Commit Graph

32881 Commits

Author SHA1 Message Date
Heikki Linnakangas
9f0d2bdc88 Don't set reachedMinRecoveryPoint during crash recovery. In crash recovery,
we don't reach consistency before replaying all of the WAL. Rename the
variable to reachedConsistency, to make its intention clearer.

In master, that was an active bug because of the recent patch to
immediately PANIC if a reference to a missing page is found in WAL after
reaching consistency, as Tom Lane's test case demonstrated. In 9.1 and 9.0,
the only consequence was a misleading "consistent recovery state reached at
%X/%X" message in the log at the beginning of crash recovery (the database
is not consistent at that point yet). In 8.4, the log message was not
printed in crash recovery, even though there was a similar
reachedMinRecoveryPoint local variable that was also set early. So,
backpatch to 9.1 and 9.0.
2011-12-09 15:21:12 +02:00
Heikki Linnakangas
5d8a894e30 Cancel running query if it is detected that the connection to the client is
lost. The only way we detect that at the moment is when write() fails when
we try to write to the socket.

Florian Pflug with small changes by me, reviewed by Greg Jaskiewicz.
2011-12-09 14:21:36 +02:00
Peter Eisentraut
d5f23af6bf Add const qualifiers to node inspection functions
Thomas Munro
2011-12-07 21:46:56 +02:00
Tom Lane
0d0ec527af Fix corner cases in readlink() usage.
Make sure all calls are protected by HAVE_READLINK, and get the buffer
overflow tests right.  Be a bit more paranoid about string length in
_tarWriteHeader(), too.
2011-12-07 13:34:13 -05:00
Magnus Hagander
0d9b09282f Better error reporting if the link target is too long
This situation won't set errno, so using %m will give an incorrect
error message.
2011-12-07 12:19:20 +01:00
Magnus Hagander
1f422db663 Avoid using readlink() on platforms that don't support it
We don't have any such platforms now, but might in the future.

Also, detect cases when a tablespace symlink points to a path that
is longer than we can handle, and give a warning.
2011-12-07 12:09:05 +01:00
Magnus Hagander
16d8e594ac Remove spclocation field from pg_tablespace
Instead, add a function pg_tablespace_location(oid) used to return
the same information, and do this by reading the symbolic link.

Doing it this way makes it possible to relocate a tablespace when the
database is down by simply changing the symbolic link.
2011-12-07 10:37:33 +01:00
Tom Lane
c6e3ac11b6 Create a "sort support" interface API for faster sorting.
This patch creates an API whereby a btree index opclass can optionally
provide non-SQL-callable support functions for sorting.  In the initial
patch, we only use this to provide a directly-callable comparator function,
which can be invoked with a bit less overhead than the traditional
SQL-callable comparator.  While that should be of value in itself, the real
reason for doing this is to provide a datatype-extensible framework for
more aggressive optimizations, as in Peter Geoghegan's recent work.

Robert Haas and Tom Lane
2011-12-07 00:19:39 -05:00
Robert Haas
d2a662182e Typo fixes for commit 2ad36c4e44.
Noted during post-commit review by by Noah Misch.
2011-12-06 15:50:02 -05:00
Robert Haas
68281e0054 Make command-line tools smarter about finding a DB to connect to.
If unable to connect to "postgres", try "template1".  This allows things to
work more smoothly in the case where the postgres database has been
dropped.  And just in case that's not good enough, also allow the user to
specify a maintenance database to be used for the initial connection, to
cover the case where neither postgres nor template1 is suitable.
2011-12-06 08:48:15 -05:00
Magnus Hagander
6ef4ae1d4e Add missing documentation for function pg_stat_get_wal_senders()
Euler Taveira de Oliveira
2011-12-06 11:02:02 +01:00
Bruce Momjian
0e8f6bf0e7 In pg_upgrade, allow tables using regclass to be upgraded because we
preserve pg_class oids since PG 9.0.
2011-12-05 16:45:19 -05:00
Tom Lane
ff68b256a5 Remove troublesome Asserts in cost_mergejoin().
While logically correct, these two Asserts could fail depending on the
vagaries of floating-point arithmetic.  In particular, on machines with
floating-point registers wider than standard "double" values, it was
possible for the compiler to compare a rounded-to-double value already
stored in memory with an unrounded long double value still in a register.
Given the preceding checks, these assertions aren't adding much, so let's
just get rid of them rather than try to find a compiler-proof fix.
Per report from Pavel Stehule.

