OSDN Git Service

pg-rex/syncrep.git
22 years agoMove disk usage section into its own section.
Bruce Momjian [Thu, 13 Jun 2002 05:15:22 +0000 (05:15 +0000)]
Move disk usage section into its own section.

22 years agoAdd section on showing disk usage.
Bruce Momjian [Thu, 13 Jun 2002 04:36:50 +0000 (04:36 +0000)]
Add section on showing disk usage.

22 years agoFurther tweaks to support display of sort keys in EXPLAIN --- initial
Tom Lane [Thu, 13 Jun 2002 03:40:49 +0000 (03:40 +0000)]
Further tweaks to support display of sort keys in EXPLAIN --- initial
implementation didn't work for Sort nodes associated with Append plans.

22 years agoRepair for bug #691 --- CREATE TABLE AS column aliases fail to be
Tom Lane [Thu, 13 Jun 2002 02:04:46 +0000 (02:04 +0000)]
Repair for bug #691 --- CREATE TABLE AS column aliases fail to be
applied when the select is a UNION (or other set-operation).

An alternative route to a fix would be to leave analyze.c alone and
change plan_set_operations in prepunion.c to take column names from
the topmost targetlist.  But I am not sure that would work in all
cases.  This patch seems the minimum-risk fix.

22 years agoMention vacuum for relpages.
Bruce Momjian [Thu, 13 Jun 2002 00:54:05 +0000 (00:54 +0000)]
Mention vacuum for relpages.

22 years agoAdd script to show disk space per db.
Bruce Momjian [Thu, 13 Jun 2002 00:21:52 +0000 (00:21 +0000)]
Add script to show disk space per db.

22 years agoImprove query.
Bruce Momjian [Wed, 12 Jun 2002 23:42:48 +0000 (23:42 +0000)]
Improve query.

22 years agoAdd mention of query showing pages used.
Bruce Momjian [Wed, 12 Jun 2002 23:41:50 +0000 (23:41 +0000)]
Add mention of query showing pages used.

22 years agoAdd to ecpg:
Bruce Momjian [Wed, 12 Jun 2002 23:39:49 +0000 (23:39 +0000)]
Add to ecpg:

>  o Allow multi-threaded use of SQLCA

22 years agoCleanup.
Bruce Momjian [Wed, 12 Jun 2002 21:20:59 +0000 (21:20 +0000)]
Cleanup.

22 years agoAdd -q option to oid2name. Add sample session to README.
Bruce Momjian [Wed, 12 Jun 2002 21:09:09 +0000 (21:09 +0000)]
Add -q option to oid2name.  Add sample session to README.

22 years agoApplied Lee Kindness' patch to fix one of memory allocation with floating point numbers.
Michael Meskes [Wed, 12 Jun 2002 12:06:53 +0000 (12:06 +0000)]
Applied Lee Kindness' patch to fix one of memory allocation with floating point numbers.

22 years ago1) Fix a bug *double error message*.
Hiroshi Inoue [Wed, 12 Jun 2002 02:27:44 +0000 (02:27 +0000)]
1) Fix a bug *double error message*.
2) Fix a bug *passowrd prompt in case of md5 authentication*.
3) Improve the DSN setup dialog.

22 years agoAllow createdb to create comments in current db, not template1.
Bruce Momjian [Tue, 11 Jun 2002 22:21:01 +0000 (22:21 +0000)]
Allow createdb to create comments in current db, not template1.

22 years agoMark as done:
Bruce Momjian [Tue, 11 Jun 2002 21:28:17 +0000 (21:28 +0000)]
Mark as done:

> * -Add SIMILAR TO to allow character classes, 'pg_[a-c]%'

22 years agoFix markup typo.
Thomas G. Lockhart [Tue, 11 Jun 2002 16:00:17 +0000 (16:00 +0000)]
Fix markup typo.

22 years agoImplement SQL99 OVERLAY(). Allows substitution of a substring in a string.
Thomas G. Lockhart [Tue, 11 Jun 2002 15:44:38 +0000 (15:44 +0000)]
Implement SQL99 OVERLAY(). Allows substitution of a substring in a string.
Implement SQL99 SIMILAR TO as a synonym for our existing operator "~".
Implement SQL99 regular expression SUBSTRING(string FROM pat FOR escape).
 Extend the definition to make the FOR clause optional.
 Define textregexsubstr() to actually implement this feature.
Update the regression test to include these new string features.
 All tests pass.
Rename the regular expression support routines from "pg95_xxx" to "pg_xxx".
Define CREATE CHARACTER SET in the parser per SQL99. No implementation yet.

