]> granicus.if.org Git - postgresql/log
postgresql
22 years agoPre-page write additions.
Bruce Momjian [Thu, 20 Dec 2001 22:01:29 +0000 (22:01 +0000)]
Pre-page write additions.

22 years agoAdd memcmp() test and new memcmp.c file, for SunOS. Tested by Tatsuo.
Bruce Momjian [Thu, 20 Dec 2001 21:23:05 +0000 (21:23 +0000)]
Add memcmp() test and new memcmp.c file, for SunOS.  Tested by Tatsuo.

22 years agoFix for SunOS4 horology test
Tatsuo Ishii [Thu, 20 Dec 2001 04:25:00 +0000 (04:25 +0000)]
Fix for SunOS4 horology test

22 years agoMIN() -> Min, fixes compile problem.
Bruce Momjian [Thu, 20 Dec 2001 02:39:26 +0000 (02:39 +0000)]
MIN() -> Min, fixes compile problem.

22 years agoReflect Tom's change
Tatsuo Ishii [Thu, 20 Dec 2001 01:56:08 +0000 (01:56 +0000)]
Reflect Tom's change

22 years agoFix buffer-overrun problem in pretty printer.
Tom Lane [Wed, 19 Dec 2001 22:35:35 +0000 (22:35 +0000)]
Fix buffer-overrun problem in pretty printer.

22 years agoFix pgstattuple to acquire a read lock on the target table. This
Tom Lane [Wed, 19 Dec 2001 20:28:41 +0000 (20:28 +0000)]
Fix pgstattuple to acquire a read lock on the target table.  This
prevents embarassments such as having the table dropped or truncated
partway through the scan.  Also, fix free space calculation to include
pages that currently contain no tuples.

22 years agoTemporarily dike out GetUndoRecPtr() in checkpoint generation, since we
Tom Lane [Wed, 19 Dec 2001 19:42:51 +0000 (19:42 +0000)]
Temporarily dike out GetUndoRecPtr() in checkpoint generation, since we
do not use the undo pointer anyway.  This is a quick-hack solution for
the three-way deadlock condition discussed in pghackers 17-Dec-01.
Need to find a better way of doing it.

22 years agoNew edition
Peter Eisentraut [Wed, 19 Dec 2001 18:53:58 +0000 (18:53 +0000)]
New edition

22 years agoBetter formatting
Peter Eisentraut [Wed, 19 Dec 2001 18:53:47 +0000 (18:53 +0000)]
Better formatting

22 years agoMinor improvements
Peter Eisentraut [Wed, 19 Dec 2001 18:50:48 +0000 (18:50 +0000)]
Minor improvements

22 years agoSome tuning
Peter Eisentraut [Wed, 19 Dec 2001 18:49:56 +0000 (18:49 +0000)]
Some tuning

22 years agoFix warning
Peter Eisentraut [Wed, 19 Dec 2001 18:49:24 +0000 (18:49 +0000)]
Fix warning

22 years agoTweak to make freebsd/alpha use the correct float8 comparison file.
Tom Lane [Wed, 19 Dec 2001 17:25:55 +0000 (17:25 +0000)]
Tweak to make freebsd/alpha use the correct float8 comparison file.

22 years agoMake sure that all variants of HeapTupleSatisfies will do the right thing
Tom Lane [Wed, 19 Dec 2001 17:18:39 +0000 (17:18 +0000)]
Make sure that all variants of HeapTupleSatisfies will do the right thing
if presented with a tuple in process of being moved by VACUUM.  Per
bug report from Brian Hirt.

22 years agoMove geqo enable into proper section.
Bruce Momjian [Mon, 17 Dec 2001 19:09:01 +0000 (19:09 +0000)]
Move geqo enable into proper section.

22 years agoUpdate TODO list.
Bruce Momjian [Sat, 15 Dec 2001 17:28:51 +0000 (17:28 +0000)]
Update TODO list.

22 years agoUpdate TODO list.
Bruce Momjian [Sat, 15 Dec 2001 16:11:57 +0000 (16:11 +0000)]
Update TODO list.

