OSDN Git Service

pg-rex/syncrep.git
16 years agoUpdate OOM wording.
Bruce Momjian [Sun, 16 Dec 2007 11:24:25 +0000 (11:24 +0000)]
Update OOM wording.

16 years agoAdd docs about OOM killer.
Bruce Momjian [Sun, 16 Dec 2007 11:22:33 +0000 (11:22 +0000)]
Add docs about OOM killer.

lst_hoe01@kwsoft.de

16 years agoMention pg_ctl kill for Win32 in docs.
Bruce Momjian [Sun, 16 Dec 2007 10:17:13 +0000 (10:17 +0000)]
Mention pg_ctl kill for Win32 in docs.

16 years agoUpdate pgpool-II mention.
Bruce Momjian [Sun, 16 Dec 2007 09:44:27 +0000 (09:44 +0000)]
Update pgpool-II mention.

16 years agoMark 8.3 as likely now January, 2008 in release notes.
Bruce Momjian [Sun, 16 Dec 2007 07:19:32 +0000 (07:19 +0000)]
Mark 8.3 as likely now January, 2008 in release notes.

16 years agoGet dllwrap name from variable instead of hardcoded.
Magnus Hagander [Sat, 15 Dec 2007 16:21:35 +0000 (16:21 +0000)]
Get dllwrap name from variable instead of hardcoded.

Per complaint from Richard Evans

16 years agoFix example archive_command for standalone backups so it doesn't return spurious...
Andrew Dunstan [Sat, 15 Dec 2007 15:41:02 +0000 (15:41 +0000)]
Fix example archive_command for standalone backups so it doesn't return spurious non-zero.

16 years agoUse clearer error message for gmake postgres.pdf:
Bruce Momjian [Sat, 15 Dec 2007 10:28:21 +0000 (10:28 +0000)]
Use clearer error message for gmake postgres.pdf:

  Makefile:171: *** Invalid target;  use postgres-A4.pdf or postgres-US.pdf as targets.  Stop.

16 years agoImprove documentation about Julian dates; in particular, point out the
Tom Lane [Sat, 15 Dec 2007 01:18:34 +0000 (01:18 +0000)]
Improve documentation about Julian dates; in particular, point out the
difference between Julian and Gregorian reckoning of when JD 0 was.

16 years agoMake error message more accurate
Peter Eisentraut [Fri, 14 Dec 2007 14:11:02 +0000 (14:11 +0000)]
Make error message more accurate

16 years agoAdd default error rules for making postgres.pdf and postgres.ps,
Bruce Momjian [Fri, 14 Dec 2007 13:20:30 +0000 (13:20 +0000)]
Add default error rules for making postgres.pdf and postgres.ps,
suggesting proper target names.

16 years agoCorrect result type of convert_to, per Pavel Stehule.
Alvaro Herrera [Thu, 13 Dec 2007 13:22:05 +0000 (13:22 +0000)]
Correct result type of convert_to, per Pavel Stehule.

16 years agoClarify log messages
Peter Eisentraut [Thu, 13 Dec 2007 11:55:44 +0000 (11:55 +0000)]
Clarify log messages

16 years agoChange a couple of examples to say ALTER MAPPING instead of ADD MAPPING,
Tom Lane [Thu, 13 Dec 2007 06:32:47 +0000 (06:32 +0000)]
Change a couple of examples to say ALTER MAPPING instead of ADD MAPPING,
per Oleg.

16 years agoMark items needing updating for beta stamping.
Bruce Momjian [Thu, 13 Dec 2007 02:02:20 +0000 (02:02 +0000)]
Mark items needing updating for beta stamping.

16 years agoImprove the method of localizing column names and other fixed strings in
Tom Lane [Wed, 12 Dec 2007 21:41:47 +0000 (21:41 +0000)]
Improve the method of localizing column names and other fixed strings in
psql's \d commands and other uses of printQuery().  Previously we would pass
these strings through gettext() and then send them to the server as literals
in the SQL query.  But the code was not set up to handle doubling of quotes in
the strings, causing failure if a translation attempted to use the wrong kind
of quote marks, as indeed is now the case for (at least) the French
translation of \dFp.  Another hazard was that gettext() would translate to
whatever encoding was implied by the client's LC_CTYPE setting, which might be
different from the client_encoding setting, which would probably cause the
server to reject the query as mis-encoded.  The new arrangement is to send the
untranslated ASCII strings to the server, and do the translations inside
printQuery() after the query results come back.  Per report from Guillaume
Lelarge and subsequent discussion.