22 years agoImplement SQL99 OVERLAY(). Allows substitution of a substring in a string.
Thomas G. Lockhart [Tue, 11 Jun 2002 15:32:33 +0000 (15:32 +0000)]
Implement SQL99 OVERLAY(). Allows substitution of a substring in a string.
Implement SQL99 SIMILAR TO as a synonym for our existing operator "~".
Implement SQL99 regular expression SUBSTRING(string FROM pat FOR escape).
 Extend the definition to make the FOR clause optional.
 Define textregexsubstr() to actually implement this feature.
Update the regression test to include these new string features.
 All tests pass.
Rename the regular expression support routines from "pg95_xxx" to "pg_xxx".
Define CREATE CHARACTER SET in the parser per SQL99. No implementation yet.

22 years agoKatherine Ward wrote:
Jan Wieck [Tue, 11 Jun 2002 13:40:53 +0000 (13:40 +0000)]
Katherine Ward wrote:
> Changes to avoid collisions with WIN32 & MFC names...
> 1.  Renamed:
>       a.  PROC => PGPROC
>       b.  GetUserName() => GetUserNameFromId()
>       c.  GetCurrentTime() => GetCurrentDateTime()
>       d.  IGNORE => IGNORE_DTF in include/utils/datetime.h & utils/adt/datetim
>
> 2.  Added _P to some lex/yacc tokens:
>       CONST, CHAR, DELETE, FLOAT, GROUP, IN, OUT

Jan

22 years agoFix link.
Bruce Momjian [Tue, 11 Jun 2002 10:36:16 +0000 (10:36 +0000)]
Fix link.

22 years agoUpdate performance section.
Bruce Momjian [Tue, 11 Jun 2002 10:29:54 +0000 (10:29 +0000)]
Update performance section.

22 years agoThe patch does the following:
Barry Lind [Tue, 11 Jun 2002 02:55:16 +0000 (02:55 +0000)]
The patch does the following:
  Allows you to set the loglevel at runtime by adding ?loglevel=X to the connection URL, where 1 = INFO and 2 = DEBUG.
  Automatically turns on logging by calling DriverManager.setPrintWriter(new PrintWriter(System.out)) if one is not already set.
Adds a Driver.info() message that prints out the version number
Adds member variables logDebug and logInfo that can be checked before making logging methods calls
Adds a build number to the version number string.  This build number will need to be manually incremented when we see fit.

----------------------------------------------------------------------
Modified Files:
  org/postgresql/Connection.java org/postgresql/Driver.java.in
  org/postgresql/fastpath/Fastpath.java
  org/postgresql/jdbc1/DatabaseMetaData.java
  org/postgresql/jdbc2/Connection.java
  org/postgresql/jdbc2/DatabaseMetaData.java
  org/postgresql/largeobject/LargeObjectManager.java
  org/postgresql/util/PSQLException.java
  org/postgresql/util/Serialize.java
----------------------------------------------------------------------

22 years agoUpdate fsync FAQ item.
Bruce Momjian [Tue, 11 Jun 2002 02:22:41 +0000 (02:22 +0000)]
Update fsync FAQ item.

22 years agoFix link.
Bruce Momjian [Mon, 10 Jun 2002 20:45:07 +0000 (20:45 +0000)]
Fix link.

22 years agoMove ruler to separate sections.
Bruce Momjian [Mon, 10 Jun 2002 19:49:59 +0000 (19:49 +0000)]
Move ruler to separate sections.

22 years agoAdd item for plpgsql temp table access.
Bruce Momjian [Mon, 10 Jun 2002 19:47:57 +0000 (19:47 +0000)]
Add item for plpgsql temp table access.

22 years agoFix spacing.
Bruce Momjian [Mon, 10 Jun 2002 16:13:54 +0000 (16:13 +0000)]
Fix spacing.

22 years agoUpdate new Russian FAQ, from Victor Vislobokov
Bruce Momjian [Sat, 8 Jun 2002 04:23:45 +0000 (04:23 +0000)]
Update new Russian FAQ, from Victor Vislobokov

22 years agoRemove DEBUG_LEVEL from postgresql.conf. Now uses CLIENT/SERVER_MIN_MESSAGES.
Bruce Momjian [Sat, 8 Jun 2002 04:08:19 +0000 (04:08 +0000)]
Remove DEBUG_LEVEL from postgresql.conf.  Now uses CLIENT/SERVER_MIN_MESSAGES.

