mirror of
https://git.postgresql.org/git/postgresql.git
synced 2024-09-09 09:59:24 +02:00
9f2ee8f287
a lot of strange behaviors that occurred in join cases. We now identify the "current" row for every joined relation in UPDATE, DELETE, and SELECT FOR UPDATE/SHARE queries. If an EvalPlanQual recheck is necessary, we jam the appropriate row into each scan node in the rechecking plan, forcing it to emit only that one row. The former behavior could rescan the whole of each joined relation for each recheck, which was terrible for performance, and what's much worse could result in duplicated output tuples. Also, the original implementation of EvalPlanQual could not re-use the recheck execution tree --- it had to go through a full executor init and shutdown for every row to be tested. To avoid this overhead, I've associated a special runtime Param with each LockRows or ModifyTable plan node, and arranged to make every scan node below such a node depend on that Param. Thus, by signaling a change in that Param, the EPQ machinery can just rescan the already-built test plan. This patch also adds a prohibition on set-returning functions in the targetlist of SELECT FOR UPDATE/SHARE. This is needed to avoid the duplicate-output-tuple problem. It seems fairly reasonable since the other restrictions on SELECT FOR UPDATE are meant to ensure that there is a unique correspondence between source tuples and result tuples, which an output SRF destroys as much as anything else does.
59 lines
2.6 KiB
C
59 lines
2.6 KiB
C
/*-------------------------------------------------------------------------
|
|
*
|
|
* catversion.h
|
|
* "Catalog version number" for PostgreSQL.
|
|
*
|
|
* The catalog version number is used to flag incompatible changes in
|
|
* the PostgreSQL system catalogs. Whenever anyone changes the format of
|
|
* a system catalog relation, or adds, deletes, or modifies standard
|
|
* catalog entries in such a way that an updated backend wouldn't work
|
|
* with an old database (or vice versa), the catalog version number
|
|
* should be changed. The version number stored in pg_control by initdb
|
|
* is checked against the version number compiled into the backend at
|
|
* startup time, so that a backend can refuse to run in an incompatible
|
|
* database.
|
|
*
|
|
* The point of this feature is to provide a finer grain of compatibility
|
|
* checking than is possible from looking at the major version number
|
|
* stored in PG_VERSION. It shouldn't matter to end users, but during
|
|
* development cycles we usually make quite a few incompatible changes
|
|
* to the contents of the system catalogs, and we don't want to bump the
|
|
* major version number for each one. What we can do instead is bump
|
|
* this internal version number. This should save some grief for
|
|
* developers who might otherwise waste time tracking down "bugs" that
|
|
* are really just code-vs-database incompatibilities.
|
|
*
|
|
* The rule for developers is: if you commit a change that requires
|
|
* an initdb, you should update the catalog version number (as well as
|
|
* notifying the pghackers mailing list, which has been the informal
|
|
* practice for a long time).
|
|
*
|
|
* The catalog version number is placed here since modifying files in
|
|
* include/catalog is the most common kind of initdb-forcing change.
|
|
* But it could be used to protect any kind of incompatible change in
|
|
* database contents or layout, such as altering tuple headers.
|
|
*
|
|
*
|
|
* Portions Copyright (c) 1996-2009, PostgreSQL Global Development Group
|
|
* Portions Copyright (c) 1994, Regents of the University of California
|
|
*
|
|
* $PostgreSQL: pgsql/src/include/catalog/catversion.h,v 1.547 2009/10/26 02:26:41 tgl Exp $
|
|
*
|
|
*-------------------------------------------------------------------------
|
|
*/
|
|
#ifndef CATVERSION_H
|
|
#define CATVERSION_H
|
|
|
|
/*
|
|
* We could use anything we wanted for version numbers, but I recommend
|
|
* following the "YYYYMMDDN" style often used for DNS zone serial numbers.
|
|
* YYYYMMDD are the date of the change, and N is the number of the change
|
|
* on that day. (Hopefully we'll never commit ten independent sets of
|
|
* catalog changes on the same day...)
|
|
*/
|
|
|
|
/* yyyymmddN */
|
|
#define CATALOG_VERSION_NO 200910251
|
|
|
|
#endif
|