas Hallgren
2004-07-06 01:14:14 +02:00
*value = guc_strdup(FATAL, &string[equal_pos + 1]);
Move PGPORT envar handling to ResetAllOptions(). Improve long options parsing to not clobber the optarg string -- so that we can bring SetOptsFile() up to speed.
2000-07-03 22:46:10 +02:00
}
pgindent run. Make it all clean.
2001-03-22 05:01:46 +01:00
else
Move PGPORT envar handling to ResetAllOptions(). Improve long options parsing to not clobber the optarg string -- so that we can bring SetOptsFile() up to speed.
2000-07-03 22:46:10 +02:00
{
Extend GUC concepts of parse_hook and assign_hook to all four supported datatypes, not only strings. parse_hook is useless for bool, I suppose, but it seems possibly useful for int and double to apply variable-specific constraints that are more complex than simple range limits. assign_hook is definitely useful for all datatypes --- we need it right now for bool to support date cache reset when changing Australian timezone rule setting. Also, clean up some residual problems with the reset all/show all patch, including memory leaks and mistaken reset of PostPortNumber. It seems best that RESET ALL not touch variables that don't have SUSET or USERSET context.
2001-06-13 00:54:06 +02:00
/* no equal sign in string */
Fix unchecked mallocs/strdups added by recent placeholder-config-vars patch. Thomas Hallgren
2004-07-06 01:14:14 +02:00
*name = guc_strdup(FATAL, string);
Move PGPORT envar handling to ResetAllOptions(). Improve long options parsing to not clobber the optarg string -- so that we can bring SetOptsFile() up to speed.
2000-07-03 22:46:10 +02:00
*value = NULL;
}
pgindent run. Make it all clean.
2001-03-22 05:01:46 +01:00
for (cp = *name; *cp; cp++)
Move PGPORT envar handling to ResetAllOptions(). Improve long options parsing to not clobber the optarg string -- so that we can bring SetOptsFile() up to speed.
2000-07-03 22:46:10 +02:00
if (*cp == '-')
*cp = '_';
}
Small polishing of syslog facility and ident settings. Only allow setting at postmaster start, rename syslog_progid to syslog_ident, since syslog itself uses that term, fix doc markup.
2000-11-14 20:13:27 +01:00
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
/*
Prevent ALTER USER f RESET ALL from removing the settings that were put there by a superuser -- "ALTER USER f RESET setting" already disallows removing such a setting. Apply the same treatment to ALTER DATABASE d RESET ALL when run by a database owner that's not superuser.
2010-03-25 15:44:34 +01:00
* Handle options fetched from pg_db_role_setting.setconfig,
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
* pg_proc.proconfig, etc. Caller must specify proper context/source/action.
Implement function-local GUC parameter settings, as per recent discussion. There are still some loose ends: I didn't do anything about the SET FROM CURRENT idea yet, and it's not real clear whether we are happy with the interaction of SET LOCAL with function-local settings. The documentation is a bit spartan, too.
2007-09-03 02:39:26 +02:00
*
Repair core dump when trying to delete an entry from an already-NULL datconfig or useconfig field. Per report from Dustin Sallings.
2002-12-02 06:20:47 +01:00
* The array parameter must be an array of TEXT (it must not be NULL).
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
*/
User and database-specific session defaults for run-time configuration variables. New commands ALTER DATABASE ... SET and ALTER USER ... SET.
2002-03-01 23:45:19 +01:00
void
Implement function-local GUC parameter settings, as per recent discussion. There are still some loose ends: I didn't do anything about the SET FROM CURRENT idea yet, and it's not real clear whether we are happy with the interaction of SET LOCAL with function-local settings. The documentation is a bit spartan, too.
2007-09-03 02:39:26 +02:00
ProcessGUCArray(ArrayType *array,
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
GucContext context, GucSource source, GucAction action)
User and database-specific session defaults for run-time configuration variables. New commands ALTER DATABASE ... SET and ALTER USER ... SET.
2002-03-01 23:45:19 +01:00
{
pgindent run.
2002-09-04 22:31:48 +02:00
int i;
User and database-specific session defaults for run-time configuration variables. New commands ALTER DATABASE ... SET and ALTER USER ... SET.
2002-03-01 23:45:19 +01:00
Merge the last few variable.c configuration variables into the generic GUC support. It's now possible to set datestyle, timezone, and client_encoding from postgresql.conf and per-database or per-user settings. Also, implement rollback of SET commands that occur in a transaction that later fails. Create a SET LOCAL var = value syntax that sets the variable only for the duration of the current transaction. All per previous discussions in pghackers.
2002-05-17 03:19:19 +02:00
Assert(array != NULL);
Modify array operations to include array's element type OID in the array header, and to compute sizing and alignment of array elements the same way normal tuple access operations do --- viz, using the tupmacs.h macros att_addlength and att_align. This makes the world safe for arrays of cstrings or intervals, and should make it much easier to write array-type-polymorphic functions; as examples see the cleanups of array_out and contrib/array_iterator. By Joe Conway and Tom Lane.
2002-08-26 19:54:02 +02:00
Assert(ARR_ELEMTYPE(array) == TEXTOID);
Assert(ARR_NDIM(array) == 1);
Assert(ARR_LBOUND(array)[0] == 1);
User and database-specific session defaults for run-time configuration variables. New commands ALTER DATABASE ... SET and ALTER USER ... SET.
2002-03-01 23:45:19 +01:00
for (i = 1; i <= ARR_DIMS(array)[0]; i++)
{
Datum d;
bool isnull;
char *s;
char *name;
char *value;
d = array_ref(array, 1, &i,
pgindent run.
2002-09-04 22:31:48 +02:00
-1 /* varlenarray */ ,
-1 /* TEXT's typlen */ ,
false /* TEXT's typbyval */ ,
'i' /* TEXT's typalign */ ,
User and database-specific session defaults for run-time configuration variables. New commands ALTER DATABASE ... SET and ALTER USER ... SET.
2002-03-01 23:45:19 +01:00
&isnull);
if (isnull)
continue;
Simplify and standardize conversions between TEXT datums and ordinary C strings. This patch introduces four support functions cstring_to_text, cstring_to_text_with_len, text_to_cstring, and text_to_cstring_buffer, and two macros CStringGetTextDatum and TextDatumGetCString. A number of existing macros that provided variants on these themes were removed. Most of the places that need to make such conversions now require just one function or macro call, in place of the multiple notational layers that used to be needed. There are no longer any direct calls of textout or textin, and we got most of the places that were using handmade conversions via memcpy (there may be a few still lurking, though). This commit doesn't make any serious effort to eliminate transient memory leaks caused by detoasting toasted text objects before they reach text_to_cstring. We changed PG_GETARG_TEXT_P to PG_GETARG_TEXT_PP in a few places where it was easy, but much more could be done. Brendan Jurd and Tom Lane
2008-03-25 23:42:46 +01:00
s = TextDatumGetCString(d);
Dept. of second thoughts: interaction between DoIt and makeDepend in set_config_option wasn't quite right. Also clean up a couple other things that could have been done better.
2002-05-17 22:32:29 +02:00
User and database-specific session defaults for run-time configuration variables. New commands ALTER DATABASE ... SET and ALTER USER ... SET.
2002-03-01 23:45:19 +01:00
ParseLongOption(s, &name, &value);
if (!value)
{
Error message editing in backend/utils (except /adt).
2003-07-25 22:18:01 +02:00
ereport(WARNING,
(errcode(ERRCODE_SYNTAX_ERROR),
Implement function-local GUC parameter settings, as per recent discussion. There are still some loose ends: I didn't do anything about the SET FROM CURRENT idea yet, and it's not real clear whether we are happy with the interaction of SET LOCAL with function-local settings. The documentation is a bit spartan, too.
2007-09-03 02:39:26 +02:00
errmsg("could not parse setting for parameter \"%s\"",
name)));
Dept. of second thoughts: interaction between DoIt and makeDepend in set_config_option wasn't quite right. Also clean up a couple other things that could have been done better.
2002-05-17 22:32:29 +02:00
free(name);
Change made to elog: o Change all current CVS messages of NOTICE to WARNING. We were going to do this just before 7.3 beta but it has to be done now, as you will see below. o Change current INFO messages that should be controlled by client_min_messages to NOTICE. o Force remaining INFO messages, like from EXPLAIN, VACUUM VERBOSE, etc. to always go to the client. o Remove INFO from the client_min_messages options and add NOTICE. Seems we do need three non-ERROR elog levels to handle the various behaviors we need for these messages. Regression passed.
2002-03-06 07:10:59 +01:00
continue;
User and database-specific session defaults for run-time configuration variables. New commands ALTER DATABASE ... SET and ALTER USER ... SET.
2002-03-01 23:45:19 +01:00
}
Arrange for SET LOCAL's effects to persist until the end of the current top transaction, unless rolled back or overridden by a SET clause for the same variable attached to a surrounding function call. Per discussion, these seem the best semantics. Note that this is an INCOMPATIBLE CHANGE: in 8.0 through 8.2, SET LOCAL's effects disappeared at subtransaction commit (leading to behavior that made little sense at the SQL level). I took advantage of the opportunity to rewrite and simplify the GUC variable save/restore logic a little bit. The old idea of a "tentative" value is gone; it was a hangover from before we had a stack. Also, we no longer need a stack entry for every nesting level, but only for those in which a variable's value actually changed.
2007-09-11 02:06:42 +02:00
(void) set_config_option(name, value, context, source, action, true);
Dept. of second thoughts: interaction between DoIt and makeDepend in set_config_option wasn't quite right. Also clean up a couple other things that could have been done better.
2002-05-17 22:32:29 +02:00
free(name);
if (value)
free(value);
Prevent ALTER USER f RESET ALL from removing the settings that were put there by a superuser -- "ALTER USER f RESET setting" already disallows removing such a setting. Apply the same treatment to ALTER DATABASE d RESET ALL when run by a database owner that's not superuser.
2010-03-25 15:44:34 +01:00
pfree(s);
User and database-specific session defaults for run-time configuration variables. New commands ALTER DATABASE ... SET and ALTER USER ... SET.
2002-03-01 23:45:19 +01:00
}
}
Repair core dump when trying to delete an entry from an already-NULL datconfig or useconfig field. Per report from Dustin Sallings.
2002-12-02 06:20:47 +01:00
/*
* Add an entry to an option array. The array parameter may be NULL
* to indicate the current table entry is NULL.
*/
User and database-specific session defaults for run-time configuration variables. New commands ALTER DATABASE ... SET and ALTER USER ... SET.
2002-03-01 23:45:19 +01:00
ArrayType *
GUCArrayAdd(ArrayType *array, const char *name, const char *value)
{
Enforce superuser permissions checks during ALTER ROLE/DATABASE SET, rather than during define_custom_variable(). This entails rejecting an ALTER command if the target variable doesn't have a known (non-placeholder) definition, unless the calling user is superuser. When the variable *is* known, we can correctly apply the rule that only superusers can issue ALTER for SUSET parameters. This allows define_custom_variable to apply ALTER's values for SUSET parameters at module load time, secure in the knowledge that only a superuser could have set the ALTER value. This change fixes a longstanding gotcha in the usage of SUSET-level custom parameters; which is a good thing to fix now that plpgsql defines such a parameter.
2010-04-21 22:54:19 +02:00
struct config_generic *record;
User and database-specific session defaults for run-time configuration variables. New commands ALTER DATABASE ... SET and ALTER USER ... SET.
2002-03-01 23:45:19 +01:00
Datum datum;
char *newval;
ArrayType *a;
Assert(name);
Assert(value);
Enforce superuser permissions checks during ALTER ROLE/DATABASE SET, rather than during define_custom_variable(). This entails rejecting an ALTER command if the target variable doesn't have a known (non-placeholder) definition, unless the calling user is superuser. When the variable *is* known, we can correctly apply the rule that only superusers can issue ALTER for SUSET parameters. This allows define_custom_variable to apply ALTER's values for SUSET parameters at module load time, secure in the knowledge that only a superuser could have set the ALTER value. This change fixes a longstanding gotcha in the usage of SUSET-level custom parameters; which is a good thing to fix now that plpgsql defines such a parameter.
2010-04-21 22:54:19 +02:00
/* test if the option is valid and we're allowed to set it */
(void) validate_option_array_item(name, value, false);
User and database-specific session defaults for run-time configuration variables. New commands ALTER DATABASE ... SET and ALTER USER ... SET.
2002-03-01 23:45:19 +01:00
Enforce superuser permissions checks during ALTER ROLE/DATABASE SET, rather than during define_custom_variable(). This entails rejecting an ALTER command if the target variable doesn't have a known (non-placeholder) definition, unless the calling user is superuser. When the variable *is* known, we can correctly apply the rule that only superusers can issue ALTER for SUSET parameters. This allows define_custom_variable to apply ALTER's values for SUSET parameters at module load time, secure in the knowledge that only a superuser could have set the ALTER value. This change fixes a longstanding gotcha in the usage of SUSET-level custom parameters; which is a good thing to fix now that plpgsql defines such a parameter.
2010-04-21 22:54:19 +02:00
/* normalize name (converts obsolete GUC names to modern spellings) */
record = find_option(name, false, WARNING);
if (record)
name = record->name;
Convert variable name to canonical spelling before checking for matches in GUCArrayAdd/GUCArrayDelete. This prevents the multiple-entry bug exhibited by Frank Lupo 28-Jan-2003.
2003-01-28 19:04:02 +01:00
Enforce superuser permissions checks during ALTER ROLE/DATABASE SET, rather than during define_custom_variable(). This entails rejecting an ALTER command if the target variable doesn't have a known (non-placeholder) definition, unless the calling user is superuser. When the variable *is* known, we can correctly apply the rule that only superusers can issue ALTER for SUSET parameters. This allows define_custom_variable to apply ALTER's values for SUSET parameters at module load time, secure in the knowledge that only a superuser could have set the ALTER value. This change fixes a longstanding gotcha in the usage of SUSET-level custom parameters; which is a good thing to fix now that plpgsql defines such a parameter.