22 years agoImprove readability of factorial, from Florian Weimer.
Bruce Momjian [Fri, 7 Jun 2002 23:43:17 +0000 (23:43 +0000)]
Improve readability of factorial, from Florian Weimer.

22 years agoFix for factorial(0::int2) returning 1, from sugita@sra.co.jp.
Bruce Momjian [Fri, 7 Jun 2002 23:41:53 +0000 (23:41 +0000)]
Fix for  factorial(0::int2) returning 1, from  sugita@sra.co.jp.

22 years agoPlease apply attached patch to contrib/intarray (7.2, 7.3).
Bruce Momjian [Fri, 7 Jun 2002 21:52:33 +0000 (21:52 +0000)]
Please apply attached patch to contrib/intarray (7.2, 7.3).

     Fixed bug with '=' operator for gist__int_ops and
     define '=' operator for gist__intbig_ops opclass.
     Now '=' operator is consistent with standard 'array' type.

     Thanks Achilleus Mantzios for bug report and suggestion.

Oleg Bartunov

22 years agoThe attached patch fixes a problem with InstallXLogFileSegment()'s use
Bruce Momjian [Fri, 7 Jun 2002 21:47:45 +0000 (21:47 +0000)]
The attached patch fixes a problem with InstallXLogFileSegment()'s use
of link() under Cygwin:

    http://archives.postgresql.org/pgsql-cygwin/2002-04/msg00072.php

Note that it appears that BeOS and Netware also have the above or
similar problem.

I have only verified that PostgreSQL builds under Cygwin with this
patch.
Since I cannot reproduce the problem, I cannot verify that the proposed
patch solves it.  Nevertheless, both Barry Pederson and David P.
Caldwell
attest that this patch solves the problem.  See the following for
details:

    http://archives.postgresql.org/pgsql-cygwin/2002-05/msg00043.php
    http://archives.postgresql.org/pgsql-cygwin/2002-05/msg00040.php

Jason Tishler

22 years agoNone.
Bruce Momjian [Fri, 7 Jun 2002 20:22:33 +0000 (20:22 +0000)]
None.

22 years agoRemove int16 from libpgeasy examples. Improve error reporting.
Bruce Momjian [Fri, 7 Jun 2002 16:53:40 +0000 (16:53 +0000)]
Remove int16 from libpgeasy examples.  Improve error reporting.

22 years agofixed bug reported by cc.ais40@wanadoo.fr where getObject was returning an Integer...
Barry Lind [Fri, 7 Jun 2002 16:46:24 +0000 (16:46 +0000)]
fixed bug reported by cc.ais40@wanadoo.fr where getObject was returning an Integer for a smallint datatype instead of a Short

22 years agoAdd:
Bruce Momjian [Fri, 7 Jun 2002 04:19:52 +0000 (04:19 +0000)]
Add:

> * Allow DEFERRABLE UNIQUE constraints

22 years agoFix interface example errors causes by backend changes.
Bruce Momjian [Fri, 7 Jun 2002 03:55:15 +0000 (03:55 +0000)]
Fix interface example errors causes by backend changes.

Permaine Cheung

22 years agoMark as done:
Bruce Momjian [Fri, 7 Jun 2002 03:47:06 +0000 (03:47 +0000)]
Mark as done:

>  o -Abort all SET changes made in an aborted transaction

22 years agochange table name to lower case in getColumns
Dave Cramer [Thu, 6 Jun 2002 14:47:52 +0000 (14:47 +0000)]
change table name to lower case in getColumns

22 years agoAdd missing win_setup.h.
Hiroshi Inoue [Thu, 6 Jun 2002 09:47:25 +0000 (09:47 +0000)]
Add missing win_setup.h.

22 years agoAdd *Int8 As* option.
Hiroshi Inoue [Thu, 6 Jun 2002 04:50:47 +0000 (04:50 +0000)]
Add *Int8 As* option.