22 years agoMore comment for libpgeasy.
Bruce Momjian [Fri, 14 Dec 2001 02:15:04 +0000 (02:15 +0000)]
More comment for libpgeasy.

22 years agoFix double-memory free in libpgeasy; problem introduced yesterday.
Bruce Momjian [Fri, 14 Dec 2001 00:52:53 +0000 (00:52 +0000)]
Fix double-memory free in libpgeasy;  problem introduced yesterday.

22 years agoClean up comment in libpgeasy.
Bruce Momjian [Fri, 14 Dec 2001 00:40:31 +0000 (00:40 +0000)]
Clean up comment in libpgeasy.

22 years agoAdd Swedish
Peter Eisentraut [Thu, 13 Dec 2001 22:06:33 +0000 (22:06 +0000)]
Add Swedish

22 years agoupdate
Peter Eisentraut [Thu, 13 Dec 2001 22:04:40 +0000 (22:04 +0000)]
update

22 years agoUpdate from Serguei Mokhov
Peter Eisentraut [Thu, 13 Dec 2001 22:04:12 +0000 (22:04 +0000)]
Update from Serguei Mokhov

22 years agoUpdate Hungarian from Kovacs Zoltan
Peter Eisentraut [Thu, 13 Dec 2001 22:02:20 +0000 (22:02 +0000)]
Update Hungarian from Kovacs Zoltan

22 years ago--with-pam was claiming to take an argument, but didn't. Help display fixed.
Peter Eisentraut [Thu, 13 Dec 2001 22:00:22 +0000 (22:00 +0000)]
--with-pam was claiming to take an argument, but didn't.  Help display fixed.

22 years agorevert last change
Peter Eisentraut [Thu, 13 Dec 2001 18:39:04 +0000 (18:39 +0000)]
revert last change

22 years ago>>I confirmed the UNION hack is no longer required. Thanks! Is it too late
Bruce Momjian [Thu, 13 Dec 2001 10:48:39 +0000 (10:48 +0000)]
>>I confirmed the UNION hack is no longer required. Thanks! Is it too late
>>to change the README in contrib/dblink?
>>
>
> No, I don't think that's a problem.  Send a patch.
>

Here's a (documentation only) patch for the contrib/dblink README.

Joe Conway

22 years agoFree libpgeasy result structure on database close; fixed memory leak.
Bruce Momjian [Thu, 13 Dec 2001 09:40:18 +0000 (09:40 +0000)]
Free libpgeasy result structure on database close; fixed memory leak.

22 years agotag it as b4, with all the changes that have gone on ... REL7_2_BETA4
PostgreSQL Daemon [Wed, 12 Dec 2001 13:30:12 +0000 (13:30 +0000)]
tag it as b4, with all the changes that have gone on ...

22 years agoDon't accept names of complex types (ie, relation types) as being
Tom Lane [Wed, 12 Dec 2001 03:28:49 +0000 (03:28 +0000)]
Don't accept names of complex types (ie, relation types) as being
requests for implicit trivial coercions.  Prevents sillinesses like
this one:
regression=# select x.int8_tbl.q1 from int8_tbl x;
ERROR:  fmgr_info: function 270997776: cache lookup failed

22 years agoApplied patch from Thomas O'Dowd that fixes timestamp parsing. The jdbc code
Barry Lind [Tue, 11 Dec 2001 04:48:05 +0000 (04:48 +0000)]
Applied patch from Thomas O'Dowd that fixes timestamp parsing.  The jdbc code
wasn't updated to handle more than two decimal digits for fractional seconds
that now are possible in 7.2.  This patch fixes the timestamp parsing logic.
I have built and tested on both jdbc1 and jdbc2.

22 years agoPatch from Ned Wolpert that fixes a bug that caused the cache of types not
Barry Lind [Tue, 11 Dec 2001 04:44:23 +0000 (04:44 +0000)]
Patch from Ned Wolpert that fixes a bug that caused the cache of types not
to be used, causing extra sql statements to be executed.  This was a
significant performance problem with the database meta data classes.
The fix is a simple one liner.