2010-04-21 22:54:19 +02:00
/* build new item for array */
User and database-specific session defaults for run-time configuration variables. New commands ALTER DATABASE ... SET and ALTER USER ... SET.
2002-03-01 23:45:19 +01:00
newval = palloc(strlen(name) + 1 + strlen(value) + 1);
sprintf(newval, "%s=%s", name, value);
Simplify and standardize conversions between TEXT datums and ordinary C strings. This patch introduces four support functions cstring_to_text, cstring_to_text_with_len, text_to_cstring, and text_to_cstring_buffer, and two macros CStringGetTextDatum and TextDatumGetCString. A number of existing macros that provided variants on these themes were removed. Most of the places that need to make such conversions now require just one function or macro call, in place of the multiple notational layers that used to be needed. There are no longer any direct calls of textout or textin, and we got most of the places that were using handmade conversions via memcpy (there may be a few still lurking, though). This commit doesn't make any serious effort to eliminate transient memory leaks caused by detoasting toasted text objects before they reach text_to_cstring. We changed PG_GETARG_TEXT_P to PG_GETARG_TEXT_PP in a few places where it was easy, but much more could be done. Brendan Jurd and Tom Lane
2008-03-25 23:42:46 +01:00
datum = CStringGetTextDatum(newval);
Change made to elog: o Change all current CVS messages of NOTICE to WARNING. We were going to do this just before 7.3 beta but it has to be done now, as you will see below. o Change current INFO messages that should be controlled by client_min_messages to NOTICE. o Force remaining INFO messages, like from EXPLAIN, VACUUM VERBOSE, etc. to always go to the client. o Remove INFO from the client_min_messages options and add NOTICE. Seems we do need three non-ERROR elog levels to handle the various behaviors we need for these messages. Regression passed.
2002-03-06 07:10:59 +01:00
User and database-specific session defaults for run-time configuration variables. New commands ALTER DATABASE ... SET and ALTER USER ... SET.
2002-03-01 23:45:19 +01:00
if (array)
{
pgindent run.
2002-09-04 22:31:48 +02:00
int index;
bool isnull;
int i;
User and database-specific session defaults for run-time configuration variables. New commands ALTER DATABASE ... SET and ALTER USER ... SET.
2002-03-01 23:45:19 +01:00
Modify array operations to include array's element type OID in the array header, and to compute sizing and alignment of array elements the same way normal tuple access operations do --- viz, using the tupmacs.h macros att_addlength and att_align. This makes the world safe for arrays of cstrings or intervals, and should make it much easier to write array-type-polymorphic functions; as examples see the cleanups of array_out and contrib/array_iterator. By Joe Conway and Tom Lane.
2002-08-26 19:54:02 +02:00
Assert(ARR_ELEMTYPE(array) == TEXTOID);
Assert(ARR_NDIM(array) == 1);
Assert(ARR_LBOUND(array)[0] == 1);
pgindent run.
2002-09-04 22:31:48 +02:00
index = ARR_DIMS(array)[0] + 1; /* add after end */
User and database-specific session defaults for run-time configuration variables. New commands ALTER DATABASE ... SET and ALTER USER ... SET.
2002-03-01 23:45:19 +01:00
for (i = 1; i <= ARR_DIMS(array)[0]; i++)
{
Datum d;
char *current;
d = array_ref(array, 1, &i,
pgindent run.
2002-09-04 22:31:48 +02:00
-1 /* varlenarray */ ,
-1 /* TEXT's typlen */ ,
false /* TEXT's typbyval */ ,
'i' /* TEXT's typalign */ ,
User and database-specific session defaults for run-time configuration variables. New commands ALTER DATABASE ... SET and ALTER USER ... SET.
2002-03-01 23:45:19 +01:00
&isnull);
Modify array operations to include array's element type OID in the array header, and to compute sizing and alignment of array elements the same way normal tuple access operations do --- viz, using the tupmacs.h macros att_addlength and att_align. This makes the world safe for arrays of cstrings or intervals, and should make it much easier to write array-type-polymorphic functions; as examples see the cleanups of array_out and contrib/array_iterator. By Joe Conway and Tom Lane.
2002-08-26 19:54:02 +02:00
if (isnull)
continue;
Simplify and standardize conversions between TEXT datums and ordinary C strings. This patch introduces four support functions cstring_to_text, cstring_to_text_with_len, text_to_cstring, and text_to_cstring_buffer, and two macros CStringGetTextDatum and TextDatumGetCString. A number of existing macros that provided variants on these themes were removed. Most of the places that need to make such conversions now require just one function or macro call, in place of the multiple notational layers that used to be needed. There are no longer any direct calls of textout or textin, and we got most of the places that were using handmade conversions via memcpy (there may be a few still lurking, though). This commit doesn't make any serious effort to eliminate transient memory leaks caused by detoasting toasted text objects before they reach text_to_cstring. We changed PG_GETARG_TEXT_P to PG_GETARG_TEXT_PP in a few places where it was easy, but much more could be done. Brendan Jurd and Tom Lane
2008-03-25 23:42:46 +01:00
current = TextDatumGetCString(d);
Enforce superuser permissions checks during ALTER ROLE/DATABASE SET, rather than during define_custom_variable(). This entails rejecting an ALTER command if the target variable doesn't have a known (non-placeholder) definition, unless the calling user is superuser. When the variable *is* known, we can correctly apply the rule that only superusers can issue ALTER for SUSET parameters. This allows define_custom_variable to apply ALTER's values for SUSET parameters at module load time, secure in the knowledge that only a superuser could have set the ALTER value. This change fixes a longstanding gotcha in the usage of SUSET-level custom parameters; which is a good thing to fix now that plpgsql defines such a parameter.
2010-04-21 22:54:19 +02:00
/* check for match up through and including '=' */
pgindent run.
2002-09-04 22:31:48 +02:00
if (strncmp(current, newval, strlen(name) + 1) == 0)
User and database-specific session defaults for run-time configuration variables. New commands ALTER DATABASE ... SET and ALTER USER ... SET.
2002-03-01 23:45:19 +01:00
{
index = i;
break;
}
}
Modify array operations to include array's element type OID in the array header, and to compute sizing and alignment of array elements the same way normal tuple access operations do --- viz, using the tupmacs.h macros att_addlength and att_align. This makes the world safe for arrays of cstrings or intervals, and should make it much easier to write array-type-polymorphic functions; as examples see the cleanups of array_out and contrib/array_iterator. By Joe Conway and Tom Lane.
2002-08-26 19:54:02 +02:00
a = array_set(array, 1, &index,
datum,
Make SQL arrays support null elements. This commit fixes the core array functionality, but I still need to make another pass looking at places that incidentally use arrays (such as ACL manipulation) to make sure they are null-safe. Contrib needs work too. I have not changed the behaviors that are still under discussion about array comparison and what to do with lower bounds.
2005-11-17 23:14:56 +01:00
false,
-1 /* varlena array */ ,
pgindent run.
2002-09-04 22:31:48 +02:00
-1 /* TEXT's typlen */ ,
false /* TEXT's typbyval */ ,
Make SQL arrays support null elements. This commit fixes the core array functionality, but I still need to make another pass looking at places that incidentally use arrays (such as ACL manipulation) to make sure they are null-safe. Contrib needs work too. I have not changed the behaviors that are still under discussion about array comparison and what to do with lower bounds.
2005-11-17 23:14:56 +01:00
'i' /* TEXT's typalign */ );
User and database-specific session defaults for run-time configuration variables. New commands ALTER DATABASE ... SET and ALTER USER ... SET.
2002-03-01 23:45:19 +01:00
}
else
Modify array operations to include array's element type OID in the array header, and to compute sizing and alignment of array elements the same way normal tuple access operations do --- viz, using the tupmacs.h macros att_addlength and att_align. This makes the world safe for arrays of cstrings or intervals, and should make it much easier to write array-type-polymorphic functions; as examples see the cleanups of array_out and contrib/array_iterator. By Joe Conway and Tom Lane.
2002-08-26 19:54:02 +02:00
a = construct_array(&datum, 1,
TEXTOID,
-1, false, 'i');
User and database-specific session defaults for run-time configuration variables. New commands ALTER DATABASE ... SET and ALTER USER ... SET.
2002-03-01 23:45:19 +01:00
return a;
}
Repair core dump when trying to delete an entry from an already-NULL datconfig or useconfig field. Per report from Dustin Sallings.
2002-12-02 06:20:47 +01:00
/*
* Delete an entry from an option array. The array parameter may be NULL
* to indicate the current table entry is NULL. Also, if the return value
* is NULL then a null should be stored.
*/
User and database-specific session defaults for run-time configuration variables. New commands ALTER DATABASE ... SET and ALTER USER ... SET.
2002-03-01 23:45:19 +01:00
ArrayType *
GUCArrayDelete(ArrayType *array, const char *name)
{
Enforce superuser permissions checks during ALTER ROLE/DATABASE SET, rather than during define_custom_variable(). This entails rejecting an ALTER command if the target variable doesn't have a known (non-placeholder) definition, unless the calling user is superuser. When the variable *is* known, we can correctly apply the rule that only superusers can issue ALTER for SUSET parameters. This allows define_custom_variable to apply ALTER's values for SUSET parameters at module load time, secure in the knowledge that only a superuser could have set the ALTER value. This change fixes a longstanding gotcha in the usage of SUSET-level custom parameters; which is a good thing to fix now that plpgsql defines such a parameter.
2010-04-21 22:54:19 +02:00
struct config_generic *record;
pgindent run.
2002-09-04 22:31:48 +02:00
ArrayType *newarray;
int i;
int index;
User and database-specific session defaults for run-time configuration variables. New commands ALTER DATABASE ... SET and ALTER USER ... SET.
2002-03-01 23:45:19 +01:00
Assert(name);
Enforce superuser permissions checks during ALTER ROLE/DATABASE SET, rather than during define_custom_variable(). This entails rejecting an ALTER command if the target variable doesn't have a known (non-placeholder) definition, unless the calling user is superuser. When the variable *is* known, we can correctly apply the rule that only superusers can issue ALTER for SUSET parameters. This allows define_custom_variable to apply ALTER's values for SUSET parameters at module load time, secure in the knowledge that only a superuser could have set the ALTER value. This change fixes a longstanding gotcha in the usage of SUSET-level custom parameters; which is a good thing to fix now that plpgsql defines such a parameter.
2010-04-21 22:54:19 +02:00
/* test if the option is valid and we're allowed to set it */
(void) validate_option_array_item(name, NULL, false);
User and database-specific session defaults for run-time configuration variables. New commands ALTER DATABASE ... SET and ALTER USER ... SET.
2002-03-01 23:45:19 +01:00
Enforce superuser permissions checks during ALTER ROLE/DATABASE SET, rather than during define_custom_variable(). This entails rejecting an ALTER command if the target variable doesn't have a known (non-placeholder) definition, unless the calling user is superuser. When the variable *is* known, we can correctly apply the rule that only superusers can issue ALTER for SUSET parameters. This allows define_custom_variable to apply ALTER's values for SUSET parameters at module load time, secure in the knowledge that only a superuser could have set the ALTER value. This change fixes a longstanding gotcha in the usage of SUSET-level custom parameters; which is a good thing to fix now that plpgsql defines such a parameter.
2010-04-21 22:54:19 +02:00
/* normalize name (converts obsolete GUC names to modern spellings) */
record = find_option(name, false, WARNING);
if (record)
name = record->name;
Convert variable name to canonical spelling before checking for matches in GUCArrayAdd/GUCArrayDelete. This prevents the multiple-entry bug exhibited by Frank Lupo 28-Jan-2003.
2003-01-28 19:04:02 +01:00
Repair core dump when trying to delete an entry from an already-NULL datconfig or useconfig field. Per report from Dustin Sallings.
2002-12-02 06:20:47 +01:00
/* if array is currently null, then surely nothing to delete */
if (!array)
return NULL;
newarray = NULL;
User and database-specific session defaults for run-time configuration variables. New commands ALTER DATABASE ... SET and ALTER USER ... SET.
2002-03-01 23:45:19 +01:00
index = 1;
for (i = 1; i <= ARR_DIMS(array)[0]; i++)
{
Datum d;
char *val;
bool isnull;
d = array_ref(array, 1, &i,
pgindent run.
2002-09-04 22:31:48 +02:00
-1 /* varlenarray */ ,
-1 /* TEXT's typlen */ ,
false /* TEXT's typbyval */ ,
'i' /* TEXT's typalign */ ,
User and database-specific session defaults for run-time configuration variables. New commands ALTER DATABASE ... SET and ALTER USER ... SET.
2002-03-01 23:45:19 +01:00
&isnull);
Modify array operations to include array's element type OID in the array header, and to compute sizing and alignment of array elements the same way normal tuple access operations do --- viz, using the tupmacs.h macros att_addlength and att_align. This makes the world safe for arrays of cstrings or intervals, and should make it much easier to write array-type-polymorphic functions; as examples see the cleanups of array_out and contrib/array_iterator. By Joe Conway and Tom Lane.
2002-08-26 19:54:02 +02:00
if (isnull)
continue;
Simplify and standardize conversions between TEXT datums and ordinary C strings. This patch introduces four support functions cstring_to_text, cstring_to_text_with_len, text_to_cstring, and text_to_cstring_buffer, and two macros CStringGetTextDatum and TextDatumGetCString. A number of existing macros that provided variants on these themes were removed. Most of the places that need to make such conversions now require just one function or macro call, in place of the multiple notational layers that used to be needed. There are no longer any direct calls of textout or textin, and we got most of the places that were using handmade conversions via memcpy (there may be a few still lurking, though). This commit doesn't make any serious effort to eliminate transient memory leaks caused by detoasting toasted text objects before they reach text_to_cstring. We changed PG_GETARG_TEXT_P to PG_GETARG_TEXT_PP in a few places where it was easy, but much more could be done. Brendan Jurd and Tom Lane
2008-03-25 23:42:46 +01:00
val = TextDatumGetCString(d);
User and database-specific session defaults for run-time configuration variables. New commands ALTER DATABASE ... SET and ALTER USER ... SET.