16 years agoClean up unportable attempt to use #if inside a macro call, also
Tom Lane [Wed, 12 Dec 2007 16:53:14 +0000 (16:53 +0000)]
Clean up unportable attempt to use #if inside a macro call, also
faulty code in third arm of #if.  Per buildfarm reports.

16 years agoProvide a more accurate, detailed log message when the archive command fails.
Peter Eisentraut [Wed, 12 Dec 2007 09:39:54 +0000 (09:39 +0000)]
Provide a more accurate, detailed log message when the archive command fails.

16 years agoAdd random I/O URL:
Bruce Momjian [Wed, 12 Dec 2007 07:06:27 +0000 (07:06 +0000)]
Add random I/O URL:

>   http://archives.postgresql.org/pgsql-hackers/2007-12/msg00027.php

16 years agoThis item was fixed in a different way, so has been removed:
Bruce Momjian [Wed, 12 Dec 2007 06:59:56 +0000 (06:59 +0000)]
This item was fixed in a different way, so has been removed:

<
< * Allow free-behind capability for large sequential scans to avoid
<   kernel cache spoiling
<
<   Posix_fadvise() can control both sequential/random file caching and
<   free-behind behavior, but it is unclear how the setting affects other
<   backends that also have the file open, and the feature is not supported
<   on all operating systems.

16 years agoAdjust release notes for HISTORY.html generation.
Bruce Momjian [Wed, 12 Dec 2007 06:58:33 +0000 (06:58 +0000)]
Adjust release notes for HISTORY.html generation.

16 years agoChange 'merge sort join', a phrase we use nowhere else, to the more
Tom Lane [Wed, 12 Dec 2007 06:23:27 +0000 (06:23 +0000)]
Change 'merge sort join', a phrase we use nowhere else, to the more
usual 'merge join'.

16 years agoMake CSV column ordering a bit more logical.
Alvaro Herrera [Tue, 11 Dec 2007 20:07:31 +0000 (20:07 +0000)]
Make CSV column ordering a bit more logical.

16 years agoMake documentation of -W options more accurate and uniform.
Tom Lane [Tue, 11 Dec 2007 19:57:32 +0000 (19:57 +0000)]
Make documentation of -W options more accurate and uniform.

16 years agoRemove the long-deprecated -u option from psql, since it does nothing very
Tom Lane [Tue, 11 Dec 2007 19:01:06 +0000 (19:01 +0000)]
Remove the long-deprecated -u option from psql, since it does nothing very
useful and confuses people who think it is the same as -U.  (Eventually
we might want to re-introduce it as being an alias for -U, but that should
not happen until the switch has actually not been there for a few releases.)
Likewise in pg_dump and pg_restore.  Per gripe from Robert Treat and
subsequent discussion.

16 years agoTypo pointed out by Erwin Brandstetter. Remove the whole makes it even
Magnus Hagander [Tue, 11 Dec 2007 18:30:20 +0000 (18:30 +0000)]
Typo pointed out by Erwin Brandstetter. Remove the whole makes it even
better, per JD.

16 years agoExtend the format of CSV logs to include the additional information supplied
Alvaro Herrera [Tue, 11 Dec 2007 15:19:05 +0000 (15:19 +0000)]
Extend the format of CSV logs to include the additional information supplied
with the logged event.  CSV logs are now a first-class citizen along plain
text logs in that they carry much of the same information.

Per complaint from depesz on bug #3799.

16 years agoUse _USE_32BIT_TIME_T when building with MSVC. Also, enforce that it's
Magnus Hagander [Tue, 11 Dec 2007 14:34:43 +0000 (14:34 +0000)]
Use _USE_32BIT_TIME_T when building with MSVC. Also, enforce that it's
used when building addons.

Dave Page

16 years agoRemove the -P options of oid2name and pgbench, as they are security
Tom Lane [Tue, 11 Dec 2007 02:31:49 +0000 (02:31 +0000)]
Remove the -P options of oid2name and pgbench, as they are security
hazards.  Instead teach these programs to prompt for a password when
necessary, just like all our other programs.
I did not bother to invent -W switches for them, since the return on
investment seems so low.