22 years agoMake sure that inlined S_UNLOCK is marked as an update of a 'volatile'
Tom Lane [Tue, 11 Dec 2001 02:58:49 +0000 (02:58 +0000)]
Make sure that inlined S_UNLOCK is marked as an update of a 'volatile'
object.  This should prevent the compiler from reordering loads and stores
into or out of a critical section.

22 years agoRepair roundoff-error problem for stddev/variance results near zero,
Tom Lane [Tue, 11 Dec 2001 02:02:12 +0000 (02:02 +0000)]
Repair roundoff-error problem for stddev/variance results near zero,
per complaint from Kemin Zhou.
Fix lack of precision in numeric stddev/variance.

22 years agoSuppress subquery pullup and pushdown when the subquery has any
Tom Lane [Mon, 10 Dec 2001 22:54:12 +0000 (22:54 +0000)]
Suppress subquery pullup and pushdown when the subquery has any
set-returning functions in its target list.  This ensures that we
won't rewrite the query in a way that places set-returning functions
into quals (WHERE clauses).  Cf. bug reports from Joe Conway.

22 years agoDeclare LWLock pointers as volatile to prevent AIX compiler from
Tom Lane [Mon, 10 Dec 2001 21:13:50 +0000 (21:13 +0000)]
Declare LWLock pointers as volatile to prevent AIX compiler from
reordering operations at its whim.  Releasing TAS lock before we've
finished updating proc structure is uncool.

22 years agoUpdate from Weiping He
Peter Eisentraut [Mon, 10 Dec 2001 18:45:57 +0000 (18:45 +0000)]
Update from Weiping He

22 years agoUpdate from Serguei Mokhov
Peter Eisentraut [Mon, 10 Dec 2001 18:44:44 +0000 (18:44 +0000)]
Update from Serguei Mokhov

22 years agoAIX 5 needs same resultmap tweaks as AIX 4.
Tom Lane [Mon, 10 Dec 2001 16:31:23 +0000 (16:31 +0000)]
AIX 5 needs same resultmap tweaks as AIX 4.

22 years agoRepair case-conversion error in to_date's handling of roman numerals.
Tom Lane [Mon, 10 Dec 2001 15:34:05 +0000 (15:34 +0000)]
Repair case-conversion error in to_date's handling of roman numerals.
From Manuel Sugawara, approved by Karel Zak.

22 years agocommitted the missing files
Michael Meskes [Mon, 10 Dec 2001 14:55:47 +0000 (14:55 +0000)]
committed the missing files

22 years agoRevert breakage
Peter Eisentraut [Mon, 10 Dec 2001 13:07:16 +0000 (13:07 +0000)]
Revert breakage

22 years agoAdd French translation
Peter Eisentraut [Mon, 10 Dec 2001 13:03:55 +0000 (13:03 +0000)]
Add French translation

22 years agoupdate
Peter Eisentraut [Mon, 10 Dec 2001 13:03:12 +0000 (13:03 +0000)]
update

22 years agoAdd new Japanese FAQ versions.
Bruce Momjian [Mon, 10 Dec 2001 11:17:51 +0000 (11:17 +0000)]
Add new Japanese FAQ versions.

22 years agoRemove duplicate lines from previous patch attempt. Trouble with my
Thomas G. Lockhart [Mon, 10 Dec 2001 00:11:06 +0000 (00:11 +0000)]
Remove duplicate lines from previous patch attempt. Trouble with my
 DSL line at home broke things right in the middle of an update. :(

22 years agoSupport ODBC-style CURRENT_TIME, CURRENT_USER, etc with trailing empty
Thomas G. Lockhart [Mon, 10 Dec 2001 00:01:14 +0000 (00:01 +0000)]
Support ODBC-style CURRENT_TIME, CURRENT_USER, etc with trailing empty
 parens. This is not SQL spec syntax, so later we will remove
 this extension from gram.y.

22 years agoFixed several bugs concerning indicators and added error messages instead of segfaults.
Michael Meskes [Sun, 9 Dec 2001 15:27:49 +0000 (15:27 +0000)]
Fixed several bugs concerning indicators and added error messages instead of segfaults.

