OSDN Git Service

pg-rex/syncrep.git
16 years agoMake casts from xml to text independent of the XML option setting, thus
Peter Eisentraut [Tue, 27 Nov 2007 12:21:05 +0000 (12:21 +0000)]
Make casts from xml to text independent of the XML option setting, thus
immutable and indexable.  Also fix the volatility settings of some other
XML-related functions.

16 years agocorrect capitalization
Peter Eisentraut [Tue, 27 Nov 2007 12:17:27 +0000 (12:17 +0000)]
correct capitalization

16 years agoWording improvement for MONEY casting.
Bruce Momjian [Tue, 27 Nov 2007 06:05:57 +0000 (06:05 +0000)]
Wording improvement for MONEY casting.

16 years agoAdd documentation about the issues of casting MONEY to/from numeric
Bruce Momjian [Tue, 27 Nov 2007 05:49:58 +0000 (05:49 +0000)]
Add documentation about the issues of casting MONEY to/from numeric
types.

16 years agoAdd:
Bruce Momjian [Mon, 26 Nov 2007 23:31:09 +0000 (23:31 +0000)]
Add:

< * -Make 64-bit version of the MONEY data type
< * Add locale-aware MONEY type, and support multiple currencies
<
<   http://archives.postgresql.org/pgsql-general/2005-08/msg01432.php
<   http://archives.postgresql.org/pgsql-hackers/2007-03/msg01181.php
<
< * MONEY dumps in a locale-specific format making it difficult to
<   restore to a system with a different locale
>
> * MONEY data type
>
>  o -Make 64-bit version of the MONEY data type
>  * Add locale-aware MONEY type, and support multiple currencies
>
>    http://archives.postgresql.org/pgsql-general/2005-08/msg01432.php
>    http://archives.postgresql.org/pgsql-hackers/2007-03/msg01181.php
>
>  * MONEY dumps in a locale-specific format making it difficult to
>    restore to a system with a different locale
>  * Allow MONEY to be cast to/from other numeric data types

16 years agoChange index_name to name for consistency.
Peter Eisentraut [Mon, 26 Nov 2007 21:36:33 +0000 (21:36 +0000)]
Change index_name to name for consistency.

16 years agoFix select_common_type() so that it can select a domain type, if all inputs
Tom Lane [Mon, 26 Nov 2007 16:46:51 +0000 (16:46 +0000)]
Fix select_common_type() so that it can select a domain type, if all inputs
to a UNION, CASE, or related construct are of the same domain type.  The
main part of this routine smashes domains to their base types, which seems
necessary because the logic involves TypeCategory() and IsPreferredType(),
neither of which work usefully on domains.  However, we can add a first
pass that just detects whether all the inputs are exactly the same type,
and if so accept that without question (so long as it's not UNKNOWN).
Per recent gripe from Dean Rasheed.

In passing, remove some tests for InvalidOid, which have clearly been dead
code for quite some time now, because getBaseType() would fail on that input.

Also, clarify the manual's not-very-precise description of the existing
algorithm's behavior.

16 years agoRequire a specific Autoconf version, instead of a lower bound only.
Peter Eisentraut [Mon, 26 Nov 2007 12:31:07 +0000 (12:31 +0000)]
Require a specific Autoconf version, instead of a lower bound only.

16 years agoAllow for X as well as x to be the prefix for hexadecimal character ref entity numbers,
Andrew Dunstan [Sun, 25 Nov 2007 19:35:41 +0000 (19:35 +0000)]
Allow for X as well as x to be the prefix for hexadecimal character ref entity numbers,
as in HTML.

16 years agoFix XML tag namespace change inadvertantly missed from previous fix. Add
Andrew Dunstan [Sun, 25 Nov 2007 15:37:11 +0000 (15:37 +0000)]
Fix XML tag namespace change inadvertantly missed from previous fix. Add
regression test for XML names and numeric entities.

16 years agoUse double quotes for quoting xml attributes.
Peter Eisentraut [Sun, 25 Nov 2007 12:08:11 +0000 (12:08 +0000)]
Use double quotes for quoting xml attributes.

16 years agoFix thinko in comment.
Tom Lane [Sun, 25 Nov 2007 02:09:46 +0000 (02:09 +0000)]
Fix thinko in comment.