16 years agoMake vacuumlo prompt for password when needed, thus making its -W
Tom Lane [Tue, 11 Dec 2007 02:08:59 +0000 (02:08 +0000)]
Make vacuumlo prompt for password when needed, thus making its -W
switch optional, as is the case for every other one of our programs.
I had already documented its -W as being optional, so this is bringing
the code into line with the docs ...

16 years agoRemove Stefan Kaltenbrunner mention from full text release note item.
Bruce Momjian [Mon, 10 Dec 2007 23:38:43 +0000 (23:38 +0000)]
Remove Stefan Kaltenbrunner mention from full text release note item.

16 years agoAdd release note introduction.
Bruce Momjian [Mon, 10 Dec 2007 23:37:49 +0000 (23:37 +0000)]
Add release note introduction.

16 years agoMention Stefan Kaltenbrunner as contributor for text search because of
Bruce Momjian [Mon, 10 Dec 2007 22:14:52 +0000 (22:14 +0000)]
Mention Stefan Kaltenbrunner as contributor for text search because of
psql work.

16 years agoWording improvement for write-cache disabling instructions.
Bruce Momjian [Mon, 10 Dec 2007 22:08:36 +0000 (22:08 +0000)]
Wording improvement for write-cache disabling instructions.

16 years agoWrite cache doc cleanups
Bruce Momjian [Mon, 10 Dec 2007 14:51:10 +0000 (14:51 +0000)]
Write cache doc cleanups

Greg Smith

16 years agoDocument how to turn off disk write cache on popular operating systems.
Bruce Momjian [Mon, 10 Dec 2007 14:05:05 +0000 (14:05 +0000)]
Document how to turn off disk write cache on popular operating systems.

16 years agoMake an editorial pass over the newly SGML-ified contrib documentation.
Tom Lane [Mon, 10 Dec 2007 05:32:51 +0000 (05:32 +0000)]
Make an editorial pass over the newly SGML-ified contrib documentation.
Fix lots of bad markup, bad English, bad explanations.

Second round of commits.  pgcrypto and pgstandby still to go...

16 years agoChange the "distributed checkpoint" release note item to Greg Smith's proposed
Alvaro Herrera [Mon, 10 Dec 2007 01:48:54 +0000 (01:48 +0000)]
Change the "distributed checkpoint" release note item to Greg Smith's proposed
wording.

16 years agoUpdate Brazilian FAQ.
Bruce Momjian [Mon, 10 Dec 2007 00:34:06 +0000 (00:34 +0000)]
Update Brazilian FAQ.

Euler Taveira de Oliveira

16 years agoFAQ wording and markup fix.
Bruce Momjian [Mon, 10 Dec 2007 00:32:25 +0000 (00:32 +0000)]
FAQ wording and markup fix.

Euler Taveira de Oliveira

16 years agoFix attribution for Rime of the Ancient Mariner (obviously it's been
Tom Lane [Mon, 10 Dec 2007 00:12:31 +0000 (00:12 +0000)]
Fix attribution for Rime of the Ancient Mariner (obviously it's been
too long since freshman English :-()

16 years agoThe E. J. Pratt verse used as a tsearch test case is unfortunately still
Tom Lane [Sun, 9 Dec 2007 21:01:18 +0000 (21:01 +0000)]
The E. J. Pratt verse used as a tsearch test case is unfortunately still
under copyright in the US and many other places.  Substitute a little
something from a poet who's more safely dead.  Per gripe from Bjorn Munch.

16 years agoFix bogus prompt for password in -u case. Per gripe from Robert Treat.
Tom Lane [Sun, 9 Dec 2007 19:04:47 +0000 (19:04 +0000)]
Fix bogus prompt for password in -u case.  Per gripe from Robert Treat.

16 years agoFix up the PQconnectionUsedPassword mess: create a separate
Tom Lane [Sun, 9 Dec 2007 19:01:40 +0000 (19:01 +0000)]
Fix up the PQconnectionUsedPassword mess: create a separate
PQconnectionNeedsPassword function that tells the right thing for whether to
prompt for a password, and improve PQconnectionUsedPassword so that it checks
whether the password used by the connection was actually supplied as a
connection argument, instead of coming from environment or a password file.
Per bug report from Mark Cave-Ayland and subsequent discussion.

16 years agoFix completely-bogus volatility markings on pg_trgm functions.
Tom Lane [Sun, 9 Dec 2007 02:22:46 +0000 (02:22 +0000)]
Fix completely-bogus volatility markings on pg_trgm functions.