2002-03-01 23:45:19 +01:00
Repair core dump when trying to delete an entry from an already-NULL datconfig or useconfig field. Per report from Dustin Sallings.
2002-12-02 06:20:47 +01:00
/* ignore entry if it's what we want to delete */
pgindent run.
2002-09-04 22:31:48 +02:00
if (strncmp(val, name, strlen(name)) == 0
User and database-specific session defaults for run-time configuration variables. New commands ALTER DATABASE ... SET and ALTER USER ... SET.
2002-03-01 23:45:19 +01:00
&& val[strlen(name)] == '=')
continue;
Repair core dump when trying to delete an entry from an already-NULL datconfig or useconfig field. Per report from Dustin Sallings.
2002-12-02 06:20:47 +01:00
/* else add it to the output array */
if (newarray)
newarray = array_set(newarray, 1, &index,
d,
Make SQL arrays support null elements. This commit fixes the core array functionality, but I still need to make another pass looking at places that incidentally use arrays (such as ACL manipulation) to make sure they are null-safe. Contrib needs work too. I have not changed the behaviors that are still under discussion about array comparison and what to do with lower bounds.
2005-11-17 23:14:56 +01:00
false,
Repair core dump when trying to delete an entry from an already-NULL datconfig or useconfig field. Per report from Dustin Sallings.
2002-12-02 06:20:47 +01:00
-1 /* varlenarray */ ,
-1 /* TEXT's typlen */ ,
false /* TEXT's typbyval */ ,
Make SQL arrays support null elements. This commit fixes the core array functionality, but I still need to make another pass looking at places that incidentally use arrays (such as ACL manipulation) to make sure they are null-safe. Contrib needs work too. I have not changed the behaviors that are still under discussion about array comparison and what to do with lower bounds.
2005-11-17 23:14:56 +01:00
'i' /* TEXT's typalign */ );
Repair core dump when trying to delete an entry from an already-NULL datconfig or useconfig field. Per report from Dustin Sallings.
2002-12-02 06:20:47 +01:00
else
newarray = construct_array(&d, 1,
TEXTOID,
-1, false, 'i');
User and database-specific session defaults for run-time configuration variables. New commands ALTER DATABASE ... SET and ALTER USER ... SET.
2002-03-01 23:45:19 +01:00
index++;
}
return newarray;
}
Add GUC variable to print original query to the server logs when there is an error, warning, etc. Gavin Sherry
2002-09-02 07:42:54 +02:00
Enforce superuser permissions checks during ALTER ROLE/DATABASE SET, rather than during define_custom_variable(). This entails rejecting an ALTER command if the target variable doesn't have a known (non-placeholder) definition, unless the calling user is superuser. When the variable *is* known, we can correctly apply the rule that only superusers can issue ALTER for SUSET parameters. This allows define_custom_variable to apply ALTER's values for SUSET parameters at module load time, secure in the knowledge that only a superuser could have set the ALTER value. This change fixes a longstanding gotcha in the usage of SUSET-level custom parameters; which is a good thing to fix now that plpgsql defines such a parameter.
2010-04-21 22:54:19 +02:00
Prevent ALTER USER f RESET ALL from removing the settings that were put there by a superuser -- "ALTER USER f RESET setting" already disallows removing such a setting. Apply the same treatment to ALTER DATABASE d RESET ALL when run by a database owner that's not superuser.
2010-03-25 15:44:34 +01:00
/*
* Given a GUC array, delete all settings from it that our permission
* level allows: if superuser, delete them all; if regular user, only
* those that are PGC_USERSET
*/
ArrayType *
GUCArrayReset(ArrayType *array)
{
ArrayType *newarray;
int i;
int index;
/* if array is currently null, nothing to do */
if (!array)
return NULL;
Enforce superuser permissions checks during ALTER ROLE/DATABASE SET, rather than during define_custom_variable(). This entails rejecting an ALTER command if the target variable doesn't have a known (non-placeholder) definition, unless the calling user is superuser. When the variable *is* known, we can correctly apply the rule that only superusers can issue ALTER for SUSET parameters. This allows define_custom_variable to apply ALTER's values for SUSET parameters at module load time, secure in the knowledge that only a superuser could have set the ALTER value. This change fixes a longstanding gotcha in the usage of SUSET-level custom parameters; which is a good thing to fix now that plpgsql defines such a parameter.
2010-04-21 22:54:19 +02:00
/* if we're superuser, we can delete everything, so just do it */
Prevent ALTER USER f RESET ALL from removing the settings that were put there by a superuser -- "ALTER USER f RESET setting" already disallows removing such a setting. Apply the same treatment to ALTER DATABASE d RESET ALL when run by a database owner that's not superuser.
2010-03-25 15:44:34 +01:00
if (superuser())
return NULL;
newarray = NULL;
index = 1;
for (i = 1; i <= ARR_DIMS(array)[0]; i++)
{
Datum d;
char *val;
char *eqsgn;
bool isnull;
d = array_ref(array, 1, &i,
-1 /* varlenarray */ ,
-1 /* TEXT's typlen */ ,
false /* TEXT's typbyval */ ,
'i' /* TEXT's typalign */ ,
&isnull);
if (isnull)
continue;
val = TextDatumGetCString(d);
eqsgn = strchr(val, '=');
*eqsgn = '\0';
Enforce superuser permissions checks during ALTER ROLE/DATABASE SET, rather than during define_custom_variable(). This entails rejecting an ALTER command if the target variable doesn't have a known (non-placeholder) definition, unless the calling user is superuser. When the variable *is* known, we can correctly apply the rule that only superusers can issue ALTER for SUSET parameters. This allows define_custom_variable to apply ALTER's values for SUSET parameters at module load time, secure in the knowledge that only a superuser could have set the ALTER value. This change fixes a longstanding gotcha in the usage of SUSET-level custom parameters; which is a good thing to fix now that plpgsql defines such a parameter.
2010-04-21 22:54:19 +02:00
/* skip if we have permission to delete it */
if (validate_option_array_item(val, NULL, true))
Prevent ALTER USER f RESET ALL from removing the settings that were put there by a superuser -- "ALTER USER f RESET setting" already disallows removing such a setting. Apply the same treatment to ALTER DATABASE d RESET ALL when run by a database owner that's not superuser.
2010-03-25 15:44:34 +01:00
continue;
/* else add it to the output array */
if (newarray)
newarray = array_set(newarray, 1, &index,
d,
false,
-1 /* varlenarray */ ,
-1 /* TEXT's typlen */ ,
false /* TEXT's typbyval */ ,
'i' /* TEXT's typalign */ );
else
newarray = construct_array(&d, 1,
TEXTOID,
-1, false, 'i');
index++;
pfree(val);
}
return newarray;
}
Enforce superuser permissions checks during ALTER ROLE/DATABASE SET, rather than during define_custom_variable(). This entails rejecting an ALTER command if the target variable doesn't have a known (non-placeholder) definition, unless the calling user is superuser. When the variable *is* known, we can correctly apply the rule that only superusers can issue ALTER for SUSET parameters. This allows define_custom_variable to apply ALTER's values for SUSET parameters at module load time, secure in the knowledge that only a superuser could have set the ALTER value. This change fixes a longstanding gotcha in the usage of SUSET-level custom parameters; which is a good thing to fix now that plpgsql defines such a parameter.
2010-04-21 22:54:19 +02:00
/*
* Validate a proposed option setting for GUCArrayAdd/Delete/Reset.
*
* name is the option name. value is the proposed value for the Add case,
pgindent run for 9.0, second run
2010-07-06 21:19:02 +02:00
* or NULL for the Delete/Reset cases. If skipIfNoPermissions is true, it's
Enforce superuser permissions checks during ALTER ROLE/DATABASE SET, rather than during define_custom_variable(). This entails rejecting an ALTER command if the target variable doesn't have a known (non-placeholder) definition, unless the calling user is superuser. When the variable *is* known, we can correctly apply the rule that only superusers can issue ALTER for SUSET parameters. This allows define_custom_variable to apply ALTER's values for SUSET parameters at module load time, secure in the knowledge that only a superuser could have set the ALTER value. This change fixes a longstanding gotcha in the usage of SUSET-level custom parameters; which is a good thing to fix now that plpgsql defines such a parameter.
2010-04-21 22:54:19 +02:00
* not an error to have no permissions to set the option.
*
* Returns TRUE if OK, FALSE if skipIfNoPermissions is true and user does not
* have permission to change this option (all other error cases result in an
* error being thrown).
*/
static bool
validate_option_array_item(const char *name, const char *value,
bool skipIfNoPermissions)
{
struct config_generic *gconf;
/*
* There are three cases to consider:
*
* name is a known GUC variable. Check the value normally, check
* permissions normally (ie, allow if variable is USERSET, or if it's
* SUSET and user is superuser).
*
Remove the custom_variable_classes parameter. This variable provides only marginal error-prevention capability (since it can only check the prefix of a qualified GUC name), and the consensus is that that isn't worth the amount of hassle that maintaining the setting creates for DBAs. So, let's just remove it. With this commit, the system will silently accept a value for any qualified GUC name at all, whether it has anything to do with any known extension or not. (Unqualified names still have to match known built-in settings, though; and you will get a WARNING at extension load time if there's an unrecognized setting with that extension's prefix.) There's still some discussion ongoing about whether to tighten that up and if so how; but if we do come up with a solution, it's not likely to look anything like custom_variable_classes.
2011-10-04 18:36:18 +02:00
* name is not known, but exists or can be created as a placeholder (i.e.,
* it has a prefixed name). We allow this case if you're a superuser,
* otherwise not. Superusers are assumed to know what they're doing.
* We can't allow it for other users, because when the placeholder is
* resolved it might turn out to be a SUSET variable;
* define_custom_variable assumes we checked that.
Enforce superuser permissions checks during ALTER ROLE/DATABASE SET, rather than during define_custom_variable(). This entails rejecting an ALTER command if the target variable doesn't have a known (non-placeholder) definition, unless the calling user is superuser. When the variable *is* known, we can correctly apply the rule that only superusers can issue ALTER for SUSET parameters. This allows define_custom_variable to apply ALTER's values for SUSET parameters at module load time, secure in the knowledge that only a superuser could have set the ALTER value. This change fixes a longstanding gotcha in the usage of SUSET-level custom parameters; which is a good thing to fix now that plpgsql defines such a parameter.
2010-04-21 22:54:19 +02:00
*
* name is not known and can't be created as a placeholder. Throw error,
Remove the custom_variable_classes parameter. This variable provides only marginal error-prevention capability (since it can only check the prefix of a qualified GUC name), and the consensus is that that isn't worth the amount of hassle that maintaining the setting creates for DBAs. So, let's just remove it. With this commit, the system will silently accept a value for any qualified GUC name at all, whether it has anything to do with any known extension or not. (Unqualified names still have to match known built-in settings, though; and you will get a WARNING at extension load time if there's an unrecognized setting with that extension's prefix.) There's still some discussion ongoing about whether to tighten that up and if so how; but if we do come up with a solution, it's not likely to look anything like custom_variable_classes.
2011-10-04 18:36:18 +02:00
* unless skipIfNoPermissions is true, in which case return FALSE.
Enforce superuser permissions checks during ALTER ROLE/DATABASE SET, rather than during define_custom_variable(). This entails rejecting an ALTER command if the target variable doesn't have a known (non-placeholder) definition, unless the calling user is superuser. When the variable *is* known, we can correctly apply the rule that only superusers can issue ALTER for SUSET parameters. This allows define_custom_variable to apply ALTER's values for SUSET parameters at module load time, secure in the knowledge that only a superuser could have set the ALTER value. This change fixes a longstanding gotcha in the usage of SUSET-level custom parameters; which is a good thing to fix now that plpgsql defines such a parameter.
2010-04-21 22:54:19 +02:00
*/
gconf = find_option(name, true, WARNING);
if (!gconf)
{
/* not known, failed to make a placeholder */
if (skipIfNoPermissions)
return false;
ereport(ERROR,
(errcode(ERRCODE_UNDEFINED_OBJECT),
Remove the custom_variable_classes parameter. This variable provides only marginal error-prevention capability (since it can only check the prefix of a qualified GUC name), and the consensus is that that isn't worth the amount of hassle that maintaining the setting creates for DBAs. So, let's just remove it. With this commit, the system will silently accept a value for any qualified GUC name at all, whether it has anything to do with any known extension or not. (Unqualified names still have to match known built-in settings, though; and you will get a WARNING at extension load time if there's an unrecognized setting with that extension's prefix.) There's still some discussion ongoing about whether to tighten that up and if so how; but if we do come up with a solution, it's not likely to look anything like custom_variable_classes.
2011-10-04 18:36:18 +02:00
errmsg("unrecognized configuration parameter \"%s\"",
name)));
Enforce superuser permissions checks during ALTER ROLE/DATABASE SET, rather than during define_custom_variable(). This entails rejecting an ALTER command if the target variable doesn't have a known (non-placeholder) definition, unless the calling user is superuser. When the variable *is* known, we can correctly apply the rule that only superusers can issue ALTER for SUSET parameters. This allows define_custom_variable to apply ALTER's values for SUSET parameters at module load time, secure in the knowledge that only a superuser could have set the ALTER value. This change fixes a longstanding gotcha in the usage of SUSET-level custom parameters; which is a good thing to fix now that plpgsql defines such a parameter.
2010-04-21 22:54:19 +02:00
}
if (gconf->flags & GUC_CUSTOM_PLACEHOLDER)
{
/*
* We cannot do any meaningful check on the value, so only permissions
* are useful to check.