22 years agofixed getImported/ExportedKeys to be simpler, and return the correct number of keys
Dave Cramer [Wed, 5 Jun 2002 19:12:01 +0000 (19:12 +0000)]
fixed getImported/ExportedKeys to be simpler, and return the correct number of keys

22 years agoAdd /dev/zero mention for anon mmap().
Bruce Momjian [Wed, 5 Jun 2002 04:01:27 +0000 (04:01 +0000)]
Add /dev/zero mention for anon mmap().

22 years agoUpdate Japanese FAQ, from Jun Kuwamura
Bruce Momjian [Tue, 4 Jun 2002 05:12:21 +0000 (05:12 +0000)]
Update Japanese FAQ, from Jun Kuwamura

22 years agoDocument that SM_* variables should be longer.
Bruce Momjian [Mon, 3 Jun 2002 22:57:47 +0000 (22:57 +0000)]
Document that SM_* variables should be longer.

22 years agoUpdate as done:
Bruce Momjian [Mon, 3 Jun 2002 22:46:09 +0000 (22:46 +0000)]
Update as done:

> * -Add deleted bit to index tuples to reduce heap access

22 years agofixed bug reported by Noel Rappin (nrappin@sockeye.com) java Array type handled Times...
Barry Lind [Mon, 3 Jun 2002 17:52:05 +0000 (17:52 +0000)]
fixed bug reported by Noel Rappin (nrappin@sockeye.com) java Array type handled Timestamps incorrectly

22 years agoRemove NOCREATE patch from TODO.detail.
Bruce Momjian [Mon, 3 Jun 2002 17:45:19 +0000 (17:45 +0000)]
Remove NOCREATE patch from TODO.detail.

22 years agoSmall patch to correct the default arraysize associated
Bruce Momjian [Mon, 3 Jun 2002 17:42:11 +0000 (17:42 +0000)]
Small patch to correct the default arraysize associated
with the Cursor object's fetchmany() method.  The API and
inline documentation state that the default is 1.  It
currently defaults to 5.

Patrick Macdonald

22 years agoUpdate index growth problem.
Bruce Momjian [Mon, 3 Jun 2002 17:41:37 +0000 (17:41 +0000)]
Update index growth problem.

22 years agoThe attached patch fixes a typo in the documentation for SET
Bruce Momjian [Mon, 3 Jun 2002 01:10:38 +0000 (01:10 +0000)]
The attached patch fixes a typo in the documentation for SET
CONSTRAINTS.

Neil Conway <neilconway@rogers.com>

22 years agoMention SM_USER should be the same size as the others.
Bruce Momjian [Mon, 3 Jun 2002 01:05:45 +0000 (01:05 +0000)]
Mention SM_USER should be the same size as the others.

22 years agoAdd PQescapeString and PQescapeBytea for Win32.
Bruce Momjian [Sun, 2 Jun 2002 22:36:30 +0000 (22:36 +0000)]
Add PQescapeString and PQescapeBytea for Win32.

22 years agoRemove mention of timestamp() in SGML docs; doesn't exist anymore.
Bruce Momjian [Sun, 2 Jun 2002 21:56:09 +0000 (21:56 +0000)]
Remove mention of timestamp() in SGML docs;  doesn't exist anymore.

22 years agoAllow pod2man 5.005p3 to work with our current sources.
Bruce Momjian [Sun, 2 Jun 2002 21:37:26 +0000 (21:37 +0000)]
Allow pod2man 5.005p3 to work with our current sources.

22 years agoAdd section explaining unspecified expression evaluation order.
Peter Eisentraut [Sat, 1 Jun 2002 20:56:55 +0000 (20:56 +0000)]
Add section explaining unspecified expression evaluation order.

22 years agoIndent verbatim environments.
Peter Eisentraut [Sat, 1 Jun 2002 20:56:00 +0000 (20:56 +0000)]
Indent verbatim environments.

22 years agoMention btree:
Bruce Momjian [Sat, 1 Jun 2002 19:53:31 +0000 (19:53 +0000)]
Mention btree:

> * Add FILLFACTOR to btree index creation

22 years agoFix timestamp to date conversion for the case where timestamp uses a double
Thomas G. Lockhart [Sat, 1 Jun 2002 15:52:15 +0000 (15:52 +0000)]
Fix timestamp to date conversion for the case where timestamp uses a double
 precision storage format. Previously applied the same math as used for the
 64-bit integer storage format case, which was wrong.
 Problem introduced recently when the 64-bit storage format was
 implemented.