16 years agoFix mergejoin cost estimation so that we consider the statistical ranges of
Tom Lane [Sat, 8 Dec 2007 21:05:11 +0000 (21:05 +0000)]
Fix mergejoin cost estimation so that we consider the statistical ranges of
the two join variables at both ends: not only trailing rows that need not be
scanned because there cannot be a match on the other side, but initial rows
that will be scanned without possibly having a match.  This allows a more
realistic estimate of startup cost to be made, per recent pgsql-performance
discussion.  In passing, fix a couple of bugs that had crept into
mergejoinscansel: it was not quite up to speed for the task of estimating
descending-order scans, which is a new requirement in 8.3.

16 years agoIn the release notes, mention that the "Overview" items are listed in
Bruce Momjian [Sat, 8 Dec 2007 17:24:03 +0000 (17:24 +0000)]
In the release notes, mention that the "Overview" items are listed in
the sections below in more detail.

16 years agoIn release notes, mention "most" updates are improved with HOT.
Bruce Momjian [Fri, 7 Dec 2007 21:22:34 +0000 (21:22 +0000)]
In release notes, mention "most" updates are improved with HOT.

16 years agoRemove completed 8.3 TODO items:
Bruce Momjian [Fri, 7 Dec 2007 17:52:15 +0000 (17:52 +0000)]
Remove completed 8.3 TODO items:

<  o -Allow commenting of variables in postgresql.conf to restore them
<    to defaults
<  o -Add a GUC variable to control the tablespace for temporary objects
<    and sort files
< Monitoring
< ==========
<
< * -Allow server log information to be output as CSV format
< * -Add ability to monitor the use of temporary sort files
< * -Allow user-defined types to accept 'typmod' parameters
<
<   http://archives.postgresql.org/pgsql-hackers/2005-08/msg01142.php
<   http://archives.postgresql.org/pgsql-hackers/2005-09/msg00012.php
<   http://archives.postgresql.org/pgsql-hackers/2006-08/msg00149.php
<
< * -Add Globally/Universally Unique Identifier (GUID/UUID)
<
<   http://archives.postgresql.org/pgsql-patches/2006-09/msg00209.php
<   http://archives.postgresql.org/pgsql-general/2007-01/msg00853.php
<
< * -Support a data type with specific enumerated values (ENUM)
<  o -Add support for arrays of complex types
<  o -Make 64-bit version of the MONEY data type
< * -Add ISO day of week format 'ID' to to_char() where Monday = 1
< * -Add a field 'isoyear' to extract(), based on the ISO week
< * -Add RESET SESSION command to reset all session state
<  o -Make CLUSTER preserve recently-dead tuples per MVCC requirements
<  o -Add more logical syntax CLUSTER table USING index;
<    support current syntax for backward compatibility
<  o -Allow UPDATE/DELETE WHERE CURRENT OF cursor
<  o -Add support for MOVE cursors
<  o -Allow PL/PythonU to return boolean rather than 1/0
<  o -Allow psql \pset boolean variables to set to fixed values, rather
<    than toggle
<  o -Add -f to pg_dumpall
< Dependency Checking
< ===================
<
< * -Flush cached query plans when the dependent objects change or
<   when new ANALYZE statistics are available
< * -Track dependencies in function bodies and recompile/invalidate
< * -Invalidate prepared queries, like INSERT, when the table definition
<   is altered
<
< * -Allow use of indexes to search for NULLs
< * -Allow the creation of indexes with mixed ascending/descending
<   specifiers
< * -Reduce checkpoint performance degredation by forcing data to disk
<   more evenly
< * -Allow sequential scans to take advantage of other concurrent
<   sequential scans, also called "Synchronised Scanning"
< * -Consider shrinking expired tuples to just their headers
< * -Allow heap reuse of UPDATEd rows if no indexed columns are changed,
<   and old and new versions are on the same heap page
< * -Reduce XID consumption of read-only queries
<  o -Turn on by default
<  o -Allow multiple vacuums so large tables do not starve small
<    tables
< * -Allow the pg_xlog directory location to be specified during initdb
<   with a symlink back to the /data location
< * -Allow buffered WAL writes and fsync
< * -Allow ORDER BY ... LIMIT # to select high/low value without sort or
<   index using a sequential scan for highest/lowest values
< * -Merge xmin/xmax/cmin/cmax back into three header fields
<  o -Support a smaller header for short variable-length fields
< * -Move NAMEDATALEN from postgres_ext.h to pg_config_manual.h
< * -Fix problem with excessive logging during SSL disconnection
<
<   http://archives.postgresql.org/pgsql-bugs/2006-12/msg00122.php
<   http://archives.postgresql.org/pgsql-bugs/2007-05/msg00065.php
<
<  o -Add long file support for binary pg_dump output