*/
if (superuser())
return true;
if (skipIfNoPermissions)
return false;
ereport(ERROR,
(errcode(ERRCODE_INSUFFICIENT_PRIVILEGE),
errmsg("permission denied to set parameter \"%s\"", name)));
}
/* manual permissions check so we can avoid an error being thrown */
if (gconf->context == PGC_USERSET)
pgindent run for 9.0, second run
2010-07-06 21:19:02 +02:00
/* ok */ ;
Enforce superuser permissions checks during ALTER ROLE/DATABASE SET, rather than during define_custom_variable(). This entails rejecting an ALTER command if the target variable doesn't have a known (non-placeholder) definition, unless the calling user is superuser. When the variable *is* known, we can correctly apply the rule that only superusers can issue ALTER for SUSET parameters. This allows define_custom_variable to apply ALTER's values for SUSET parameters at module load time, secure in the knowledge that only a superuser could have set the ALTER value. This change fixes a longstanding gotcha in the usage of SUSET-level custom parameters; which is a good thing to fix now that plpgsql defines such a parameter.
2010-04-21 22:54:19 +02:00
else if (gconf->context == PGC_SUSET && superuser())
pgindent run for 9.0, second run
2010-07-06 21:19:02 +02:00
/* ok */ ;
Enforce superuser permissions checks during ALTER ROLE/DATABASE SET, rather than during define_custom_variable(). This entails rejecting an ALTER command if the target variable doesn't have a known (non-placeholder) definition, unless the calling user is superuser. When the variable *is* known, we can correctly apply the rule that only superusers can issue ALTER for SUSET parameters. This allows define_custom_variable to apply ALTER's values for SUSET parameters at module load time, secure in the knowledge that only a superuser could have set the ALTER value. This change fixes a longstanding gotcha in the usage of SUSET-level custom parameters; which is a good thing to fix now that plpgsql defines such a parameter.
2010-04-21 22:54:19 +02:00
else if (skipIfNoPermissions)
return false;
/* if a permissions error should be thrown, let set_config_option do it */
/* test for permissions and valid option value */
Restructure error handling in reading of postgresql.conf. This patch has two distinct purposes: to report multiple problems in postgresql.conf rather than always bailing out after the first one, and to change the policy for whether changes are applied when there are unrelated errors in postgresql.conf. Formerly the policy was to apply no changes if any errors could be detected, but that had a significant consistency problem, because in some cases specific values might be seen as valid by some processes but invalid by others. This meant that the latter processes would fail to adopt changes in other parameters even though the former processes had done so. The new policy is that during SIGHUP, the file is rejected as a whole if there are any errors in the "name = value" syntax, or if any lines attempt to set nonexistent built-in parameters, or if any lines attempt to set custom parameters whose prefix is not listed in (the new value of) custom_variable_classes. These tests should always give the same results in all processes, and provide what seems a reasonably robust defense against loading values from badly corrupted config files. If these tests pass, all processes will apply all settings that they individually see as good, ignoring (but logging) any they don't. In addition, the postmaster does not abandon reading a configuration file after the first syntax error, but continues to read the file and report syntax errors (up to a maximum of 100 syntax errors per file). The postmaster will still refuse to start up if the configuration file contains any errors at startup time, but these changes allow multiple errors to be detected and reported before quitting. Alexey Klyukin, reviewed by Andy Colson and av (Alexander ?) with some additional hacking by Tom Lane
2011-10-02 22:50:04 +02:00
(void) set_config_option(name, value,
superuser() ? PGC_SUSET : PGC_USERSET,
PGC_S_TEST, GUC_ACTION_SET, false);
Enforce superuser permissions checks during ALTER ROLE/DATABASE SET, rather than during define_custom_variable(). This entails rejecting an ALTER command if the target variable doesn't have a known (non-placeholder) definition, unless the calling user is superuser. When the variable *is* known, we can correctly apply the rule that only superusers can issue ALTER for SUSET parameters. This allows define_custom_variable to apply ALTER's values for SUSET parameters at module load time, secure in the knowledge that only a superuser could have set the ALTER value. This change fixes a longstanding gotcha in the usage of SUSET-level custom parameters; which is a good thing to fix now that plpgsql defines such a parameter.
2010-04-21 22:54:19 +02:00
return true;
}
In the continuing saga of FE/BE protocol revisions, add reporting of initial values and runtime changes in selected parameters. This gets rid of the need for an initial 'select pg_client_encoding()' query in libpq, bringing us back to one message transmitted in each direction for a standard connection startup. To allow server version to be sent using the same GUC mechanism that handles other parameters, invent the concept of a never-settable GUC parameter: you can 'show server_version' but it's not settable by any GUC input source. Create 'lc_collate' and 'lc_ctype' never-settable parameters so that people can find out these settings without need for pg_controldata. (These side ideas were all discussed some time ago in pgsql-hackers, but not yet implemented.)
2003-04-25 21:45:10 +02:00
/*
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
* Called by check_hooks that want to override the normal
* ERRCODE_INVALID_PARAMETER_VALUE SQLSTATE for check hook failures.
*
* Note that GUC_check_errmsg() etc are just macros that result in a direct
pgindent run before PG 9.1 beta 1.
2011-04-10 17:42:00 +02:00
* assignment to the associated variables. That is ugly, but forced by the
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
* limitations of C's macro mechanisms.
In the continuing saga of FE/BE protocol revisions, add reporting of initial values and runtime changes in selected parameters. This gets rid of the need for an initial 'select pg_client_encoding()' query in libpq, bringing us back to one message transmitted in each direction for a standard connection startup. To allow server version to be sent using the same GUC mechanism that handles other parameters, invent the concept of a never-settable GUC parameter: you can 'show server_version' but it's not settable by any GUC input source. Create 'lc_collate' and 'lc_ctype' never-settable parameters so that people can find out these settings without need for pg_controldata. (These side ideas were all discussed some time ago in pgsql-hackers, but not yet implemented.)
2003-04-25 21:45:10 +02:00
*/
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
void
GUC_check_errcode(int sqlerrcode)
{
GUC_check_errcode_value = sqlerrcode;
}
In the continuing saga of FE/BE protocol revisions, add reporting of initial values and runtime changes in selected parameters. This gets rid of the need for an initial 'select pg_client_encoding()' query in libpq, bringing us back to one message transmitted in each direction for a standard connection startup. To allow server version to be sent using the same GUC mechanism that handles other parameters, invent the concept of a never-settable GUC parameter: you can 'show server_version' but it's not settable by any GUC input source. Create 'lc_collate' and 'lc_ctype' never-settable parameters so that people can find out these settings without need for pg_controldata. (These side ideas were all discussed some time ago in pgsql-hackers, but not yet implemented.)
2003-04-25 21:45:10 +02:00
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
/*
* Convenience functions to manage calling a variable's check_hook.
* These mostly take care of the protocol for letting check hooks supply
* portions of the error report on failure.
*/
static bool
pgindent run before PG 9.1 beta 1.
2011-04-10 17:42:00 +02:00
call_bool_check_hook(struct config_bool * conf, bool *newval, void **extra,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
GucSource source, int elevel)
{
/* Quick success if no hook */
if (!conf->check_hook)
return true;
/* Reset variables that might be set by hook */
GUC_check_errcode_value = ERRCODE_INVALID_PARAMETER_VALUE;
GUC_check_errmsg_string = NULL;
GUC_check_errdetail_string = NULL;
GUC_check_errhint_string = NULL;
if (!(*conf->check_hook) (newval, extra, source))
{
ereport(elevel,
(errcode(GUC_check_errcode_value),
GUC_check_errmsg_string ?
Replace errdetail("%s", ...) with errdetail_internal("%s", ...). There may be some other places where we should use errdetail_internal, but they'll have to be evaluated case-by-case. This commit just hits a bunch of places where invoking gettext is obviously a waste of cycles.
2011-07-16 20:21:12 +02:00
errmsg_internal("%s", GUC_check_errmsg_string) :
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
errmsg("invalid value for parameter \"%s\": %d",
conf->gen.name, (int) *newval),
GUC_check_errdetail_string ?
Replace errdetail("%s", ...) with errdetail_internal("%s", ...). There may be some other places where we should use errdetail_internal, but they'll have to be evaluated case-by-case. This commit just hits a bunch of places where invoking gettext is obviously a waste of cycles.
2011-07-16 20:21:12 +02:00
errdetail_internal("%s", GUC_check_errdetail_string) : 0,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
GUC_check_errhint_string ?
errhint("%s", GUC_check_errhint_string) : 0));
/* Flush any strings created in ErrorContext */
FlushErrorState();
return false;
}
return true;
}
static bool
pgindent run before PG 9.1 beta 1.
2011-04-10 17:42:00 +02:00
call_int_check_hook(struct config_int * conf, int *newval, void **extra,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
GucSource source, int elevel)
{
/* Quick success if no hook */
if (!conf->check_hook)
return true;
/* Reset variables that might be set by hook */
GUC_check_errcode_value = ERRCODE_INVALID_PARAMETER_VALUE;
GUC_check_errmsg_string = NULL;
GUC_check_errdetail_string = NULL;
GUC_check_errhint_string = NULL;
if (!(*conf->check_hook) (newval, extra, source))
{
ereport(elevel,
(errcode(GUC_check_errcode_value),
GUC_check_errmsg_string ?
Replace errdetail("%s", ...) with errdetail_internal("%s", ...). There may be some other places where we should use errdetail_internal, but they'll have to be evaluated case-by-case. This commit just hits a bunch of places where invoking gettext is obviously a waste of cycles.
2011-07-16 20:21:12 +02:00
errmsg_internal("%s", GUC_check_errmsg_string) :
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
errmsg("invalid value for parameter \"%s\": %d",
conf->gen.name, *newval),
GUC_check_errdetail_string ?
Replace errdetail("%s", ...) with errdetail_internal("%s", ...). There may be some other places where we should use errdetail_internal, but they'll have to be evaluated case-by-case. This commit just hits a bunch of places where invoking gettext is obviously a waste of cycles.
2011-07-16 20:21:12 +02:00
errdetail_internal("%s", GUC_check_errdetail_string) : 0,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
GUC_check_errhint_string ?
errhint("%s", GUC_check_errhint_string) : 0));
/* Flush any strings created in ErrorContext */
FlushErrorState();
return false;
}
return true;
}
static bool
pgindent run before PG 9.1 beta 1.
2011-04-10 17:42:00 +02:00
call_real_check_hook(struct config_real * conf, double *newval, void **extra,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
GucSource source, int elevel)
{
/* Quick success if no hook */
if (!conf->check_hook)
return true;
/* Reset variables that might be set by hook */
GUC_check_errcode_value = ERRCODE_INVALID_PARAMETER_VALUE;
GUC_check_errmsg_string = NULL;
GUC_check_errdetail_string = NULL;
GUC_check_errhint_string = NULL;
if (!(*conf->check_hook) (newval, extra, source))
{
ereport(elevel,
(errcode(GUC_check_errcode_value),
GUC_check_errmsg_string ?
Replace errdetail("%s", ...) with errdetail_internal("%s", ...). There may be some other places where we should use errdetail_internal, but they'll have to be evaluated case-by-case. This commit just hits a bunch of places where invoking gettext is obviously a waste of cycles.
2011-07-16 20:21:12 +02:00
errmsg_internal("%s", GUC_check_errmsg_string) :
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
errmsg("invalid value for parameter \"%s\": %g",
conf->gen.name, *newval),
GUC_check_errdetail_string ?
Replace errdetail("%s", ...) with errdetail_internal("%s", ...). There may be some other places where we should use errdetail_internal, but they'll have to be evaluated case-by-case. This commit just hits a bunch of places where invoking gettext is obviously a waste of cycles.
2011-07-16 20:21:12 +02:00
errdetail_internal("%s", GUC_check_errdetail_string) : 0,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
GUC_check_errhint_string ?
errhint("%s", GUC_check_errhint_string) : 0));
/* Flush any strings created in ErrorContext */
FlushErrorState();
return false;
}
return true;
}
static bool
pgindent run before PG 9.1 beta 1.
2011-04-10 17:42:00 +02:00
call_string_check_hook(struct config_string * conf, char **newval, void **extra,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
GucSource source, int elevel)
{
/* Quick success if no hook */
if (!conf->check_hook)
return true;
/* Reset variables that might be set by hook */
GUC_check_errcode_value = ERRCODE_INVALID_PARAMETER_VALUE;
GUC_check_errmsg_string = NULL;
GUC_check_errdetail_string = NULL;
GUC_check_errhint_string = NULL;
if (!(*conf->check_hook) (newval, extra, source))
{
ereport(elevel,
(errcode(GUC_check_errcode_value),
GUC_check_errmsg_string ?
Replace errdetail("%s", ...) with errdetail_internal("%s", ...). There may be some other places where we should use errdetail_internal, but they'll have to be evaluated case-by-case. This commit just hits a bunch of places where invoking gettext is obviously a waste of cycles.
2011-07-16 20:21:12 +02:00
errmsg_internal("%s", GUC_check_errmsg_string) :
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
errmsg("invalid value for parameter \"%s\": \"%s\"",
conf->gen.name, *newval ? *newval : ""),
GUC_check_errdetail_string ?
Replace errdetail("%s", ...) with errdetail_internal("%s", ...). There may be some other places where we should use errdetail_internal, but they'll have to be evaluated case-by-case. This commit just hits a bunch of places where invoking gettext is obviously a waste of cycles.
2011-07-16 20:21:12 +02:00
errdetail_internal("%s", GUC_check_errdetail_string) : 0,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
GUC_check_errhint_string ?
errhint("%s", GUC_check_errhint_string) : 0));
/* Flush any strings created in ErrorContext */
FlushErrorState();
return false;
}
return true;
}
static bool
pgindent run before PG 9.1 beta 1.
2011-04-10 17:42:00 +02:00
call_enum_check_hook(struct config_enum * conf, int *newval, void **extra,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
GucSource source, int elevel)
{
/* Quick success if no hook */
if (!conf->check_hook)
return true;
/* Reset variables that might be set by hook */
GUC_check_errcode_value = ERRCODE_INVALID_PARAMETER_VALUE;
GUC_check_errmsg_string = NULL;
GUC_check_errdetail_string = NULL;
GUC_check_errhint_string = NULL;
if (!(*conf->check_hook) (newval, extra, source))
{
ereport(elevel,
(errcode(GUC_check_errcode_value),
GUC_check_errmsg_string ?