22 years agoImprove manual's discussion of locking and MVCC.
Tom Lane [Thu, 30 May 2002 20:45:18 +0000 (20:45 +0000)]
Improve manual's discussion of locking and MVCC.

22 years agosmall fix to testExportedKeys
Dave Cramer [Thu, 30 May 2002 16:39:26 +0000 (16:39 +0000)]
small fix to testExportedKeys

22 years agoadded imported/exported key testDatabaseMetaDataTest.java
Dave Cramer [Thu, 30 May 2002 16:26:55 +0000 (16:26 +0000)]
added imported/exported key testDatabaseMetaDataTest.java

22 years agoAdd a note about the interpretation of amcanmulticol and amindexnulls:
Tom Lane [Wed, 29 May 2002 17:36:40 +0000 (17:36 +0000)]
Add a note about the interpretation of amcanmulticol and amindexnulls:
a multicolumn-capable index AM *must* support nulls in index columns
after the first one.

22 years agoAnother place that needs schema qualification.
Tom Lane [Wed, 29 May 2002 01:49:57 +0000 (01:49 +0000)]
Another place that needs schema qualification.

22 years agoFix some more not-schema-aware queries in pg_dump. Also fix some places
Tom Lane [Wed, 29 May 2002 01:38:56 +0000 (01:38 +0000)]
Fix some more not-schema-aware queries in pg_dump.  Also fix some places
that would do the wrong thing with BLOB OIDs exceeding 2G.

22 years agoRemove 32-bit mention:
Bruce Momjian [Wed, 29 May 2002 00:44:11 +0000 (00:44 +0000)]
Remove 32-bit mention:

> * Add BSD-licensed qsort() for Solaris

22 years agoRearrange LOG_CONNECTIONS code so that two log messages are made:
Tom Lane [Tue, 28 May 2002 23:56:51 +0000 (23:56 +0000)]
Rearrange LOG_CONNECTIONS code so that two log messages are made:
one immediately upon forking to handle a new connection, and one after
the authentication cycle is finished.  Per today's pggeneral discussion.

22 years agoRework pg_dump namespace search criteria so that dumping of user objects
Tom Lane [Tue, 28 May 2002 22:26:57 +0000 (22:26 +0000)]
Rework pg_dump namespace search criteria so that dumping of user objects
having names conflicting with system objects will work --- the search
path is now user-schema, pg_catalog rather than implicitly the other way
around.  Note this requires being careful to explicitly qualify references
to system names whenever pg_catalog is not first in the search path.
Also, add support for dumping ACLs of schemas.

22 years agoQueries used by ruleutils were not schema-proof.
Tom Lane [Tue, 28 May 2002 22:16:15 +0000 (22:16 +0000)]
Queries used by ruleutils were not schema-proof.

22 years agoIn default nextval('foo') expression for a SERIAL column, use double
Tom Lane [Tue, 28 May 2002 22:15:42 +0000 (22:15 +0000)]
In default nextval('foo') expression for a SERIAL column, use double
quotes only when necessary.

22 years agoChange PL/Perl and Pg interface build to use configured compiler and
Peter Eisentraut [Tue, 28 May 2002 16:57:53 +0000 (16:57 +0000)]
Change PL/Perl and Pg interface build to use configured compiler and
Makefile.shlib system, not MakeMaker.

22 years agoFix a bug with building rtree_gist indexes.
Tom Lane [Tue, 28 May 2002 15:24:53 +0000 (15:24 +0000)]
Fix a bug with building rtree_gist indexes.
Patch from Teodor Sigaev.

22 years agoRepair error with not adjusting active scans properly after gistSplit.
Tom Lane [Tue, 28 May 2002 15:22:33 +0000 (15:22 +0000)]
Repair error with not adjusting active scans properly after gistSplit.
Patch from Teodor Sigaev.

22 years agoUpdate:
Bruce Momjian [Tue, 28 May 2002 01:06:23 +0000 (01:06 +0000)]
Update:

> * Add BSD-licensed qsort() for 32-bit Solaris

22 years agoUpdate dblink to work with qualified relation names.
Tom Lane [Mon, 27 May 2002 21:59:12 +0000 (21:59 +0000)]
Update dblink to work with qualified relation names.
From Joe Conway.