16 years agoUpdate HOT description to make index column changes clearer.
Bruce Momjian [Fri, 7 Dec 2007 17:42:49 +0000 (17:42 +0000)]
Update HOT description to make index column changes clearer.

16 years agoChange documentation, change "distributed" checkpoints to "smoothed"
Bruce Momjian [Fri, 7 Dec 2007 16:53:31 +0000 (16:53 +0000)]
Change documentation, change "distributed" checkpoints to "smoothed"
checkpoints.

16 years agoClean up a couple of problems in crosstab_hash's use of a hash table.
Tom Lane [Fri, 7 Dec 2007 16:44:58 +0000 (16:44 +0000)]
Clean up a couple of problems in crosstab_hash's use of a hash table.
The original coding leaked memory (at least 8K per crosstab_hash call)
because it allowed the hash table to be allocated as a child of
TopMemoryContext and then never freed it.  Fix that by putting the
hash table under per_query_ctx, instead.  Also get rid of use
of a static variable to point to the hash table.  Aside from being
ugly, that would actively do the wrong thing in the case of re-entrant
calls to crosstab_hash, which are at least theoretically possible
since it was expecting the static variable to stay valid across
a SPI_execute call.

16 years agoRealign the running text in this file to 79 characters wide. Some other
Peter Eisentraut [Fri, 7 Dec 2007 16:44:56 +0000 (16:44 +0000)]
Realign the running text in this file to 79 characters wide.  Some other
copy-editing.

16 years agoFix typo in pg_standby docs. Arul Shaji.
Neil Conway [Fri, 7 Dec 2007 08:25:38 +0000 (08:25 +0000)]
Fix typo in pg_standby docs. Arul Shaji.

16 years agoAdd missing table entry for lastval(), per Jeff Davis.
Tom Lane [Thu, 6 Dec 2007 17:47:25 +0000 (17:47 +0000)]
Add missing table entry for lastval(), per Jeff Davis.

16 years agoDisable asserts by default on msvc build (to make it the same
Magnus Hagander [Thu, 6 Dec 2007 17:07:37 +0000 (17:07 +0000)]
Disable asserts by default on msvc build (to make it the same
as in unix)

16 years agoChange wording of logged message when cancelling an autovacuum task, using
Alvaro Herrera [Thu, 6 Dec 2007 14:32:54 +0000 (14:32 +0000)]
Change wording of logged message when cancelling an autovacuum task, using
american speling (unlike this commit message).  Per complaint from Mike C. on
bug #3790 and subsequent discussion.

16 years agoMake an editorial pass over the newly SGML-ified contrib documentation.
Tom Lane [Thu, 6 Dec 2007 04:12:10 +0000 (04:12 +0000)]
Make an editorial pass over the newly SGML-ified contrib documentation.
Fix lots of bad markup, bad English, bad explanations.

This commit covers only about half the contrib modules, but I grow weary...

16 years agoRemove unused, occasionally confusing graphics building support.
Peter Eisentraut [Wed, 5 Dec 2007 10:06:08 +0000 (10:06 +0000)]
Remove unused, occasionally confusing graphics building support.

16 years agoCorrect misattribution of earthdistance module --- according to its
Tom Lane [Tue, 4 Dec 2007 23:45:01 +0000 (23:45 +0000)]
Correct misattribution of earthdistance module --- according to its
old README file, Bruno Wolff did the bulk of the work, but for some
reason only Hal Snyder was credited here.

16 years agoDon't send an empty SSPI negotiation packet at the end of the negotiation.
Magnus Hagander [Tue, 4 Dec 2007 13:02:53 +0000 (13:02 +0000)]
Don't send an empty SSPI negotiation packet at the end of the negotiation.

Fixes bug #3750

16 years agoEntity-ify a passel of & < > characters. Per gripe from Devrim.
Tom Lane [Mon, 3 Dec 2007 23:49:51 +0000 (23:49 +0000)]
Entity-ify a passel of & < > characters.  Per gripe from Devrim.