22 years agoAllow variable (unrestricted) precision for TIME and TIMESTAMP types in parser.
Thomas G. Lockhart [Sun, 9 Dec 2001 04:39:39 +0000 (04:39 +0000)]
Allow variable (unrestricted) precision for TIME and TIMESTAMP types in parser.

22 years agoHonor the typmod field for time zone intervals.
Thomas G. Lockhart [Sun, 9 Dec 2001 04:37:50 +0000 (04:37 +0000)]
Honor the typmod field for time zone intervals.

22 years agoFix ecpg to allow pointer to structs.
Michael Meskes [Sat, 8 Dec 2001 20:43:35 +0000 (20:43 +0000)]
Fix ecpg to allow pointer to structs.

22 years agoUse PostgreSQL consistantly throughout docs. Before, usage was split evenly
Thomas G. Lockhart [Sat, 8 Dec 2001 03:24:40 +0000 (03:24 +0000)]
Use PostgreSQL consistantly throughout docs. Before, usage was split evenly
 between Postgres and PostgreSQL.

22 years agoUpdate list of currently supported platforms.
Thomas G. Lockhart [Sat, 8 Dec 2001 03:24:23 +0000 (03:24 +0000)]
Update list of currently supported platforms.
Mention SQL9x precision syntax for date/time types.
Use PostgreSQL consistantly throughout docs. Before, usage was split evenly
 between Postgres and PostgreSQL.

22 years agoFix a lo_manage()'s bug.
Hiroshi Inoue [Fri, 7 Dec 2001 04:18:31 +0000 (04:18 +0000)]
Fix a lo_manage()'s bug.

22 years ago - Removed debug message from preproc.y.
Michael Meskes [Thu, 6 Dec 2001 14:46:20 +0000 (14:46 +0000)]
    - Removed debug message from preproc.y.
        - Fixed some bugs in exec sql var and exec sql type command.

22 years agoUpdate TODO list.
Bruce Momjian [Thu, 6 Dec 2001 10:19:24 +0000 (10:19 +0000)]
Update TODO list.

22 years agoAdd to java.
Bruce Momjian [Wed, 5 Dec 2001 23:12:16 +0000 (23:12 +0000)]
Add to java.

22 years agoUpdate TODO list.
Bruce Momjian [Wed, 5 Dec 2001 23:11:30 +0000 (23:11 +0000)]
Update TODO list.

22 years agoUpdate TODO list.
Bruce Momjian [Wed, 5 Dec 2001 21:25:28 +0000 (21:25 +0000)]
Update TODO list.

22 years agoUpdate TODO list.
Bruce Momjian [Wed, 5 Dec 2001 21:23:45 +0000 (21:23 +0000)]
Update TODO list.

22 years agoAdd mention of CREATE TABLE and atttypmod problems.
Bruce Momjian [Wed, 5 Dec 2001 21:07:23 +0000 (21:07 +0000)]
Add mention of CREATE TABLE and atttypmod problems.

22 years agoUpdate TODO list.
Bruce Momjian [Wed, 5 Dec 2001 21:06:40 +0000 (21:06 +0000)]
Update TODO list.

22 years agoAdd mention of aggregates and updates causing problems.
Bruce Momjian [Wed, 5 Dec 2001 21:02:43 +0000 (21:02 +0000)]
Add mention of aggregates and updates causing problems.

22 years agoWe forgot to mention in README.rtree_gist we implemented new
Bruce Momjian [Wed, 5 Dec 2001 18:25:28 +0000 (18:25 +0000)]
We forgot to mention in README.rtree_gist we implemented new
picksplit algorithm for Rtree. Please apply patch below to CVS

Oleg Bartunov

22 years agoAttached is the FAQ_MSWIN patch as promised. Hopefully, it is not too
Bruce Momjian [Wed, 5 Dec 2001 18:23:40 +0000 (18:23 +0000)]
Attached is the FAQ_MSWIN patch as promised.  Hopefully, it is not too
little, too late...