Replace errdetail("%s", ...) with errdetail_internal("%s", ...). There may be some other places where we should use errdetail_internal, but they'll have to be evaluated case-by-case. This commit just hits a bunch of places where invoking gettext is obviously a waste of cycles.
2011-07-16 20:21:12 +02:00
errmsg_internal("%s", GUC_check_errmsg_string) :
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
errmsg("invalid value for parameter \"%s\": \"%s\"",
conf->gen.name,
config_enum_lookup_by_value(conf, *newval)),
GUC_check_errdetail_string ?
Replace errdetail("%s", ...) with errdetail_internal("%s", ...). There may be some other places where we should use errdetail_internal, but they'll have to be evaluated case-by-case. This commit just hits a bunch of places where invoking gettext is obviously a waste of cycles.
2011-07-16 20:21:12 +02:00
errdetail_internal("%s", GUC_check_errdetail_string) : 0,
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
GUC_check_errhint_string ?
errhint("%s", GUC_check_errhint_string) : 0));
/* Flush any strings created in ErrorContext */
FlushErrorState();
return false;
}
return true;
}
/*
* check_hook, assign_hook and show_hook subroutines
*/
static bool
check_log_destination(char **newval, void **extra, GucSource source)
Remove 'syslog' GUC variable, and add more logical 'log_destination' variable to control logoutput location on Unix and Win32. Magnus Hagander
2004-04-05 05:02:11 +02:00
{
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
char *rawstring;
List *elemlist;
ListCell *l;
int newlogdest = 0;
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
int *myextra;
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
Remove 'syslog' GUC variable, and add more logical 'log_destination' variable to control logoutput location on Unix and Win32. Magnus Hagander
2004-04-05 05:02:11 +02:00
/* Need a modifiable copy of string */
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
rawstring = pstrdup(*newval);
Remove 'syslog' GUC variable, and add more logical 'log_destination' variable to control logoutput location on Unix and Win32. Magnus Hagander
2004-04-05 05:02:11 +02:00
/* Parse string into list of identifiers */
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
if (!SplitIdentifierString(rawstring, ',', &elemlist))
Remove 'syslog' GUC variable, and add more logical 'log_destination' variable to control logoutput location on Unix and Win32. Magnus Hagander
2004-04-05 05:02:11 +02:00
{
/* syntax error in list */
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
GUC_check_errdetail("List syntax is invalid.");
Remove 'syslog' GUC variable, and add more logical 'log_destination' variable to control logoutput location on Unix and Win32. Magnus Hagander
2004-04-05 05:02:11 +02:00
pfree(rawstring);
Use the new List API function names throughout the backend, and disable the list compatibility API by default. While doing this, I decided to keep the llast() macro around and introduce llast_int() and llast_oid() variants.
2004-05-31 01:40:41 +02:00
list_free(elemlist);
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
return false;
Remove 'syslog' GUC variable, and add more logical 'log_destination' variable to control logoutput location on Unix and Win32. Magnus Hagander
2004-04-05 05:02:11 +02:00
}
foreach(l, elemlist)
{
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
char *tok = (char *) lfirst(l);
if (pg_strcasecmp(tok, "stderr") == 0)
Remove 'syslog' GUC variable, and add more logical 'log_destination' variable to control logoutput location on Unix and Win32. Magnus Hagander
2004-04-05 05:02:11 +02:00
newlogdest |= LOG_DESTINATION_STDERR;
Provide for logfiles in machine readable CSV format. In consequence, rename redirect_stderr to logging_collector. Original patch from Arul Shaji, subsequently modified by Greg Smith, and then heavily modified by me.
2007-08-19 03:41:25 +02:00
else if (pg_strcasecmp(tok, "csvlog") == 0)
pgindent run for 8.3.
2007-11-15 22:14:46 +01:00
newlogdest |= LOG_DESTINATION_CSVLOG;
Remove 'syslog' GUC variable, and add more logical 'log_destination' variable to control logoutput location on Unix and Win32. Magnus Hagander
2004-04-05 05:02:11 +02:00
#ifdef HAVE_SYSLOG
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
else if (pg_strcasecmp(tok, "syslog") == 0)
Remove 'syslog' GUC variable, and add more logical 'log_destination' variable to control logoutput location on Unix and Win32. Magnus Hagander
2004-04-05 05:02:11 +02:00
newlogdest |= LOG_DESTINATION_SYSLOG;
#endif
#ifdef WIN32
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
else if (pg_strcasecmp(tok, "eventlog") == 0)
Remove 'syslog' GUC variable, and add more logical 'log_destination' variable to control logoutput location on Unix and Win32. Magnus Hagander
2004-04-05 05:02:11 +02:00
newlogdest |= LOG_DESTINATION_EVENTLOG;
#endif
Pgindent run for 8.0.
2004-08-29 07:07:03 +02:00
else
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
GUC_check_errdetail("Unrecognized key word: \"%s\".", tok);
Remove 'syslog' GUC variable, and add more logical 'log_destination' variable to control logoutput location on Unix and Win32. Magnus Hagander
2004-04-05 05:02:11 +02:00
pfree(rawstring);
Use the new List API function names throughout the backend, and disable the list compatibility API by default. While doing this, I decided to keep the llast() macro around and introduce llast_int() and llast_oid() variants.
2004-05-31 01:40:41 +02:00
list_free(elemlist);
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
return false;
Remove 'syslog' GUC variable, and add more logical 'log_destination' variable to control logoutput location on Unix and Win32. Magnus Hagander
2004-04-05 05:02:11 +02:00
}
}
pfree(rawstring);
Use the new List API function names throughout the backend, and disable the list compatibility API by default. While doing this, I decided to keep the llast() macro around and introduce llast_int() and llast_oid() variants.
2004-05-31 01:40:41 +02:00
list_free(elemlist);
Remove 'syslog' GUC variable, and add more logical 'log_destination' variable to control logoutput location on Unix and Win32. Magnus Hagander
2004-04-05 05:02:11 +02:00
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
myextra = (int *) guc_malloc(ERROR, sizeof(int));
*myextra = newlogdest;
*extra = (void *) myextra;
return true;
Remove 'syslog' GUC variable, and add more logical 'log_destination' variable to control logoutput location on Unix and Win32. Magnus Hagander
2004-04-05 05:02:11 +02:00
}
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
static void
assign_log_destination(const char *newval, void *extra)
{
Log_destination = *((int *) extra);
}
static void
assign_syslog_facility(int newval, void *extra)
In the continuing saga of FE/BE protocol revisions, add reporting of initial values and runtime changes in selected parameters. This gets rid of the need for an initial 'select pg_client_encoding()' query in libpq, bringing us back to one message transmitted in each direction for a standard connection startup. To allow server version to be sent using the same GUC mechanism that handles other parameters, invent the concept of a never-settable GUC parameter: you can 'show server_version' but it's not settable by any GUC input source. Create 'lc_collate' and 'lc_ctype' never-settable parameters so that people can find out these settings without need for pg_controldata. (These side ideas were all discussed some time ago in pgsql-hackers, but not yet implemented.)
2003-04-25 21:45:10 +02:00
{
Make GUC variables for syslog and SSL always visible Make the variables visible (but not used) even when support is not compiled in.
2010-12-18 16:53:59 +01:00
#ifdef HAVE_SYSLOG
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
set_syslog_parameters(syslog_ident_str ? syslog_ident_str : "postgres",
newval);
Make GUC variables for syslog and SSL always visible Make the variables visible (but not used) even when support is not compiled in.
2010-12-18 16:53:59 +01:00
#endif
/* Without syslog support, just ignore it */
In the continuing saga of FE/BE protocol revisions, add reporting of initial values and runtime changes in selected parameters. This gets rid of the need for an initial 'select pg_client_encoding()' query in libpq, bringing us back to one message transmitted in each direction for a standard connection startup. To allow server version to be sent using the same GUC mechanism that handles other parameters, invent the concept of a never-settable GUC parameter: you can 'show server_version' but it's not settable by any GUC input source. Create 'lc_collate' and 'lc_ctype' never-settable parameters so that people can find out these settings without need for pg_controldata. (These side ideas were all discussed some time ago in pgsql-hackers, but not yet implemented.)
2003-04-25 21:45:10 +02:00
}
Fix syslog bug: if any messages are emitted to write_syslog before the facility has been set, the facility gets set to LOCAL0 and cannot be changed later. This seems reasonably plausible to happen, particularly at higher debug log levels, though I am not certain it explains Han Holl's recent report. Easiest fix is to teach the code how to change the value on-the-fly, which is nicer anyway. I made the settings PGC_SIGHUP to conform with log_destination.
2005-10-14 22:53:56 +02:00
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
static void
assign_syslog_ident(const char *newval, void *extra)
Fix syslog bug: if any messages are emitted to write_syslog before the facility has been set, the facility gets set to LOCAL0 and cannot be changed later. This seems reasonably plausible to happen, particularly at higher debug log levels, though I am not certain it explains Han Holl's recent report. Easiest fix is to teach the code how to change the value on-the-fly, which is nicer anyway. I made the settings PGC_SIGHUP to conform with log_destination.
2005-10-14 22:53:56 +02:00
{
Make GUC variables for syslog and SSL always visible Make the variables visible (but not used) even when support is not compiled in.
2010-12-18 16:53:59 +01:00
#ifdef HAVE_SYSLOG
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
set_syslog_parameters(newval, syslog_facility);
Make GUC variables for syslog and SSL always visible Make the variables visible (but not used) even when support is not compiled in.
2010-12-18 16:53:59 +01:00
#endif
/* Without syslog support, it will always be set to "none", so ignore */
Fix syslog bug: if any messages are emitted to write_syslog before the facility has been set, the facility gets set to LOCAL0 and cannot be changed later. This seems reasonably plausible to happen, particularly at higher debug log levels, though I am not certain it explains Han Holl's recent report. Easiest fix is to teach the code how to change the value on-the-fly, which is nicer anyway. I made the settings PGC_SIGHUP to conform with log_destination.
2005-10-14 22:53:56 +02:00
}
In the continuing saga of FE/BE protocol revisions, add reporting of initial values and runtime changes in selected parameters. This gets rid of the need for an initial 'select pg_client_encoding()' query in libpq, bringing us back to one message transmitted in each direction for a standard connection startup. To allow server version to be sent using the same GUC mechanism that handles other parameters, invent the concept of a never-settable GUC parameter: you can 'show server_version' but it's not settable by any GUC input source. Create 'lc_collate' and 'lc_ctype' never-settable parameters so that people can find out these settings without need for pg_controldata. (These side ideas were all discussed some time ago in pgsql-hackers, but not yet implemented.)
2003-04-25 21:45:10 +02:00
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
static void
assign_session_replication_role(int newval, void *extra)
Changes pg_trigger and extend pg_rewrite in order to allow triggers and rules to be defined with different, per session controllable, behaviors for replication purposes. This will allow replication systems like Slony-I and, as has been stated on pgsql-hackers, other products to control the firing mechanism of triggers and rewrite rules without modifying the system catalog directly. The firing mechanisms are controlled by a new superuser-only GUC variable, session_replication_role, together with a change to pg_trigger.tgenabled and a new column pg_rewrite.ev_enabled. Both columns are a single char data type now (tgenabled was a bool before). The possible values in these attributes are: 'O' - Trigger/Rule fires when session_replication_role is "origin" (default) or "local". This is the default behavior. 'D' - Trigger/Rule is disabled and fires never 'A' - Trigger/Rule fires always regardless of the setting of session_replication_role 'R' - Trigger/Rule fires when session_replication_role is "replica" The GUC variable can only be changed as long as the system does not have any cached query plans. This will prevent changing the session role and accidentally executing stored procedures or functions that have plans cached that expand to the wrong query set due to differences in the rule firing semantics. The SQL syntax for changing a triggers/rules firing semantics is ALTER TABLE <tabname> <when> TRIGGER|RULE <name>; <when> ::= ENABLE | ENABLE ALWAYS | ENABLE REPLICA | DISABLE psql's \d command as well as pg_dump are extended in a backward compatible fashion. Jan
2007-03-20 00:38:32 +01:00
{
Minor editorialization: don't flush plan cache without need.
2007-06-05 23:50:19 +02:00
/*
* Must flush the plan cache when changing replication role; but don't
* flush unnecessarily.
*/
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
if (SessionReplicationRole != newval)
Minor editorialization: don't flush plan cache without need.
2007-06-05 23:50:19 +02:00
ResetPlanCache();
Changes pg_trigger and extend pg_rewrite in order to allow triggers and rules to be defined with different, per session controllable, behaviors for replication purposes. This will allow replication systems like Slony-I and, as has been stated on pgsql-hackers, other products to control the firing mechanism of triggers and rewrite rules without modifying the system catalog directly. The firing mechanisms are controlled by a new superuser-only GUC variable, session_replication_role, together with a change to pg_trigger.tgenabled and a new column pg_rewrite.ev_enabled. Both columns are a single char data type now (tgenabled was a bool before). The possible values in these attributes are: 'O' - Trigger/Rule fires when session_replication_role is "origin" (default) or "local". This is the default behavior. 'D' - Trigger/Rule is disabled and fires never 'A' - Trigger/Rule fires always regardless of the setting of session_replication_role 'R' - Trigger/Rule fires when session_replication_role is "replica" The GUC variable can only be changed as long as the system does not have any cached query plans. This will prevent changing the session role and accidentally executing stored procedures or functions that have plans cached that expand to the wrong query set due to differences in the rule firing semantics. The SQL syntax for changing a triggers/rules firing semantics is ALTER TABLE <tabname> <when> TRIGGER|RULE <name>; <when> ::= ENABLE | ENABLE ALWAYS | ENABLE REPLICA | DISABLE psql's \d command as well as pg_dump are extended in a backward compatible fashion. Jan
2007-03-20 00:38:32 +01:00
}
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
static bool
check_temp_buffers(int *newval, void **extra, GucSource source)
Add temp_buffers GUC variable to allow users to determine the size of the local buffer arena for temporary table access.