16 years agoFix build_minmax_path() to cope if an IS NULL clause turns up in the
Tom Lane [Mon, 3 Dec 2007 22:37:17 +0000 (22:37 +0000)]
Fix build_minmax_path() to cope if an IS NULL clause turns up in the
indexable-clauses list for a btree index.  Formerly it just Asserted that
all such clauses were opclauses, but that's no longer true in 8.3.
Per bug #3796 from Matthias Schoeneich.

16 years agoRemove contrib/spi/preprocessor, which has been uninteresting ever
Tom Lane [Mon, 3 Dec 2007 19:20:40 +0000 (19:20 +0000)]
Remove contrib/spi/preprocessor, which has been uninteresting ever
since we supported standard FOREIGN KEY constraint syntax.  It was
harmless enough just sitting there, but the prospect of having to
document it is surely more work than it's worth.

16 years agoDefault to integer timestamps being off until 8.4, to correspond
Magnus Hagander [Mon, 3 Dec 2007 17:17:05 +0000 (17:17 +0000)]
Default to integer timestamps being off until 8.4, to correspond
to the settings used for other platforms.

16 years agoDon't try to install README.spi
Magnus Hagander [Mon, 3 Dec 2007 15:42:58 +0000 (15:42 +0000)]
Don't try to install README.spi

Hannes Eder

16 years agoAdd missing documentation for SSPI packets.
Magnus Hagander [Mon, 3 Dec 2007 13:40:11 +0000 (13:40 +0000)]
Add missing documentation for SSPI packets.

16 years agoImprove partitioning example, per Itagaki Takahiro.
Tom Lane [Mon, 3 Dec 2007 04:59:55 +0000 (04:59 +0000)]
Improve partitioning example, per Itagaki Takahiro.

16 years agoRemove obsoleted README files.
Tom Lane [Mon, 3 Dec 2007 04:22:54 +0000 (04:22 +0000)]
Remove obsoleted README files.

16 years agoAdd SGML documentation for contrib/spi and contrib/test_parser.
Tom Lane [Mon, 3 Dec 2007 04:18:47 +0000 (04:18 +0000)]
Add SGML documentation for contrib/spi and contrib/test_parser.
The spi documentation is pretty rudimentary, but it's a start.

16 years agoUpdate SGML contrib documentation to catch up with recent changes to
Tom Lane [Mon, 3 Dec 2007 03:05:47 +0000 (03:05 +0000)]
Update SGML contrib documentation to catch up with recent changes to
the former README files.  Albert Cervera i Areny

16 years agoStamp 8.3beta4.
Tom Lane [Mon, 3 Dec 2007 00:11:01 +0000 (00:11 +0000)]
Stamp 8.3beta4.

16 years agoRevert COPY OUT to follow the pre-8.3 handling of ASCII control characters,
Tom Lane [Mon, 3 Dec 2007 00:03:05 +0000 (00:03 +0000)]
Revert COPY OUT to follow the pre-8.3 handling of ASCII control characters,
namely that \r, \n, \t, \b, \f, \v are dumped as those two-character
representations rather than a backslash and the literal control character.
I had made it do the other to save some code, but this was ill-advised,
because dump files in which these characters appear literally are prone to
newline mangling.  Fortunately, doing it the old way should only cost a few
more lines of code, and not slow down the copy loop materially.
Per bug #3795 from Lou Duchez.

16 years agoEnsure that all <sect1> and <refentry> tags have IDs. This is needed
Tom Lane [Sun, 2 Dec 2007 22:33:20 +0000 (22:33 +0000)]
Ensure that all <sect1> and <refentry> tags have IDs.  This is needed
to ensure that the resulting webpages have predictable URLs, instead of
ever-changing numeric IDs.  The new contrib docs were the biggest
offender, but some old stuff had the problem too.  Also, rename a couple
of new contrib sgml files for consistency's sake.

16 years agoRemove README files now migrated to SGML docs.
Tom Lane [Sun, 2 Dec 2007 21:15:38 +0000 (21:15 +0000)]
Remove README files now migrated to SGML docs.