22 years agoDistinguish between MaxHeapAttributeNumber and MaxTupleAttributeNumber,
Tom Lane [Mon, 27 May 2002 19:53:33 +0000 (19:53 +0000)]
Distinguish between MaxHeapAttributeNumber and MaxTupleAttributeNumber,
where the latter is made slightly larger to allow for in-memory tuples
containing resjunk attributes.  Responds to today's complaint that one
cannot UPDATE a table containing the allegedly-legal maximum number of
columns.

Also, apply Manfred Koizar's recent patch to avoid extra alignment padding
when there is a null bitmap.  This saves bytes in some cases while not
creating any backward-compatibility problem AFAICS.

22 years agoAdd
Bruce Momjian [Sat, 25 May 2002 22:44:13 +0000 (22:44 +0000)]
Add

> * Add getpid() function to backend

22 years agoRemove AMI_OVERRIDE tests from tqual.c routines; they aren't necessary
Tom Lane [Sat, 25 May 2002 20:00:12 +0000 (20:00 +0000)]
Remove AMI_OVERRIDE tests from tqual.c routines; they aren't necessary
and just slow down normal operations (only fractionally, but a cycle saved
is a cycle earned).  Improve documentation of AMI_OVERRIDE behavior.

22 years agoAlterDatabaseSet() forgot to update the indexes on pg_database.
Tom Lane [Sat, 25 May 2002 16:30:59 +0000 (16:30 +0000)]
AlterDatabaseSet() forgot to update the indexes on pg_database.

22 years agoFix coding error in UTF conversion.
Tom Lane [Fri, 24 May 2002 21:04:34 +0000 (21:04 +0000)]
Fix coding error in UTF conversion.

22 years agoAdd 'volatile' to suppress gcc warning. Not sure why this warning
Tom Lane [Fri, 24 May 2002 19:58:04 +0000 (19:58 +0000)]
Add 'volatile' to suppress gcc warning.  Not sure why this warning
wasn't seen before, maybe the Tcl compiler flags were less strict.

22 years agoWups, managed to break ANALYZE with one aspect of that heap_fetch change.
Tom Lane [Fri, 24 May 2002 19:52:43 +0000 (19:52 +0000)]
Wups, managed to break ANALYZE with one aspect of that heap_fetch change.

22 years agoMark index entries "killed" when they are no longer visible to any
Tom Lane [Fri, 24 May 2002 18:57:57 +0000 (18:57 +0000)]
Mark index entries "killed" when they are no longer visible to any
transaction, so as to avoid returning them out of the index AM.  Saves
repeated heap_fetch operations on frequently-updated rows.  Also detect
queries on unique keys (equality to all columns of a unique index), and
don't bother continuing scan once we have found first match.

Killing is implemented in the btree and hash AMs, but not yet in rtree
or gist, because there isn't an equally convenient place to do it in
those AMs (the outer amgetnext routine can't do it without re-pinning
the index page).

Did some small cleanup on APIs of HeapTupleSatisfies, heap_fetch, and
index_insert to make this a little easier.

22 years agoChange PL/Tcl build to use configured compiler and Makefile.shlib
Peter Eisentraut [Fri, 24 May 2002 18:10:17 +0000 (18:10 +0000)]
Change PL/Tcl build to use configured compiler and Makefile.shlib
system, not Tcl-provided one.

Make sure export file, if any, is cleaned.

Tcl configuration is now read directly in configure and recorded in
Makefile.global.  This eliminates some duplicate efforts and allows
for easier hand-editing of the results, if necessary.

22 years agoAdd $(LDFLAGS) to Windows make rule for postgres executable. Needed
Tom Lane [Wed, 22 May 2002 21:46:40 +0000 (21:46 +0000)]
Add $(LDFLAGS) to Windows make rule for postgres executable.  Needed
to do profiling on Cygwin, per report from Dave Page.

22 years agoModify sequence state storage to eliminate dangling-pointer problem
Tom Lane [Wed, 22 May 2002 21:40:55 +0000 (21:40 +0000)]
Modify sequence state storage to eliminate dangling-pointer problem
exemplified by bug #671.  Moving the storage to relcache turned out to
be a bad idea because relcache might decide to discard the info.  Instead,
open and close the relcache entry on each sequence operation, and use
a record of the current XID to discover whether we already hold
AccessShareLock on the sequence.

22 years agoMinor kibitzing.
Tom Lane [Wed, 22 May 2002 18:33:15 +0000 (18:33 +0000)]
Minor kibitzing.