Jason Tishler

22 years agoUpdate TODO list.
Bruce Momjian [Wed, 5 Dec 2001 18:20:42 +0000 (18:20 +0000)]
Update TODO list.

22 years ago - Fixed variable handling in AT statement.
Michael Meskes [Wed, 5 Dec 2001 15:32:07 +0000 (15:32 +0000)]
    - Fixed variable handling in AT statement.
        - Fixed bug that caused segfault when given incorrect DB name.
        - Fixed bug in ecpglib causing indicator to list the size of the
          variable instead of the size of the data.

22 years agoUpdate FAQ.
Bruce Momjian [Wed, 5 Dec 2001 05:41:13 +0000 (05:41 +0000)]
Update FAQ.

22 years agoUpdate FAQ.
Bruce Momjian [Wed, 5 Dec 2001 05:38:51 +0000 (05:38 +0000)]
Update FAQ.

22 years agoUpdate FAQ.
Bruce Momjian [Wed, 5 Dec 2001 05:33:48 +0000 (05:33 +0000)]
Update FAQ.

22 years agoFix for usage of spirntf in more portable way.
Tatsuo Ishii [Wed, 5 Dec 2001 02:06:19 +0000 (02:06 +0000)]
Fix for usage of spirntf in more portable way.

22 years agoRe-enable SunOS4 port. Replace
Tatsuo Ishii [Wed, 5 Dec 2001 02:03:59 +0000 (02:03 +0000)]
Re-enable SunOS4 port. Replace
RTLD_LAZY | RTLD_GLOBAL to 1.
It seems sunos4.h was accidentally modified between 7.1 and 7.2.

22 years agoEnforce restriction that COPY DELIMITERS string must be exactly one
Tom Lane [Tue, 4 Dec 2001 21:19:57 +0000 (21:19 +0000)]
Enforce restriction that COPY DELIMITERS string must be exactly one
character; replace strchr() search with simple comparison to speed up
COPY IN.  Per discussion in pghackers.

22 years agoChange appendStringInfoChar to appendStringInfoCharMacro in a couple of
Tom Lane [Tue, 4 Dec 2001 20:57:22 +0000 (20:57 +0000)]
Change appendStringInfoChar to appendStringInfoCharMacro in a couple of
hot spots --- buys about 10% in Verner's INSERT example.

22 years agoUpdate possible release date 2001-12-??
Bruce Momjian [Tue, 4 Dec 2001 20:25:01 +0000 (20:25 +0000)]
Update possible release date 2001-12-??

22 years agoUpdate HISTORY to current.
Bruce Momjian [Tue, 4 Dec 2001 20:15:41 +0000 (20:15 +0000)]
Update HISTORY to current.

22 years agoReplace pq_getbytes(&ch, 1) calls with pq_getbyte(), which is easier
Tom Lane [Tue, 4 Dec 2001 19:40:17 +0000 (19:40 +0000)]
Replace pq_getbytes(&ch, 1) calls with pq_getbyte(), which is easier
to use and significantly faster.  This tweak saves 25% (!) of the runtime
of COPY IN in a test with 8000-character lines.  I wouldn't normally
commit a performance improvement this late in the cycle, but 25% got
my attention...

22 years agoRemove now-dead code for processing CONSTR_UNIQUE nodes in
Tom Lane [Tue, 4 Dec 2001 17:19:48 +0000 (17:19 +0000)]
Remove now-dead code for processing CONSTR_UNIQUE nodes in
AlterTableAddConstraint.  Earlier reorganization of the parser's
processing of ALTER TABLE means that these node types no longer get here.

22 years agoUpdate FAQ_DEV.
Bruce Momjian [Tue, 4 Dec 2001 17:12:25 +0000 (17:12 +0000)]
Update FAQ_DEV.

22 years agoHi Bruce, here is a first version of the hu.po file for the backend. I
Bruce Momjian [Tue, 4 Dec 2001 17:00:51 +0000 (17:00 +0000)]
Hi Bruce, here is a first version of the hu.po file for the backend. I
tested it with 7.2b3 and worked fine.