2005-03-20 00:27:11 +01:00
{
/*
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
* Once local buffers have been initialized, it's too late to change this.
Add temp_buffers GUC variable to allow users to determine the size of the local buffer arena for temporary table access.
2005-03-20 00:27:11 +01:00
*/
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
if (NLocBuffer && NLocBuffer != *newval)
{
Message style and spelling improvements
2011-06-21 23:33:20 +02:00
GUC_check_errdetail("\"temp_buffers\" cannot be changed after any temporary tables have been accessed in the session.");
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
return false;
}
return true;
Add temp_buffers GUC variable to allow users to determine the size of the local buffer arena for temporary table access.
2005-03-20 00:27:11 +01:00
}
Backend support for autocommit removed, per recent discussions. The only remnant of this failed experiment is that the server will take SET AUTOCOMMIT TO ON. Still TODO: provide some client-side autocommit logic in libpq.
2003-05-14 05:26:03 +02:00
static bool
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
check_phony_autocommit(bool *newval, void **extra, GucSource source)
Backend support for autocommit removed, per recent discussions. The only remnant of this failed experiment is that the server will take SET AUTOCOMMIT TO ON. Still TODO: provide some client-side autocommit logic in libpq.
2003-05-14 05:26:03 +02:00
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
if (!*newval)
Backend support for autocommit removed, per recent discussions. The only remnant of this failed experiment is that the server will take SET AUTOCOMMIT TO ON. Still TODO: provide some client-side autocommit logic in libpq.
2003-05-14 05:26:03 +02:00
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
GUC_check_errcode(ERRCODE_FEATURE_NOT_SUPPORTED);
GUC_check_errmsg("SET AUTOCOMMIT TO OFF is no longer supported");
Backend support for autocommit removed, per recent discussions. The only remnant of this failed experiment is that the server will take SET AUTOCOMMIT TO ON. Still TODO: provide some client-side autocommit logic in libpq.
2003-05-14 05:26:03 +02:00
return false;
}
return true;
}
Make all command-line options of postmaster and postgres the same. See http://archives.postgresql.org/pgsql-hackers/2006-01/msg00151.php for the complete plan.
2006-01-05 11:07:46 +01:00
static bool
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
check_debug_assertions(bool *newval, void **extra, GucSource source)
Make all command-line options of postmaster and postgres the same. See http://archives.postgresql.org/pgsql-hackers/2006-01/msg00151.php for the complete plan.
2006-01-05 11:07:46 +01:00
{
#ifndef USE_ASSERT_CHECKING
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
if (*newval)
Improve consistency of error reporting in GUC assign_hook routines. Some were reporting ERROR for interactive assignments and LOG for other cases, some were saying nothing for non-interactive cases, and a few did yet other things. Make them use a new function GUC_complaint_elevel() to establish a reasonably uniform policy about how to report. There are still a few edge cases such as assign_search_path(), but it's much better than before. Per gripe from Devrim Gunduz and subsequent discussion. As noted by Alvaro, it'd be better to fold these custom messages into the standard "invalid parameter value" complaint from guc.c, perhaps as the DETAIL field. However that will require more redesign than seems prudent for 8.3. This is a relatively safe, low-impact change that we can afford to risk now.
2007-12-28 01:23:23 +01:00
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
GUC_check_errmsg("assertion checking is not supported by this build");
Improve consistency of error reporting in GUC assign_hook routines. Some were reporting ERROR for interactive assignments and LOG for other cases, some were saying nothing for non-interactive cases, and a few did yet other things. Make them use a new function GUC_complaint_elevel() to establish a reasonably uniform policy about how to report. There are still a few edge cases such as assign_search_path(), but it's much better than before. Per gripe from Devrim Gunduz and subsequent discussion. As noted by Alvaro, it'd be better to fold these custom messages into the standard "invalid parameter value" complaint from guc.c, perhaps as the DETAIL field. However that will require more redesign than seems prudent for 8.3. This is a relatively safe, low-impact change that we can afford to risk now.
2007-12-28 01:23:23 +01:00
return false;
}
Make all command-line options of postmaster and postgres the same. See http://archives.postgresql.org/pgsql-hackers/2006-01/msg00151.php for the complete plan.
2006-01-05 11:07:46 +01:00
#endif
return true;
}
Add a boolean GUC parameter "bonjour" to control whether a Bonjour-enabled build actually attempts to advertise itself via Bonjour. Formerly it always did so, which meant that packagers had to decide for their users whether this behavior was wanted or not. The default is "off" to be on the safe side, though this represents a change in the default behavior of a Bonjour-enabled build. Per discussion.
2009-09-08 19:08:36 +02:00
static bool
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
check_bonjour(bool *newval, void **extra, GucSource source)
Add a boolean GUC parameter "bonjour" to control whether a Bonjour-enabled build actually attempts to advertise itself via Bonjour. Formerly it always did so, which meant that packagers had to decide for their users whether this behavior was wanted or not. The default is "off" to be on the safe side, though this represents a change in the default behavior of a Bonjour-enabled build. Per discussion.
2009-09-08 19:08:36 +02:00
{
#ifndef USE_BONJOUR
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
if (*newval)
Add a boolean GUC parameter "bonjour" to control whether a Bonjour-enabled build actually attempts to advertise itself via Bonjour. Formerly it always did so, which meant that packagers had to decide for their users whether this behavior was wanted or not. The default is "off" to be on the safe side, though this represents a change in the default behavior of a Bonjour-enabled build. Per discussion.
2009-09-08 19:08:36 +02:00
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
GUC_check_errmsg("Bonjour is not supported by this build");
Add a boolean GUC parameter "bonjour" to control whether a Bonjour-enabled build actually attempts to advertise itself via Bonjour. Formerly it always did so, which meant that packagers had to decide for their users whether this behavior was wanted or not. The default is "off" to be on the safe side, though this represents a change in the default behavior of a Bonjour-enabled build. Per discussion.
2009-09-08 19:08:36 +02:00
return false;
}
#endif
return true;
}
Disallow setting ssl = on if SSL is not compiled in.
2005-12-30 01:13:50 +01:00
static bool
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
check_ssl(bool *newval, void **extra, GucSource source)
Disallow setting ssl = on if SSL is not compiled in.
2005-12-30 01:13:50 +01:00
{
#ifndef USE_SSL
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
if (*newval)
Improve consistency of error reporting in GUC assign_hook routines. Some were reporting ERROR for interactive assignments and LOG for other cases, some were saying nothing for non-interactive cases, and a few did yet other things. Make them use a new function GUC_complaint_elevel() to establish a reasonably uniform policy about how to report. There are still a few edge cases such as assign_search_path(), but it's much better than before. Per gripe from Devrim Gunduz and subsequent discussion. As noted by Alvaro, it'd be better to fold these custom messages into the standard "invalid parameter value" complaint from guc.c, perhaps as the DETAIL field. However that will require more redesign than seems prudent for 8.3. This is a relatively safe, low-impact change that we can afford to risk now.
2007-12-28 01:23:23 +01:00
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
GUC_check_errmsg("SSL is not supported by this build");
Improve consistency of error reporting in GUC assign_hook routines. Some were reporting ERROR for interactive assignments and LOG for other cases, some were saying nothing for non-interactive cases, and a few did yet other things. Make them use a new function GUC_complaint_elevel() to establish a reasonably uniform policy about how to report. There are still a few edge cases such as assign_search_path(), but it's much better than before. Per gripe from Devrim Gunduz and subsequent discussion. As noted by Alvaro, it'd be better to fold these custom messages into the standard "invalid parameter value" complaint from guc.c, perhaps as the DETAIL field. However that will require more redesign than seems prudent for 8.3. This is a relatively safe, low-impact change that we can afford to risk now.
2007-12-28 01:23:23 +01:00
return false;
}
Disallow setting ssl = on if SSL is not compiled in.
2005-12-30 01:13:50 +01:00
#endif
return true;
}
Throw error if log_statement_stats is used with confliction options for per-query stage stats.
2004-05-07 03:34:08 +02:00
static bool
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
check_stage_log_stats(bool *newval, void **extra, GucSource source)
Throw error if log_statement_stats is used with confliction options for per-query stage stats.
2004-05-07 03:34:08 +02:00
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
if (*newval && log_statement_stats)
Throw error if log_statement_stats is used with confliction options for per-query stage stats.
2004-05-07 03:34:08 +02:00
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
GUC_check_errdetail("Cannot enable parameter when \"log_statement_stats\" is true.");
return false;
Throw error if log_statement_stats is used with confliction options for per-query stage stats.
2004-05-07 03:34:08 +02:00
}
return true;
}
static bool
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
check_log_stats(bool *newval, void **extra, GucSource source)
Throw error if log_statement_stats is used with confliction options for per-query stage stats.
2004-05-07 03:34:08 +02:00
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
if (*newval &&
Code review for various recent GUC hacking. Don't elog(ERROR) when not supposed to (fixes problem with postmaster aborting due to mistaken postgresql.conf change); don't call superuser() when not inside a transaction (fixes coredump when, eg, try to set log_statement from PGOPTIONS); some message style guidelines enforcement.
2004-08-31 21:28:51 +02:00
(log_parser_stats || log_planner_stats || log_executor_stats))
Throw error if log_statement_stats is used with confliction options for per-query stage stats.
2004-05-07 03:34:08 +02:00
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
GUC_check_errdetail("Cannot enable \"log_statement_stats\" when "
"\"log_parser_stats\", \"log_planner_stats\", "
"or \"log_executor_stats\" is true.");
return false;
Throw error if log_statement_stats is used with confliction options for per-query stage stats.
2004-05-07 03:34:08 +02:00
}
return true;
}
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
static bool
check_canonical_path(char **newval, void **extra, GucSource source)
Use canonicalize_path for -D, GUC paths, and paths coming in from environment variables.
2004-07-11 23:34:04 +02:00
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
/*
pgindent run before PG 9.1 beta 1.
2011-04-10 17:42:00 +02:00
* Since canonicalize_path never enlarges the string, we can just modify
* newval in-place. But watch out for NULL, which is the default value
* for external_pid_file.
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
*/
if (*newval)
canonicalize_path(*newval);
return true;
Use canonicalize_path for -D, GUC paths, and paths coming in from environment variables.
2004-07-11 23:34:04 +02:00
}
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
static bool
check_timezone_abbreviations(char **newval, void **extra, GucSource source)
Remove hard-wired lists of timezone abbreviations in favor of providing configuration files that can be altered by a DBA. The australian_timezones GUC setting disappears, replaced by a timezone_abbreviations setting (set this to 'Australia' to get the effect of australian_timezones). The list of zone names defined by default has undergone a bit of cleanup, too. Documentation still needs some work --- in particular, should we fix Table B-4, or just get rid of it? Joachim Wieland, with some editorializing by moi.
2006-07-25 05:51:23 +02:00
{
Adjust initialization sequence for timezone_abbreviations so that it's handled just about like timezone; in particular, don't try to read anything during InitializeGUCOptions. Should solve current startup failure on Windows, and avoid wasted cycles if a nondefault setting is specified in postgresql.conf too. Possibly we need to think about a more general solution for handling 'expensive to set' GUC options.
2006-07-29 05:02:56 +02:00
/*
pgindent run before PG 9.1 beta 1.
2011-04-10 17:42:00 +02:00
* The boot_val given above for timezone_abbreviations is NULL. When we
* see this we just do nothing. If this value isn't overridden from the
* config file then pg_timezone_abbrev_initialize() will eventually
* replace it with "Default". This hack has two purposes: to avoid
* wasting cycles loading values that might soon be overridden from the
* config file, and to avoid trying to read the timezone abbrev files
pgindent run for 8.2.
2006-10-04 02:30:14 +02:00
* during InitializeGUCOptions(). The latter doesn't work in an
* EXEC_BACKEND subprocess because my_exec_path hasn't been set yet and so
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
* we can't locate PGSHAREDIR.
Adjust initialization sequence for timezone_abbreviations so that it's handled just about like timezone; in particular, don't try to read anything during InitializeGUCOptions. Should solve current startup failure on Windows, and avoid wasted cycles if a nondefault setting is specified in postgresql.conf too. Possibly we need to think about a more general solution for handling 'expensive to set' GUC options.
2006-07-29 05:02:56 +02:00
*/
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
if (*newval == NULL)
Adjust initialization sequence for timezone_abbreviations so that it's handled just about like timezone; in particular, don't try to read anything during InitializeGUCOptions. Should solve current startup failure on Windows, and avoid wasted cycles if a nondefault setting is specified in postgresql.conf too. Possibly we need to think about a more general solution for handling 'expensive to set' GUC options.
2006-07-29 05:02:56 +02:00
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
Assert(source == PGC_S_DEFAULT);
return true;
Adjust initialization sequence for timezone_abbreviations so that it's handled just about like timezone; in particular, don't try to read anything during InitializeGUCOptions. Should solve current startup failure on Windows, and avoid wasted cycles if a nondefault setting is specified in postgresql.conf too. Possibly we need to think about a more general solution for handling 'expensive to set' GUC options.
2006-07-29 05:02:56 +02:00
}
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
/* OK, load the file and produce a malloc'd TimeZoneAbbrevTable */
*extra = load_tzoffsets(*newval);
Remove hard-wired lists of timezone abbreviations in favor of providing configuration files that can be altered by a DBA. The australian_timezones GUC setting disappears, replaced by a timezone_abbreviations setting (set this to 'Australia' to get the effect of australian_timezones). The list of zone names defined by default has undergone a bit of cleanup, too. Documentation still needs some work --- in particular, should we fix Table B-4, or just get rid of it? Joachim Wieland, with some editorializing by moi.