22 years agoAdd optional "validator" function to languages that can validate the
Peter Eisentraut [Wed, 22 May 2002 17:21:02 +0000 (17:21 +0000)]
Add optional "validator" function to languages that can validate the
function body (and other properties) as a function in the language
is created.  This generalizes ad hoc code that already existed for
the built-in languages.

The validation now happens after the pg_proc tuple of the new function
is created, so it is possible to define recursive SQL functions.

Add some regression test cases that cover bogus function definition
attempts.

22 years agoMake RelationForgetRelation error out if the relcache entry has nonzero
Tom Lane [Wed, 22 May 2002 15:57:40 +0000 (15:57 +0000)]
Make RelationForgetRelation error out if the relcache entry has nonzero
reference count.  This avoids leaving dangling pointers around, as in
recent bug report against sequences (bug# 671).

22 years agoUn-break table creation.
Tom Lane [Wed, 22 May 2002 15:35:43 +0000 (15:35 +0000)]
Un-break table creation.

22 years agoAllow
Hiroshi Inoue [Wed, 22 May 2002 07:46:58 +0000 (07:46 +0000)]
Allow
  CREATE VIEW as SELECT CTID, ....
  SELECT currtid( a view, ..).

22 years ago1) Support Keyset Driven driver cursors.
Hiroshi Inoue [Wed, 22 May 2002 05:51:03 +0000 (05:51 +0000)]
1) Support Keyset Driven driver cursors.
2) Supprt ARD precision/scale and SQL_C_NUEMRIC.
3) Minimal implementation of SQLGetDiagField().
4) SQLRowCount() reports the result of SQLSetPos and SQLBulkOperation.
5) int8 -> SQL_NUMERIC for Microsoft Jet.
6) Support isolation level change.
7) ODBC3.0 SQLSTATE code.
8) Append mode log files.

22 years agoSince COPY fires triggers, it seems like a good idea for it to use
Tom Lane [Tue, 21 May 2002 22:59:01 +0000 (22:59 +0000)]
Since COPY fires triggers, it seems like a good idea for it to use
a frozen (copied) snapshot too.  Move execMain's snapshot copying code
out into a subroutine in case we find other places that need it.

22 years agoRemove SetQuerySnapshot() from FETCH processing. No longer necessary
Tom Lane [Tue, 21 May 2002 22:18:08 +0000 (22:18 +0000)]
Remove SetQuerySnapshot() from FETCH processing.  No longer necessary
or appropriate, since the snapshot that will be used by the cursor was
frozen when ExecutorStart was run for it.

22 years agoRemove global variable scanCommandId in favor of storing a command ID
Tom Lane [Tue, 21 May 2002 22:05:55 +0000 (22:05 +0000)]
Remove global variable scanCommandId in favor of storing a command ID
in snapshots, per my proposal of a few days ago.  Also, tweak heapam.c
routines (heap_insert, heap_update, heap_delete, heap_mark4update) to
be passed the command ID to use, instead of doing GetCurrentCommandID.
For catalog updates they'll still get passed current command ID, but
for updates generated from the main executor they'll get passed the
command ID saved in the snapshot the query is using.  This should fix
some corner cases associated with functions and triggers that advance
current command ID while an outer query is still in progress.

22 years agoSpell long option with -c not --, for compatibility's sake.
Tom Lane [Tue, 21 May 2002 19:06:00 +0000 (19:06 +0000)]
Spell long option with -c not --, for compatibility's sake.

22 years agoRepair OPEN cursor(args), which I broke on 11/29/01 with a change to
Tom Lane [Tue, 21 May 2002 18:50:16 +0000 (18:50 +0000)]
Repair OPEN cursor(args), which I broke on 11/29/01 with a change to
be smarter about parentheses in read_sql_construct().  Sigh.

22 years agoRestructure indexscan API (index_beginscan, index_getnext) per
Tom Lane [Mon, 20 May 2002 23:51:44 +0000 (23:51 +0000)]
Restructure indexscan API (index_beginscan, index_getnext) per
yesterday's proposal to pghackers.  Also remove unnecessary parameters
to heap_beginscan, heap_rescan.  I modified pg_proc.h to reflect the
new numbers of parameters for the AM interface routines, but did not
force an initdb because nothing actually looks at those fields.

22 years agoMark as done:
Bruce Momjian [Mon, 20 May 2002 22:22:27 +0000 (22:22 +0000)]
Mark as done:

>  o -Disallow missing columns in INSERT ... VALUES, per ANSI