16 years agoAdd SGML docs for contrib/dict_int and contrib/dict_xsyn.
Tom Lane [Sun, 2 Dec 2007 21:13:34 +0000 (21:13 +0000)]
Add SGML docs for contrib/dict_int and contrib/dict_xsyn.
Albert Cervera i Areny

16 years agoImprove the manual's discussion of partitioning. Recommend using a
Tom Lane [Sun, 2 Dec 2007 19:20:32 +0000 (19:20 +0000)]
Improve the manual's discussion of partitioning.  Recommend using a
trigger instead of a rule to redirect insertions, use NEW.* notation
where appropriate, some other updates and adjustments.  David Fetter
and Tom Lane

16 years agoMention hash opclasses in 'System Dependencies on Operator Classes',
Tom Lane [Sun, 2 Dec 2007 04:36:40 +0000 (04:36 +0000)]
Mention hash opclasses in 'System Dependencies on Operator Classes',
which previously only talked about btree opclasses.

16 years agoSync release notes with CVS HEAD; various editorial improvements.
Tom Lane [Sun, 2 Dec 2007 03:47:11 +0000 (03:47 +0000)]
Sync release notes with CVS HEAD; various editorial improvements.

16 years agoCode review for LIKE ... INCLUDING INDEXES patch. Fix failure to propagate
Tom Lane [Sat, 1 Dec 2007 23:44:44 +0000 (23:44 +0000)]
Code review for LIKE ... INCLUDING INDEXES patch.  Fix failure to propagate
constraint status of copied indexes (bug #3774), as well as various other
small bugs such as failure to pstrdup when needed.  Allow INCLUDING INDEXES
indexes to be merged with identical declared indexes (perhaps not real useful,
but the code is there and having it not apply to LIKE indexes seems pretty
unorthogonal).  Avoid useless work in generateClonedIndexStmt().  Undo some
poorly chosen API changes, and put a couple of routines in modules that seem
to be better places for them.

16 years agoSuppress compiler warnings in recent plperl patch. Avoid uselessly expensive
Tom Lane [Sat, 1 Dec 2007 17:58:42 +0000 (17:58 +0000)]
Suppress compiler warnings in recent plperl patch.  Avoid uselessly expensive
lookup of the well-known OID of textout().

16 years agoRevert inadvertantly committed change.
Andrew Dunstan [Sat, 1 Dec 2007 15:30:09 +0000 (15:30 +0000)]
Revert inadvertantly committed change.

16 years agoWorkaround for perl problem where evaluating UTF8 regexes can cause
Andrew Dunstan [Sat, 1 Dec 2007 15:20:34 +0000 (15:20 +0000)]
Workaround for perl problem where evaluating UTF8 regexes can cause
implicit loading of modules, thereby breaking Safe rules.
We compile and call a tiny perl function on trusted interpreter init, after which
the problem does not occur.

16 years agoAvoid incrementing the CommandCounter when CommandCounterIncrement is called
Tom Lane [Fri, 30 Nov 2007 21:22:54 +0000 (21:22 +0000)]
Avoid incrementing the CommandCounter when CommandCounterIncrement is called
but no database changes have been made since the last CommandCounterIncrement.
This should result in a significant improvement in the number of "commands"
that can typically be performed within a transaction before hitting the 2^32
CommandId size limit.  In particular this buys back (and more) the possible
adverse consequences of my previous patch to fix plan caching behavior.

The implementation requires tracking whether the current CommandCounter
value has been "used" to mark any tuples.  CommandCounter values stored into
snapshots are presumed not to be used for this purpose.  This requires some
small executor changes, since the executor used to conflate the curcid of
the snapshot it was using with the command ID to mark output tuples with.
Separating these concepts allows some small simplifications in executor APIs.

Something for the TODO list: look into having CommandCounterIncrement not do
AcceptInvalidationMessages.  It seems fairly bogus to be doing it there,
but exactly where to do it instead isn't clear, and I'm disinclined to mess
with asynchronous behavior during late beta.

16 years agoRepair bug that allowed RevalidateCachedPlan to attempt to rebuild a cached
Tom Lane [Fri, 30 Nov 2007 18:38:34 +0000 (18:38 +0000)]
Repair bug that allowed RevalidateCachedPlan to attempt to rebuild a cached
plan before the effects of DDL executed in an immediately prior SPI operation
had been absorbed.  Per report from Chris Wood.

This patch has an unpleasant side effect of causing the number of
CommandCounterIncrement()s done by a typical plpgsql function to
approximately double.  Amelioration of the consequences of that
will be undertaken in a separate patch.