2006-07-25 05:51:23 +02:00
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
/* tzparser.c returns NULL on failure, reporting via GUC_check_errmsg */
if (!*extra)
return false;
return true;
}
static void
assign_timezone_abbreviations(const char *newval, void *extra)
{
/* Do nothing for the boot_val default of NULL */
if (!extra)
return;
InstallTimeZoneAbbrevs((TimeZoneAbbrevTable *) extra);
Remove hard-wired lists of timezone abbreviations in favor of providing configuration files that can be altered by a DBA. The australian_timezones GUC setting disappears, replaced by a timezone_abbreviations setting (set this to 'Australia' to get the effect of australian_timezones). The list of zone names defined by default has undergone a bit of cleanup, too. Documentation still needs some work --- in particular, should we fix Table B-4, or just get rid of it? Joachim Wieland, with some editorializing by moi.
2006-07-25 05:51:23 +02:00
}
Adjust initialization sequence for timezone_abbreviations so that it's handled just about like timezone; in particular, don't try to read anything during InitializeGUCOptions. Should solve current startup failure on Windows, and avoid wasted cycles if a nondefault setting is specified in postgresql.conf too. Possibly we need to think about a more general solution for handling 'expensive to set' GUC options.
2006-07-29 05:02:56 +02:00
/*
* pg_timezone_abbrev_initialize --- set default value if not done already
*
* This is called after initial loading of postgresql.conf. If no
* timezone_abbreviations setting was found therein, select default.
Split PGC_S_DEFAULT into two values, for true boot_val vs computed default. Failure to distinguish these cases is the real cause behind the recent reports of Windows builds crashing on 'infinity'::timestamp, which was directly due to failure to establish a value of timezone_abbreviations in postmaster child processes. The postmaster had the desired value, but write_one_nondefault_variable() didn't transmit it to backends. To fix that, invent a new value PGC_S_DYNAMIC_DEFAULT, and be sure to use that or PGC_S_ENV_VAR (as appropriate) for "default" settings that are computed during initialization. (We need both because there's at least one variable that could receive a value from either source.) This commit also fixes ProcessConfigFile's failure to restore the correct default value for certain GUC variables if they are set in postgresql.conf and then removed/commented out of the file. We have to recompute and reinstall the value for any GUC variable that could have received a value from PGC_S_DYNAMIC_DEFAULT or PGC_S_ENV_VAR sources, and there were a number of oversights. (That whole thing is a crock that needs to be redesigned, but not today.) However, I intentionally didn't make it work "exactly right" for the cases of timezone and log_timezone. The exactly right behavior would involve running select_default_timezone, which we'd have to do independently in each postgres process, causing the whole database to become entirely unresponsive for as much as several seconds. That didn't seem like a good idea, especially since the variable's removal from postgresql.conf might be just an accidental edit. Instead the behavior is to adopt the previously active setting as if it were default. Note that this patch creates an ABI break for extensions that use any of the PGC_S_XXX constants; they'll need to be recompiled.
2011-05-12 01:57:38 +02:00
* If a non-default value is already installed, nothing will happen.
Simplify handling of the timezone GUC by making initdb choose the default. We were doing some amazingly complicated things in order to avoid running the very expensive identify_system_timezone() procedure during GUC initialization. But there is an obvious fix for that, which is to do it once during initdb and have initdb install the system-specific default into postgresql.conf, as it already does for most other GUC variables that need system-environment-dependent defaults. This means that the timezone (and log_timezone) settings no longer have any magic behavior in the server. Per discussion.
2011-09-09 23:59:11 +02:00
*
* This can also be called from ProcessConfigFile to establish the default
* value after a postgresql.conf entry for it is removed.
Adjust initialization sequence for timezone_abbreviations so that it's handled just about like timezone; in particular, don't try to read anything during InitializeGUCOptions. Should solve current startup failure on Windows, and avoid wasted cycles if a nondefault setting is specified in postgresql.conf too. Possibly we need to think about a more general solution for handling 'expensive to set' GUC options.
2006-07-29 05:02:56 +02:00
*/
Simplify handling of the timezone GUC by making initdb choose the default. We were doing some amazingly complicated things in order to avoid running the very expensive identify_system_timezone() procedure during GUC initialization. But there is an obvious fix for that, which is to do it once during initdb and have initdb install the system-specific default into postgresql.conf, as it already does for most other GUC variables that need system-environment-dependent defaults. This means that the timezone (and log_timezone) settings no longer have any magic behavior in the server. Per discussion.
2011-09-09 23:59:11 +02:00
static void
Adjust initialization sequence for timezone_abbreviations so that it's handled just about like timezone; in particular, don't try to read anything during InitializeGUCOptions. Should solve current startup failure on Windows, and avoid wasted cycles if a nondefault setting is specified in postgresql.conf too. Possibly we need to think about a more general solution for handling 'expensive to set' GUC options.
2006-07-29 05:02:56 +02:00
pg_timezone_abbrev_initialize(void)
{
Split PGC_S_DEFAULT into two values, for true boot_val vs computed default. Failure to distinguish these cases is the real cause behind the recent reports of Windows builds crashing on 'infinity'::timestamp, which was directly due to failure to establish a value of timezone_abbreviations in postmaster child processes. The postmaster had the desired value, but write_one_nondefault_variable() didn't transmit it to backends. To fix that, invent a new value PGC_S_DYNAMIC_DEFAULT, and be sure to use that or PGC_S_ENV_VAR (as appropriate) for "default" settings that are computed during initialization. (We need both because there's at least one variable that could receive a value from either source.) This commit also fixes ProcessConfigFile's failure to restore the correct default value for certain GUC variables if they are set in postgresql.conf and then removed/commented out of the file. We have to recompute and reinstall the value for any GUC variable that could have received a value from PGC_S_DYNAMIC_DEFAULT or PGC_S_ENV_VAR sources, and there were a number of oversights. (That whole thing is a crock that needs to be redesigned, but not today.) However, I intentionally didn't make it work "exactly right" for the cases of timezone and log_timezone. The exactly right behavior would involve running select_default_timezone, which we'd have to do independently in each postgres process, causing the whole database to become entirely unresponsive for as much as several seconds. That didn't seem like a good idea, especially since the variable's removal from postgresql.conf might be just an accidental edit. Instead the behavior is to adopt the previously active setting as if it were default. Note that this patch creates an ABI break for extensions that use any of the PGC_S_XXX constants; they'll need to be recompiled.
2011-05-12 01:57:38 +02:00
SetConfigOption("timezone_abbreviations", "Default",
PGC_POSTMASTER, PGC_S_DYNAMIC_DEFAULT);
Adjust initialization sequence for timezone_abbreviations so that it's handled just about like timezone; in particular, don't try to read anything during InitializeGUCOptions. Should solve current startup failure on Windows, and avoid wasted cycles if a nondefault setting is specified in postgresql.conf too. Possibly we need to think about a more general solution for handling 'expensive to set' GUC options.
2006-07-29 05:02:56 +02:00
}
Minor improvements in backup and recovery: - create a separate archive_mode GUC, on which archive_command is dependent - %r option in recovery.conf sends last restartpoint to recovery command - %r used in pg_standby, updated README - minor other code cleanup in pg_standby - doc on Warm Standby now mentions pg_standby and %r - log_restartpoints recovery option emits LOG message at each restartpoint - end of recovery now displays last transaction end time, as requested by Warren Little; also shown at each restartpoint - restart archiver if needed to carry away WAL files at shutdown Simon Riggs
2007-09-27 00:36:30 +02:00
static const char *
show_archive_command(void)
{
Introduce wal_level GUC to explicitly control if information needed for archival or hot standby should be WAL-logged, instead of deducing that from other options like archive_mode. This replaces recovery_connections GUC in the primary, where it now has no effect, but it's still used in the standby to enable/disable hot standby. Remove the WAL-logging of "unlogged operations", like creating an index without WAL-logging and fsyncing it at the end. Instead, we keep a copy of the wal_mode setting and the settings that affect how much shared memory a hot standby server needs to track master transactions (max_connections, max_prepared_xacts, max_locks_per_xact) in pg_control. Whenever the settings change, at server restart, write a WAL record noting the new settings and update pg_control. This allows us to notice the change in those settings in the standby at the right moment, they used to be included in checkpoint records, but that meant that a changed value was not reflected in the standby until the first checkpoint after the change. Bump PG_CONTROL_VERSION and XLOG_PAGE_MAGIC. Whack XLOG_PAGE_MAGIC back to the sequence it used to follow, before hot standby and subsequent patches changed it to 0x9003.
2010-04-28 18:10:43 +02:00
if (XLogArchivingActive())
Minor improvements in backup and recovery: - create a separate archive_mode GUC, on which archive_command is dependent - %r option in recovery.conf sends last restartpoint to recovery command - %r used in pg_standby, updated README - minor other code cleanup in pg_standby - doc on Warm Standby now mentions pg_standby and %r - log_restartpoints recovery option emits LOG message at each restartpoint - end of recovery now displays last transaction end time, as requested by Warren Little; also shown at each restartpoint - restart archiver if needed to carry away WAL files at shutdown Simon Riggs
2007-09-27 00:36:30 +02:00
return XLogArchiveCommand;
else
return "(disabled)";
}
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
static void
assign_tcp_keepalives_idle(int newval, void *extra)
Add GUC variables to control keep-alive times for idle, interval, and count. Oliver Jowett
2005-07-30 17:17:26 +02:00
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
/*
pgindent run before PG 9.1 beta 1.
2011-04-10 17:42:00 +02:00
* The kernel API provides no way to test a value without setting it; and
* once we set it we might fail to unset it. So there seems little point
* in fully implementing the check-then-assign GUC API for these
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
* variables. Instead we just do the assignment on demand. pqcomm.c
* reports any problems via elog(LOG).
*
pgindent run before PG 9.1 beta 1.
2011-04-10 17:42:00 +02:00
* This approach means that the GUC value might have little to do with the
* actual kernel value, so we use a show_hook that retrieves the kernel
* value rather than trusting GUC's copy.
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
*/
(void) pq_setkeepalivesidle(newval, MyProcPort);
Add GUC variables to control keep-alive times for idle, interval, and count. Oliver Jowett
2005-07-30 17:17:26 +02:00
}
static const char *
show_tcp_keepalives_idle(void)
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
/* See comments in assign_tcp_keepalives_idle */
Tweak TCP-keepalive code so that an invalid setting doesn't cause us to drop connections unceremoniously. Also some other marginal cleanups: don't query getsockopt() repeatedly if it fails, and avoid having the apparent definition of struct Port depend on which system headers you might have included or not. Oliver Jowett and Tom Lane.
2005-09-12 04:26:33 +02:00
static char nbuf[16];
snprintf(nbuf, sizeof(nbuf), "%d", pq_getkeepalivesidle(MyProcPort));
Add GUC variables to control keep-alive times for idle, interval, and count. Oliver Jowett
2005-07-30 17:17:26 +02:00
return nbuf;
}
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
static void
assign_tcp_keepalives_interval(int newval, void *extra)
Add GUC variables to control keep-alive times for idle, interval, and count. Oliver Jowett
2005-07-30 17:17:26 +02:00
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
/* See comments in assign_tcp_keepalives_idle */
(void) pq_setkeepalivesinterval(newval, MyProcPort);
Add GUC variables to control keep-alive times for idle, interval, and count. Oliver Jowett
2005-07-30 17:17:26 +02:00
}
static const char *
show_tcp_keepalives_interval(void)
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
/* See comments in assign_tcp_keepalives_idle */
Tweak TCP-keepalive code so that an invalid setting doesn't cause us to drop connections unceremoniously. Also some other marginal cleanups: don't query getsockopt() repeatedly if it fails, and avoid having the apparent definition of struct Port depend on which system headers you might have included or not. Oliver Jowett and Tom Lane.
2005-09-12 04:26:33 +02:00
static char nbuf[16];
snprintf(nbuf, sizeof(nbuf), "%d", pq_getkeepalivesinterval(MyProcPort));
Add GUC variables to control keep-alive times for idle, interval, and count. Oliver Jowett
2005-07-30 17:17:26 +02:00
return nbuf;
}
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
static void
assign_tcp_keepalives_count(int newval, void *extra)
Add GUC variables to control keep-alive times for idle, interval, and count. Oliver Jowett
2005-07-30 17:17:26 +02:00
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
/* See comments in assign_tcp_keepalives_idle */
(void) pq_setkeepalivescount(newval, MyProcPort);
Add GUC variables to control keep-alive times for idle, interval, and count. Oliver Jowett
2005-07-30 17:17:26 +02:00
}
static const char *
show_tcp_keepalives_count(void)
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
/* See comments in assign_tcp_keepalives_idle */
Tweak TCP-keepalive code so that an invalid setting doesn't cause us to drop connections unceremoniously. Also some other marginal cleanups: don't query getsockopt() repeatedly if it fails, and avoid having the apparent definition of struct Port depend on which system headers you might have included or not. Oliver Jowett and Tom Lane.
2005-09-12 04:26:33 +02:00
static char nbuf[16];
snprintf(nbuf, sizeof(nbuf), "%d", pq_getkeepalivescount(MyProcPort));
Add GUC variables to control keep-alive times for idle, interval, and count. Oliver Jowett
2005-07-30 17:17:26 +02:00
return nbuf;
}
Adjust configuration-files GUC behavior as per my recent proposal. The vars are renamed to data_directory, config_file, hba_file, and ident_file, and are guaranteed to be set to accurate absolute paths during postmaster startup. This commit does not yet do anything about hiding path values from non-superusers.
2004-10-10 01:13:22 +02:00
Add a multi-worker capability to autovacuum. This allows multiple worker processes to be running simultaneously. Also, now autovacuum processes do not count towards the max_connections limit; they are counted separately from regular processes, and are limited by the new GUC variable autovacuum_max_workers. The launcher now has intelligence to launch workers on each database every autovacuum_naptime seconds, limited only on the max amount of worker slots available. Also, the global worker I/O utilization is limited by the vacuum cost-based delay feature. Workers are "balanced" so that the total I/O consumption does not exceed the established limit. This part of the patch was contributed by ITAGAKI Takahiro. Per discussion.