Given the lack of previous complaints, and the fact that only developers
would be likely to trip over it, I'm only going to change this in HEAD,
even though the code has been like this for a long time.
2011-12-05 15:50:06 -05:00
Peter Eisentraut
89e850e6fd plpython: Add SPI cursor support
Add a function plpy.cursor that is similar to plpy.execute but uses an
SPI cursor to avoid fetching the entire result set into memory.

Jan Urbański, reviewed by Steve Singer
2011-12-05 19:52:15 +02:00
Andrew Dunstan
e6d9e2106f Add a \setenv command to psql.
This can be used to set (or unset) environment variables that will
affect programs called by psql (such as the PAGER), probably most
usefully in a .psqlrc file.

Andrew Dunstan, reviewed by Josh Kupershmidt.
2011-12-04 11:43:38 -05:00
Magnus Hagander
eb06416ba4 Remove incorrect instructions to run CREATE FOREIGN DATA WRAPPER
The CREATE EXTENSION step does this automatically. Doing it again will
cause an error.
2011-12-04 16:59:24 +01:00
Michael Meskes
f2ae9f9c30 Applied another patch by Zoltan to fix memory alignement issues in ecpg's sqlda
code.
2011-12-04 04:43:58 +01:00
Magnus Hagander
5b40677986 Treat ENOTDIR as ENOENT when looking for client certificate file
This makes it possible to use a libpq app with home directory set
to /dev/null, for example - treating it the same as if the file
doesn't exist (which it doesn't).

Per bug #6302, reported by Diego Elio Petteno
2011-12-03 15:05:24 +01:00
Bruce Momjian
155e56bd97 Update documentation suggestions for debugging the backend.
Tom Lane, with minor adjustments by me.
2011-12-02 17:09:56 -05:00
Peter Eisentraut
b6f9834a9e psql: Make temporary editor files have .sql extension
This gives editors a better chance to treat these files as the SQL
files that they are.
2011-12-02 23:38:33 +02:00
Tom Lane
0de93a9c69 Add some weasel wording about threaded usage of PGresults.
PGresults used to be read-only from the application's viewpoint, but now
that we've exposed various functions that allow modification of a PGresult,
that sweeping statement is no longer accurate.  Noted by Dmitriy Igrishin.
2011-12-02 11:33:53 -05:00
Heikki Linnakangas
1e616f6391 During recovery, if we reach consistent state and still have entries in the
invalid-page hash table, PANIC immediately. Immediate PANIC is much better
than waiting for end-of-recovery, which is what we did before, because the
end-of-recovery might not come until months later if this is a standby
server.

Also refrain from creating a restartpoint if there are invalid-page entries
in the hash table. Restarting recovery from such a restartpoint would not
see the invalid references, and wouldn't be able to cross-check them when
consistency is reached. That wouldn't matter when things are going smoothly,
but the more sanity checks you have the better.

Fujii Masao
2011-12-02 10:49:54 +02:00
Tom Lane
15a5006aac Clarify documentation about SQL:2008 variant of LIMIT/OFFSET syntax.
The point that you need parentheses for non-constant expressions apparently
needs to be brought out a bit more clearly, per bug #6315.
2011-12-01 16:38:59 -05:00
Robert Haas
3b86b4653c Make pgcompinclude/pgrminclude less tied to Bruce's personal machine.
Not everyone has /pg linked to the src subdirectory of their PostgreSQL
tree.  Also, cc isn't the way to invoke the compiler everywhere.
2011-12-01 14:44:32 -05:00
Tom Lane
65d9aedb1b Fix getTypeIOParam to support type record[].
Since record[] uses array_in, it needs to have its element type passed
as typioparam.  In HEAD and 9.1, this fix essentially reverts commit
9bc933b212, which was a hack that is no
longer needed since domains don't set their typelem anymore.  Before
that, adjust the logic so that only domains are excluded from being
treated like arrays, rather than assuming that only base types should
be included.  Add a regression test to demonstrate the need for this.
Per report from Maxim Boguk.