16 years agoReduce routine successful-archiving message from LOG to DEBUG1 grade,
Tom Lane [Sat, 24 Nov 2007 21:37:04 +0000 (21:37 +0000)]
Reduce routine successful-archiving message from LOG to DEBUG1 grade,
to reduce chattiness of logs.  Per discussion.
Simon Riggs

16 years agoFix (I think) broken usage of MultiByteToWideChar. I had missed the
Tom Lane [Sat, 24 Nov 2007 21:20:07 +0000 (21:20 +0000)]
Fix (I think) broken usage of MultiByteToWideChar.  I had missed the
subtlety that this function only returns a null terminator if it's
fed input that includes one; which, in the usage here, it's not.
This probably fixes bugs reported by Thomas Haegi.

16 years agoSlightly more paranoia and slightly better comments for use of
Tom Lane [Sat, 24 Nov 2007 21:16:55 +0000 (21:16 +0000)]
Slightly more paranoia and slightly better comments for use of
Windows-specific MultiByteToWideChar/WideCharToMultiByte calls.

16 years agoAdd missed entry for new money regression test.
Tom Lane [Sat, 24 Nov 2007 20:41:35 +0000 (20:41 +0000)]
Add missed entry for new money regression test.

16 years agoMove resetting of currSchema to a spot where it's not vulnerable to getting
Tom Lane [Sat, 24 Nov 2007 20:26:49 +0000 (20:26 +0000)]
Move resetting of currSchema to a spot where it's not vulnerable to getting
missed in possible future partial-drop scenarios.  Improve comment.
Per report from Zoltan Boszormenyi.

16 years agoAdd regression tests for MONEY type.
D'Arcy J.M. Cain [Sat, 24 Nov 2007 19:49:23 +0000 (19:49 +0000)]
Add regression tests for MONEY type.

16 years agoSave another little bit of planner overhead on simple queries, by having
Tom Lane [Sat, 24 Nov 2007 19:08:51 +0000 (19:08 +0000)]
Save another little bit of planner overhead on simple queries, by having
clauselist_selectivity skip some analysis that's useless when there's only
one clause in the given list.  Actually this can win even for not-so-simple
queries, because we also apply clauselist_selectivity to sublists such as the
quals matching an index; which are likely to have only a single entry even
when the total query is quite complicated.

16 years agoIf pg_dump drops a schema, make sure 'search_path' it SET the next time
Bruce Momjian [Sat, 24 Nov 2007 17:45:32 +0000 (17:45 +0000)]
If pg_dump drops a schema, make sure 'search_path' it SET the next time
it is needed --- basically don't remember the previous 'search_path'
after a drop schema because it might have failed but be created later.

16 years agoClarify how MONEY trims off trailing thousands separator.
Bruce Momjian [Sat, 24 Nov 2007 16:18:48 +0000 (16:18 +0000)]
Clarify how MONEY trims off trailing thousands separator.

16 years agoMake the MONEY data type have a thousands separator != decimal symbol,
Bruce Momjian [Sat, 24 Nov 2007 15:28:02 +0000 (15:28 +0000)]
Make the MONEY data type have a thousands separator != decimal symbol,
if the locale has the thousands separator as "".  This now matches the
to_char and psql numericlocale behavior.  (Previously this data type was
basically useless for such setups.)

16 years agoBorland BCC does not support SSPI, per cnliou9@fastmail.fm.
Bruce Momjian [Sat, 24 Nov 2007 01:55:26 +0000 (01:55 +0000)]
Borland BCC does not support SSPI, per cnliou9@fastmail.fm.

16 years agoBorland CC 5.5.1 needs ssize_t, per cnliou9@fastmail.fm.
Bruce Momjian [Sat, 24 Nov 2007 01:32:48 +0000 (01:32 +0000)]
Borland CC 5.5.1 needs ssize_t, per cnliou9@fastmail.fm.

16 years agoChange fix_scan_expr() to avoid copying the input node tree in the common case
Tom Lane [Sat, 24 Nov 2007 00:39:44 +0000 (00:39 +0000)]
Change fix_scan_expr() to avoid copying the input node tree in the common case
where rtoffset == 0.  In that case there is no need to change Var nodes,
and since filling in unset opfuncid fields is always safe, scribbling on the
input tree to that extent is not objectionable.  This brings the cost of this
operation back down to what it was in 8.2 for simple queries.  Per
investigation of performance gripe from Guillaume Smet.

16 years agoAvoid uselessly building a duplicate of the original clause in trivial cases
Tom Lane [Fri, 23 Nov 2007 19:57:44 +0000 (19:57 +0000)]
Avoid uselessly building a duplicate of the original clause in trivial cases
where the EquivalenceClass machinery is unable to deduce anything more from a
simple "var = const" qual clause.  There are probably some more cases where
this could be done, but this seems to take care of most of the added overhead
for simple queries.  Per gripe from Guillaume Smet.