2007-04-16 20:30:04 +02:00
static bool
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
check_maxconnections(int *newval, void **extra, GucSource source)
Add a multi-worker capability to autovacuum. This allows multiple worker processes to be running simultaneously. Also, now autovacuum processes do not count towards the max_connections limit; they are counted separately from regular processes, and are limited by the new GUC variable autovacuum_max_workers. The launcher now has intelligence to launch workers on each database every autovacuum_naptime seconds, limited only on the max amount of worker slots available. Also, the global worker I/O utilization is limited by the vacuum cost-based delay feature. Workers are "balanced" so that the total I/O consumption does not exceed the established limit. This part of the patch was contributed by ITAGAKI Takahiro. Per discussion.
2007-04-16 20:30:04 +02:00
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
if (*newval + autovacuum_max_workers + 1 > MAX_BACKENDS)
Fix thinko in multi-autovac-workers code: validity checks made by GUC assign hooks are supposed to be made whether doit is true or not.
2007-08-08 18:00:46 +02:00
return false;
Add a multi-worker capability to autovacuum. This allows multiple worker processes to be running simultaneously. Also, now autovacuum processes do not count towards the max_connections limit; they are counted separately from regular processes, and are limited by the new GUC variable autovacuum_max_workers. The launcher now has intelligence to launch workers on each database every autovacuum_naptime seconds, limited only on the max amount of worker slots available. Also, the global worker I/O utilization is limited by the vacuum cost-based delay feature. Workers are "balanced" so that the total I/O consumption does not exceed the established limit. This part of the patch was contributed by ITAGAKI Takahiro. Per discussion.
2007-04-16 20:30:04 +02:00
return true;
}
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
static void
assign_maxconnections(int newval, void *extra)
{
MaxBackends = newval + autovacuum_max_workers + 1;
}
Add a multi-worker capability to autovacuum. This allows multiple worker processes to be running simultaneously. Also, now autovacuum processes do not count towards the max_connections limit; they are counted separately from regular processes, and are limited by the new GUC variable autovacuum_max_workers. The launcher now has intelligence to launch workers on each database every autovacuum_naptime seconds, limited only on the max amount of worker slots available. Also, the global worker I/O utilization is limited by the vacuum cost-based delay feature. Workers are "balanced" so that the total I/O consumption does not exceed the established limit. This part of the patch was contributed by ITAGAKI Takahiro. Per discussion.
2007-04-16 20:30:04 +02:00
static bool
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
check_autovacuum_max_workers(int *newval, void **extra, GucSource source)
Add a multi-worker capability to autovacuum. This allows multiple worker processes to be running simultaneously. Also, now autovacuum processes do not count towards the max_connections limit; they are counted separately from regular processes, and are limited by the new GUC variable autovacuum_max_workers. The launcher now has intelligence to launch workers on each database every autovacuum_naptime seconds, limited only on the max amount of worker slots available. Also, the global worker I/O utilization is limited by the vacuum cost-based delay feature. Workers are "balanced" so that the total I/O consumption does not exceed the established limit. This part of the patch was contributed by ITAGAKI Takahiro. Per discussion.
2007-04-16 20:30:04 +02:00
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
if (MaxConnections + *newval + 1 > MAX_BACKENDS)
Fix thinko in multi-autovac-workers code: validity checks made by GUC assign hooks are supposed to be made whether doit is true or not.
2007-08-08 18:00:46 +02:00
return false;
Add a multi-worker capability to autovacuum. This allows multiple worker processes to be running simultaneously. Also, now autovacuum processes do not count towards the max_connections limit; they are counted separately from regular processes, and are limited by the new GUC variable autovacuum_max_workers. The launcher now has intelligence to launch workers on each database every autovacuum_naptime seconds, limited only on the max amount of worker slots available. Also, the global worker I/O utilization is limited by the vacuum cost-based delay feature. Workers are "balanced" so that the total I/O consumption does not exceed the established limit. This part of the patch was contributed by ITAGAKI Takahiro. Per discussion.
2007-04-16 20:30:04 +02:00
return true;
}
Tweak TCP-keepalive code so that an invalid setting doesn't cause us to drop connections unceremoniously. Also some other marginal cleanups: don't query getsockopt() repeatedly if it fails, and avoid having the apparent definition of struct Port depend on which system headers you might have included or not. Oliver Jowett and Tom Lane.
2005-09-12 04:26:33 +02:00
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
static void
assign_autovacuum_max_workers(int newval, void *extra)
{
MaxBackends = MaxConnections + newval + 1;
}
Implement prefetching via posix_fadvise() for bitmap index scans. A new GUC variable effective_io_concurrency controls how many concurrent block prefetch requests will be issued. (The best way to handle this for plain index scans is still under debate, so that part is not applied yet --- tgl) Greg Stark
2009-01-12 06:10:45 +01:00
static bool
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
check_effective_io_concurrency(int *newval, void **extra, GucSource source)
Implement prefetching via posix_fadvise() for bitmap index scans. A new GUC variable effective_io_concurrency controls how many concurrent block prefetch requests will be issued. (The best way to handle this for plain index scans is still under debate, so that part is not applied yet --- tgl) Greg Stark
2009-01-12 06:10:45 +01:00
{
#ifdef USE_PREFETCH
double new_prefetch_pages = 0.0;
int i;
/*----------
* The user-visible GUC parameter is the number of drives (spindles),
* which we need to translate to a number-of-pages-to-prefetch target.
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
* The target value is stashed in *extra and then assigned to the actual
* variable by assign_effective_io_concurrency.
Implement prefetching via posix_fadvise() for bitmap index scans. A new GUC variable effective_io_concurrency controls how many concurrent block prefetch requests will be issued. (The best way to handle this for plain index scans is still under debate, so that part is not applied yet --- tgl) Greg Stark
2009-01-12 06:10:45 +01:00
*
* The expected number of prefetch pages needed to keep N drives busy is:
*
* drives | I/O requests
* -------+----------------
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
* 1 | 1
* 2 | 2/1 + 2/2 = 3
* 3 | 3/1 + 3/2 + 3/3 = 5 1/2
* 4 | 4/1 + 4/2 + 4/3 + 4/4 = 8 1/3
* n | n * H(n)
Implement prefetching via posix_fadvise() for bitmap index scans. A new GUC variable effective_io_concurrency controls how many concurrent block prefetch requests will be issued. (The best way to handle this for plain index scans is still under debate, so that part is not applied yet --- tgl) Greg Stark
2009-01-12 06:10:45 +01:00
*
* This is called the "coupon collector problem" and H(n) is called the
* harmonic series. This could be approximated by n * ln(n), but for
* reasonable numbers of drives we might as well just compute the series.
*
* Alternatively we could set the target to the number of pages necessary
* so that the expected number of active spindles is some arbitrary
* percentage of the total. This sounds the same but is actually slightly
* different. The result ends up being ln(1-P)/ln((n-1)/n) where P is
* that desired fraction.
*
* Experimental results show that both of these formulas aren't aggressive
* enough, but we don't really have any better proposals.
*
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
* Note that if *newval = 0 (disabled), we must set target = 0.
Implement prefetching via posix_fadvise() for bitmap index scans. A new GUC variable effective_io_concurrency controls how many concurrent block prefetch requests will be issued. (The best way to handle this for plain index scans is still under debate, so that part is not applied yet --- tgl) Greg Stark
2009-01-12 06:10:45 +01:00
*----------
*/
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
for (i = 1; i <= *newval; i++)
new_prefetch_pages += (double) *newval / (double) i;
Implement prefetching via posix_fadvise() for bitmap index scans. A new GUC variable effective_io_concurrency controls how many concurrent block prefetch requests will be issued. (The best way to handle this for plain index scans is still under debate, so that part is not applied yet --- tgl) Greg Stark
2009-01-12 06:10:45 +01:00
/* This range check shouldn't fail, but let's be paranoid */
if (new_prefetch_pages >= 0.0 && new_prefetch_pages < (double) INT_MAX)
{
pgindent run before PG 9.1 beta 1.
2011-04-10 17:42:00 +02:00
int *myextra = (int *) guc_malloc(ERROR, sizeof(int));
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
*myextra = (int) rint(new_prefetch_pages);
*extra = (void *) myextra;
Implement prefetching via posix_fadvise() for bitmap index scans. A new GUC variable effective_io_concurrency controls how many concurrent block prefetch requests will be issued. (The best way to handle this for plain index scans is still under debate, so that part is not applied yet --- tgl) Greg Stark
2009-01-12 06:10:45 +01:00
return true;
}
else
return false;
#else
return true;
8.4 pgindent run, with new combined Linux/FreeBSD/MinGW typedef list provided by Andrew.
2009-06-11 16:49:15 +02:00
#endif /* USE_PREFETCH */
Implement prefetching via posix_fadvise() for bitmap index scans. A new GUC variable effective_io_concurrency controls how many concurrent block prefetch requests will be issued. (The best way to handle this for plain index scans is still under debate, so that part is not applied yet --- tgl) Greg Stark
2009-01-12 06:10:45 +01:00
}
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
static void
assign_effective_io_concurrency(int newval, void *extra)
Make the temporary directory for pgstat files configurable by the GUC variable stats_temp_directory, instead of requiring the admin to mount/symlink the pg_stat_tmp directory manually. For now the config variable is PGC_POSTMASTER. Room for further improvment that would allow it to be changed on-the-fly.
2008-08-15 10:37:41 +02:00
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
#ifdef USE_PREFETCH
target_prefetch_pages = *((int *) extra);
#endif /* USE_PREFETCH */
}
Fix assign_pgstat_temp_directory() to ensure the directory path is canonicalized. Avoid the need to elog(FATAL) on out-of-memory.
2009-05-03 22:09:54 +02:00
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
static void
assign_pgstat_temp_directory(const char *newval, void *extra)
{
/* check_canonical_path already canonicalized newval for us */
char *tname;
char *fname;
Make the temporary directory for pgstat files configurable by the GUC variable stats_temp_directory, instead of requiring the admin to mount/symlink the pg_stat_tmp directory manually. For now the config variable is PGC_POSTMASTER. Room for further improvment that would allow it to be changed on-the-fly.
2008-08-15 10:37:41 +02:00
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
tname = guc_malloc(ERROR, strlen(newval) + 12); /* /pgstat.tmp */
sprintf(tname, "%s/pgstat.tmp", newval);
fname = guc_malloc(ERROR, strlen(newval) + 13); /* /pgstat.stat */
sprintf(fname, "%s/pgstat.stat", newval);
if (pgstat_stat_tmpname)
free(pgstat_stat_tmpname);
pgstat_stat_tmpname = tname;
if (pgstat_stat_filename)
free(pgstat_stat_filename);
pgstat_stat_filename = fname;
Make the temporary directory for pgstat files configurable by the GUC variable stats_temp_directory, instead of requiring the admin to mount/symlink the pg_stat_tmp directory manually. For now the config variable is PGC_POSTMASTER. Room for further improvment that would allow it to be changed on-the-fly.
2008-08-15 10:37:41 +02:00
}
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
static bool
check_application_name(char **newval, void **extra, GucSource source)
Add support for an application_name parameter, which is displayed in pg_stat_activity and recorded in log entries. Dave Page, reviewed by Andres Freund
2009-11-29 00:38:08 +01:00
{
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
/* Only allow clean ASCII chars in the application name */
char *p;
Add support for an application_name parameter, which is displayed in pg_stat_activity and recorded in log entries. Dave Page, reviewed by Andres Freund
2009-11-29 00:38:08 +01:00
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
for (p = *newval; *p; p++)
{
if (*p < 32 || *p > 126)
*p = '?';
}
Add support for an application_name parameter, which is displayed in pg_stat_activity and recorded in log entries. Dave Page, reviewed by Andres Freund
2009-11-29 00:38:08 +01:00
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
return true;
}
Add support for an application_name parameter, which is displayed in pg_stat_activity and recorded in log entries. Dave Page, reviewed by Andres Freund
2009-11-29 00:38:08 +01:00
Revise the API for GUC variable assign hooks. The previous functions of assign hooks are now split between check hooks and assign hooks, where the former can fail but the latter shouldn't. Aside from being conceptually clearer, this approach exposes the "canonicalized" form of the variable value to guc.c without having to do an actual assignment. And that lets us fix the problem recently noted by Bernd Helmle that the auto-tune patch for wal_buffers resulted in bogus log messages about "parameter "wal_buffers" cannot be changed without restarting the server". There may be some speed advantage too, because this design lets hook functions avoid re-parsing variable values when restoring a previous state after a rollback (they can store a pre-parsed representation of the value instead). This patch also resolves a longstanding annoyance about custom error messages from variable assign hooks: they should modify, not appear separately from, guc.c's own message about "invalid parameter value".
2011-04-07 06:11:01 +02:00
static void
assign_application_name(const char *newval, void *extra)
{
/* Update the pg_stat_activity view */
pgstat_report_appname(newval);
Add support for an application_name parameter, which is displayed in pg_stat_activity and recorded in log entries. Dave Page, reviewed by Andres Freund
2009-11-29 00:38:08 +01:00
}
Add a log_file_mode GUC that allows control of the file permissions set on log files created by the syslogger process. In passing, make unix_file_permissions display its value in octal, same as log_file_mode now does. Martin Pihlak
2010-07-17 00:25:51 +02:00
static const char *
show_unix_socket_permissions(void)
{
static char buf[8];
snprintf(buf, sizeof(buf), "%04o", Unix_socket_permissions);
return buf;
}
static const char *
show_log_file_mode(void)
{
static char buf[8];
snprintf(buf, sizeof(buf), "%04o", Log_file_mode);
return buf;
}
Arrange to compile flex output files as inclusions into other files (usually bison output files), not as standalone files. This hack works around flex's insistence on including <stdio.h> before we are able to include postgres.h; postgres.h will already be read before the compiler starts to read the flex output file. Needed for largefile support on some platforms.
2002-11-01 23:52:34 +01:00
#include "guc-file.c"