Regards, Zoltan

22 years agoSuppress data directory access-permissions check when __CYGWIN__,
Tom Lane [Tue, 4 Dec 2001 16:17:48 +0000 (16:17 +0000)]
Suppress data directory access-permissions check when __CYGWIN__,
per recent pghackers discussions.  Improving this should be a TODO
for 7.3.

22 years ago Fixed dumping of structs without indicators.
Michael Meskes [Tue, 4 Dec 2001 12:33:15 +0000 (12:33 +0000)]
Fixed dumping of structs without indicators.

22 years agoUpdate FAQ.
Bruce Momjian [Tue, 4 Dec 2001 07:12:08 +0000 (07:12 +0000)]
Update FAQ.

22 years agoAnother test of double-commit. It used only commit one file.
Bruce Momjian [Tue, 4 Dec 2001 07:10:45 +0000 (07:10 +0000)]
Another test of double-commit.  It used only commit one file.

22 years agoTest of double-commit.
Bruce Momjian [Tue, 4 Dec 2001 07:09:53 +0000 (07:09 +0000)]
Test of double-commit.

22 years agoUpdate FAQ_DEV.
Bruce Momjian [Tue, 4 Dec 2001 06:44:45 +0000 (06:44 +0000)]
Update FAQ_DEV.

22 years agoUpdate FAQ_DEV.
Bruce Momjian [Tue, 4 Dec 2001 06:40:09 +0000 (06:40 +0000)]
Update FAQ_DEV.

22 years agoUpdate FAQ_DEV.
Bruce Momjian [Tue, 4 Dec 2001 06:37:26 +0000 (06:37 +0000)]
Update FAQ_DEV.

22 years agoUpdate FAQ_DEV.
Bruce Momjian [Tue, 4 Dec 2001 06:35:51 +0000 (06:35 +0000)]
Update FAQ_DEV.

22 years agoUpdate FAQ_DEV.
Bruce Momjian [Tue, 4 Dec 2001 06:26:59 +0000 (06:26 +0000)]
Update FAQ_DEV.

22 years agoUpdate FAQ_DEV.
Bruce Momjian [Tue, 4 Dec 2001 06:20:53 +0000 (06:20 +0000)]
Update FAQ_DEV.

22 years agoUpdate FAQ_DEV.
Bruce Momjian [Tue, 4 Dec 2001 06:20:09 +0000 (06:20 +0000)]
Update FAQ_DEV.

22 years agoUpdate FAQ_DEV.
Bruce Momjian [Tue, 4 Dec 2001 06:14:44 +0000 (06:14 +0000)]
Update FAQ_DEV.

22 years agoUpdate FAQ_DEV.
Bruce Momjian [Tue, 4 Dec 2001 06:14:04 +0000 (06:14 +0000)]
Update FAQ_DEV.

22 years agoRecommend --enable-debug with somewhat more enthusiasm.
Tom Lane [Tue, 4 Dec 2001 02:21:35 +0000 (02:21 +0000)]
Recommend --enable-debug with somewhat more enthusiasm.

22 years agoFix trigger example for PG 7.2 change: count(*) now returns int8.
Tom Lane [Tue, 4 Dec 2001 02:07:11 +0000 (02:07 +0000)]
Fix trigger example for PG 7.2 change: count(*) now returns int8.

22 years agoDocument max-processes-per-user limit as a possible cause of failures
Tom Lane [Tue, 4 Dec 2001 01:49:17 +0000 (01:49 +0000)]
Document max-processes-per-user limit as a possible cause of failures
in the parallel regression tests.  Suggest editing parallel_schedule
as a workaround if one cannot fix the system limits.

22 years agoAdd some more material to the section about partial indexes.
Tom Lane [Tue, 4 Dec 2001 01:22:13 +0000 (01:22 +0000)]
Add some more material to the section about partial indexes.

22 years agoUpdate to get around backslash problems
Peter Eisentraut [Mon, 3 Dec 2001 19:56:43 +0000 (19:56 +0000)]
Update to get around backslash problems