16 years agoAdd win32error.c to libpq, needed to resolve _dosmaperr.
Magnus Hagander [Fri, 30 Nov 2007 16:21:19 +0000 (16:21 +0000)]
Add win32error.c to libpq, needed to resolve _dosmaperr.

16 years agoUse _dosmaperr() to deal with errors opening files in pgwin32_open().
Magnus Hagander [Fri, 30 Nov 2007 11:16:43 +0000 (11:16 +0000)]
Use _dosmaperr() to deal with errors opening files in pgwin32_open().

Per complaint from Alvaro and subsequent discussion.

16 years agoUpdate release notes to current CVS.
Bruce Momjian [Fri, 30 Nov 2007 03:27:48 +0000 (03:27 +0000)]
Update release notes to current CVS.

16 years agoUpdate Japanese FAQ.
Bruce Momjian [Thu, 29 Nov 2007 17:09:21 +0000 (17:09 +0000)]
Update Japanese FAQ.

Jun Kuwamura

16 years agoTranslation updates
Peter Eisentraut [Thu, 29 Nov 2007 08:50:36 +0000 (08:50 +0000)]
Translation updates

16 years agoAdd additional PITR documentation.
Bruce Momjian [Wed, 28 Nov 2007 22:35:54 +0000 (22:35 +0000)]
Add additional PITR documentation.

Simon Riggs

16 years agoMake a cleanup pass over error reports in tsearch code. Use ereport
Tom Lane [Wed, 28 Nov 2007 21:56:30 +0000 (21:56 +0000)]
Make a cleanup pass over error reports in tsearch code.  Use ereport
for user-facing errors, fix some poor choices of errcode, adhere to
message style guide.

16 years agoRETURN NEXT doc improvement, per Alvaro.
Bruce Momjian [Wed, 28 Nov 2007 21:47:39 +0000 (21:47 +0000)]
RETURN NEXT doc improvement, per Alvaro.

16 years agoUpdate RETURN NEXT documentation for plpgsql.
Bruce Momjian [Wed, 28 Nov 2007 20:56:35 +0000 (20:56 +0000)]
Update RETURN NEXT documentation for plpgsql.

16 years agoImprove test coverage of CLOBBER_CACHE_ALWAYS by having it also force
Tom Lane [Wed, 28 Nov 2007 20:44:26 +0000 (20:44 +0000)]
Improve test coverage of CLOBBER_CACHE_ALWAYS by having it also force
reloading of operator class information on each use of LookupOpclassInfo.
Had this been in place a year ago, it would have helped me find a bug
in the then-new 'operator family' code.  Now that we have a build farm
member testing CLOBBER_CACHE_ALWAYS on a regular basis, it seems worth
expending a little bit of effort here.

16 years agoPlpgsql doc cleanup.
Bruce Momjian [Wed, 28 Nov 2007 20:13:06 +0000 (20:13 +0000)]
Plpgsql doc cleanup.

Guillaume Lelarge

16 years agoAdjust the names of a couple of tsearch index support functions that had
Tom Lane [Wed, 28 Nov 2007 19:33:05 +0000 (19:33 +0000)]
Adjust the names of a couple of tsearch index support functions that had
inappropriately generic-sounding names.  This is more or less free since
we already forced initdb for the next beta, and it may prevent confusion or
name conflicts (particularly at the C-global-symbol level) down the road.
Per my proposal yesterday.

16 years agoInstall a lookaside cache to speed up repeated lookups of the same operator
Tom Lane [Wed, 28 Nov 2007 18:47:56 +0000 (18:47 +0000)]
Install a lookaside cache to speed up repeated lookups of the same operator
by short-circuiting schema search path and ambiguous-operator resolution
computations.  Remarkably, this buys as much as 45% speedup of repetitive
simple queries that involve operators that are not an exact match to the
input datatypes.  It should be marginally faster even for exact-match
cases, though I've not had success in proving an improvement in benchmark
tests.  Per report from Guillame Smet and subsequent discussion.

16 years agoProper capitalization of Ispell
Peter Eisentraut [Wed, 28 Nov 2007 15:42:46 +0000 (15:42 +0000)]
Proper capitalization of Ispell

16 years agospell checker run
Peter Eisentraut [Wed, 28 Nov 2007 15:42:31 +0000 (15:42 +0000)]
spell checker run