Back-patch to 8.4, where type record[] was added.
2011-12-01 12:44:16 -05:00
Bruce Momjian
1be6f93792 Add file-fdw documentation example.
Josh Berkus
2011-12-01 09:33:59 -05:00
Bruce Momjian
ebbcba75b4 Add documentation mention that 7 != NULL also returns NULL. 2011-12-01 09:17:51 -05:00
Tom Lane
bc9306f4c5 Update information about configuring SysV IPC parameters on NetBSD.
Per Emmanuel Kasper, sysctl works fine as of NetBSD 5.0.
2011-11-30 20:55:00 -05:00
Tom Lane
56d609c3b1 Draft release notes for 9.1.2, 9.0.6, 8.4.10, 8.3.17, 8.2.23. 2011-11-30 19:34:47 -05:00
Tom Lane
895d328a16 Update time zone data files to tzdata release 2011n.
DST law changes in Brazil, Cuba, Fiji, Palestine, Russia, Samoa.
Historical corrections for Alaska and British East Africa.
2011-11-30 11:48:05 -05:00
Robert Haas
2ad36c4e44 Improve table locking behavior in the face of current DDL.
In the previous coding, callers were faced with an awkward choice:
look up the name, do permissions checks, and then lock the table; or
look up the name, lock the table, and then do permissions checks.
The first choice was wrong because the results of the name lookup
and permissions checks might be out-of-date by the time the table
lock was acquired, while the second allowed a user with no privileges
to interfere with access to a table by users who do have privileges
(e.g. if a malicious backend queues up for an AccessExclusiveLock on
a table on which AccessShareLock is already held, further attempts
to access the table will be blocked until the AccessExclusiveLock
is obtained and the malicious backend's transaction rolls back).

To fix, allow callers of RangeVarGetRelid() to pass a callback which
gets executed after performing the name lookup but before acquiring
the relation lock.  If the name lookup is retried (because
invalidation messages are received), the callback will be re-executed
as well, so we get the best of both worlds.  RangeVarGetRelid() is
renamed to RangeVarGetRelidExtended(); callers not wishing to supply
a callback can continue to invoke it as RangeVarGetRelid(), which is
now a macro.  Since the only one caller that uses nowait = true now
passes a callback anyway, the RangeVarGetRelid() macro defaults nowait
as well.  The callback can also be used for supplemental locking - for
example, REINDEX INDEX needs to acquire the table lock before the index
lock to reduce deadlock possibilities.

There's a lot more work to be done here to fix all the cases where this
can be a problem, but this commit provides the general infrastructure
and fixes the following specific cases: REINDEX INDEX, REINDEX TABLE,
LOCK TABLE, and and DROP TABLE/INDEX/SEQUENCE/VIEW/FOREIGN TABLE.

Per discussion with Noah Misch and Alvaro Herrera.
2011-11-30 10:27:00 -05:00
Tom Lane
a87ebace19 Tweak previous patch to ensure edata->filename always gets initialized.
On a platform that isn't supplying __FILE__, previous coding would either
crash or give a stale result for the filename string.  Not sure how likely
that is, but the original code catered for it, so let's keep doing so.
2011-11-30 00:37:06 -05:00
Peter Eisentraut
dd136052bc Strip file names reported in error messages in vpath builds
In vpath builds, the __FILE__ macro that is used in verbose error
reports contains the full absolute file name, which makes the error
messages excessively verbose.  So keep only the base name, thus
matching the behavior of non-vpath builds.
2011-11-30 06:56:18 +02:00
Tom Lane
be2f909200 Remove duplicate definition of 'progname'.
Per buildfarm.
2011-11-29 23:19:46 -05:00
Tom Lane
9c6aa8cca0 Move EXTRA_CLEAN to where it actually works. 2011-11-29 23:14:16 -05:00
Tom Lane
73d1bfd0b5 Prevent autovacuum transactions from running in serializable mode.
Force the transaction isolation level to READ COMMITTED in autovacuum
worker and launcher processes.  There is no benefit to using a higher
isolation level, and doing so could result in delaying foreground
transactions (or maybe even causing unnecessary serialization failures?).
Noted by Dan Ports.