In passing, fix a problem that was exposed by this change:
reconsider_outer_join_clause and friends were passing the wrong relids to
build_implied_join_equality, resulting in RestrictInfos with the wrong
required_relids.  This mistake was masked in typical cases since the bogus
RestrictInfos would never have escaped from the EquivalenceClass machinery,
but I think there might be corner cases involving "broken" ECs where there
would have been a visible failure even without the new optimization.  In any
case the code was certainly not operating as intended.

16 years agoFix white space in MONEY type code. Rename 'comma' to more generic
Bruce Momjian [Fri, 23 Nov 2007 19:54:39 +0000 (19:54 +0000)]
Fix white space in MONEY type code.  Rename 'comma' to more generic
'ssymbol' as used in previous function.

16 years agoUpdate text:
Bruce Momjian [Fri, 23 Nov 2007 16:37:55 +0000 (16:37 +0000)]
Update text:

< * Prevent long-lived temporary tables from causing frozen-Xid advancement
> * Prevent long-lived temporary tables from causing frozen-xid advancement
>
>    The problem is that autovacuum cannot vacuum them to set frozen xids;
>    only the session that created them can do that.
>
>
>

16 years agoFix buggy usage of vsnprintf in PL/Python by removing it altogether, instead
Alvaro Herrera [Fri, 23 Nov 2007 01:46:34 +0000 (01:46 +0000)]
Fix buggy usage of vsnprintf in PL/Python by removing it altogether, instead
relying on stringinfo.c.  This fixes a problem reported by Marko Kreen, but I
didn't use his patch, per subsequent discussion.

16 years agoFix "Overall Page Layout" table. The second row should be ItemIdData, not
Tatsuo Ishii [Fri, 23 Nov 2007 00:24:12 +0000 (00:24 +0000)]
Fix "Overall Page Layout" table. The second row should be ItemIdData, not
ItemPointerData.

16 years agoActually ... it's pretty silly that parse_oper.c doesn't set up the
Tom Lane [Thu, 22 Nov 2007 19:40:25 +0000 (19:40 +0000)]
Actually ... it's pretty silly that parse_oper.c doesn't set up the
opfuncid of an OpExpr initially, considering that it has the information
at hand already.  We'll still treat opfuncid as a cache rather than a
guaranteed-valid value, but this change saves one more syscache lookup
in the normal code path.

16 years agoSave one syscache lookup when examining volatility or strictness of
Tom Lane [Thu, 22 Nov 2007 19:09:23 +0000 (19:09 +0000)]
Save one syscache lookup when examining volatility or strictness of
OpExpr and related nodes.  We're going to have to set the opfuncid of
such nodes eventually (if we haven't already), so we might as well
exploit the opportunity to cache the function OID.  Buys back some
of the extra planner overhead noted by Guillaume Smet, though I still
need to fool with equivclass.c to really respond to that.

16 years agoAdd more comments about thousands separator handling.
Bruce Momjian [Thu, 22 Nov 2007 17:51:39 +0000 (17:51 +0000)]
Add more comments about thousands separator handling.

16 years agoPrevent Perl from introducing a possibly-incompatible definition of type
Tom Lane [Thu, 22 Nov 2007 17:47:28 +0000 (17:47 +0000)]
Prevent Perl from introducing a possibly-incompatible definition of type
"bool" into plperl.c.  This has always been a hazard since Perl allows a
platform-specific choice to define bool as int rather than char, but
evidently this didn't happen on any platform we support ... until OS X 10.5.
Per report from Brandon Maust.

Back-patch as far as 8.0 --- a bit arbitrary, but it seems unlikely anyone
will be trying to port 7.x onto new platforms.

16 years agoAdd comments about thousands separator logic.
Bruce Momjian [Thu, 22 Nov 2007 15:10:05 +0000 (15:10 +0000)]
Add comments about thousands separator logic.

16 years agoDocument that doc builds now build indexes by default, unless "draft" is
Bruce Momjian [Thu, 22 Nov 2007 15:04:52 +0000 (15:04 +0000)]
Document that doc builds now build indexes by default, unless "draft" is
used.

16 years agoWSATYPE_NOT_FOUND was already defined for BCC so don't redefine it
Bruce Momjian [Wed, 21 Nov 2007 23:13:36 +0000 (23:13 +0000)]
WSATYPE_NOT_FOUND was already defined for BCC so don't redefine it
(conflicting values).

16 years agoFix build of libpq for Borland CC.
Bruce Momjian [Wed, 21 Nov 2007 23:03:44 +0000 (23:03 +0000)]
Fix build of libpq for Borland CC.