Also, make sure we disable zero_damaged_pages and statement_timeout in
the autovac launcher, not only workers.  Now that the launcher can run
transactions, these settings could affect its behavior, and it seems
like the same arguments apply to the launcher as the workers.
2011-11-29 22:40:18 -05:00
Bruce Momjian
99f5e47df5 Change installation docs to mention general debugging options. 2011-11-29 22:35:48 -05:00
Tom Lane
0195e5c4ab Clean up after recent pg_dump patches.
Fix entirely broken handling of va_list printing routines, update some
out-of-date comments, fix some bogus inclusion orders, fix NLS declarations,
fix missed realloc calls.
2011-11-29 20:41:54 -05:00
Bruce Momjian
2ff36abeec In docs, suggest "-O0 -g" only if using a debugger. 2011-11-29 19:12:38 -05:00
Tom Lane
f225e4bc54 When a row fails a not-null constraint, show row's contents in errdetail.
Simple extension of previous patch for CHECK constraints.
2011-11-29 18:29:18 -05:00
Bruce Momjian
8b08deb0d1 Simplify the pg_dump/pg_restore error reporting macros, and allow
pg_dumpall to use the same memory allocation functions as the others.
2011-11-29 16:34:45 -05:00
Bruce Momjian
b60f37bf44 Suggest configure options for server developers.
Greg Smith
2011-11-29 16:32:38 -05:00
Tom Lane
f1e13001b2 When a row fails a CHECK constraint, show row's contents in errdetail.
This should make it easier to identify which row is problematic when an
insert or update is processing many rows.

The formatting is similar to that for unique-index violation messages,
except that we limit field widths to 64 bytes since otherwise the message
could get unreasonably long.  (In particular, there's currently no attempt
to quote or escape field values that contain commas etc.)

Jan Kundrát, reviewed by Royce Ausburn, somewhat rewritten by me.
2011-11-29 15:02:49 -05:00
Peter Eisentraut
9922fc5f9f pg_dump: Add gettext plural support to error message 2011-11-29 19:49:02 +02:00
Peter Eisentraut
b258ceb3ad pg_upgrade: Disable installcheck
Disabled for now because some build farm members with low resources
are not prepared to handle it.
2011-11-29 06:57:09 +02:00
Peter Eisentraut
80ac853f05 python.m4: Remove useless "import string" calls
They have been unneeded since the use of the string module has been
removed in a65ed83f8a.
2011-11-29 06:50:11 +02:00
Peter Eisentraut
04e5cb629b plpython: Fix sed expression in python3 build
The old expression sed 's,$(srcdir),python3,' would normally resolve
as sed 's,.,python3,', which is not really what we wanted.  While it
doesn't actually break anything right now, it's still wrong, so put in
a bit more work to make it more robust.
2011-11-29 06:39:05 +02:00
Peter Eisentraut
cc534f8b3c pg_dump: Fix clean rule after file renaming shuffle 2011-11-29 06:30:45 +02:00
Bruce Momjian
7260a0d00a Document that perl needs to be indented during the pgindent run. 2011-11-28 21:56:58 -05:00
Tom Lane
43dc4adf58 Make some minor formatting improvements to what pgindent did.
Moving the code two full tab stops to the right requires rethinking of
cosmetic code layout choices, which pgindent isn't really able to do for
us.  Whitespace and comment adjustments only, no code changes.
2011-11-28 20:19:29 -05:00