16 years agoWhen setting default thousands separator when locale has "", use logic
Bruce Momjian [Wed, 21 Nov 2007 22:28:18 +0000 (22:28 +0000)]
When setting default thousands separator when locale has "", use logic
so new thousands separator doesn't match decimal symbol.

16 years agoFix typo in comment.
Bruce Momjian [Wed, 21 Nov 2007 21:49:22 +0000 (21:49 +0000)]
Fix typo in comment.

16 years agoClarify the syntax rules for raw tsvector input.
Tom Lane [Wed, 21 Nov 2007 04:01:37 +0000 (04:01 +0000)]
Clarify the syntax rules for raw tsvector input.

16 years agoFix bogus length calculation that could lead to crash if the string
Tom Lane [Tue, 20 Nov 2007 23:14:41 +0000 (23:14 +0000)]
Fix bogus length calculation that could lead to crash if the string
happened to be right up against the end of memory, per report from
Matt Magoffin.  While at it, avoid useless multiple copying of string
by not depending on xmlStrncatNew.

16 years agoFix pg_ctl -t to register -t value, per Alvaro.
Bruce Momjian [Tue, 20 Nov 2007 19:24:26 +0000 (19:24 +0000)]
Fix pg_ctl -t to register -t value, per Alvaro.

16 years agoSplit release notes up into smaller sections; easier to read.
Bruce Momjian [Tue, 20 Nov 2007 17:10:25 +0000 (17:10 +0000)]
Split release notes up into smaller sections; easier to read.

16 years agoFix some missed usages of 'HTML tag' and 'HTML entity'.
Tom Lane [Tue, 20 Nov 2007 15:58:52 +0000 (15:58 +0000)]
Fix some missed usages of 'HTML tag' and 'HTML entity'.

16 years agoMake "contrib/" tag consistent.
Bruce Momjian [Tue, 20 Nov 2007 05:23:20 +0000 (05:23 +0000)]
Make "contrib/" tag consistent.

16 years agoSplit out data types and function entries into separate sections in the
Bruce Momjian [Tue, 20 Nov 2007 05:16:31 +0000 (05:16 +0000)]
Split out data types and function entries into separate sections in the
release notes.

16 years agoRelease note adjustments suggested by IRC users.
Bruce Momjian [Tue, 20 Nov 2007 05:00:24 +0000 (05:00 +0000)]
Release note adjustments suggested by IRC users.

16 years agoAdjust contrib/tsearch2 regression results to use XML tag and XML entity descriptions...
Andrew Dunstan [Tue, 20 Nov 2007 04:23:10 +0000 (04:23 +0000)]
Adjust contrib/tsearch2 regression results to use XML tag and XML entity descriptions, as now used by core text search default parser.

16 years agoReorder a few release notes items for consistency.
Bruce Momjian [Tue, 20 Nov 2007 04:00:27 +0000 (04:00 +0000)]
Reorder a few release notes items for consistency.

16 years agoTrim down release section headings (no need for Changes).
Bruce Momjian [Tue, 20 Nov 2007 03:34:14 +0000 (03:34 +0000)]
Trim down release section headings (no need for Changes).

Re-order some items as migration changes.

16 years agoChange descriptions of entity and tag objects to "XML entity" and "XML tag".
Andrew Dunstan [Tue, 20 Nov 2007 02:25:22 +0000 (02:25 +0000)]
Change descriptions of entity and tag objects to "XML entity" and "XML tag".
Allow tag and entity names that follow XML rules. Provide for hexadecimal
as well as decimal numeric entities. Adjust code names to coincide with
new descriptions.

16 years agoSplit "server changes" up into smaller sections in release notes.
Bruce Momjian [Tue, 20 Nov 2007 01:42:37 +0000 (01:42 +0000)]
Split "server changes" up into smaller sections in release notes.

16 years agoMarkup fix.
Bruce Momjian [Tue, 20 Nov 2007 01:19:30 +0000 (01:19 +0000)]
Markup fix.

16 years agoMore release notes wording cleanups.
Bruce Momjian [Tue, 20 Nov 2007 01:13:00 +0000 (01:13 +0000)]
More release notes wording cleanups.

16 years agoUpdate COMMENT item:
Bruce Momjian [Tue, 20 Nov 2007 00:18:46 +0000 (00:18 +0000)]
Update COMMENT item:

<  o Prevent COMMENT ON dbname from issuing a warning when loading
<    into a database with a different name, perhaps using COMMENT ON
<    CURRENT DATABASE
>  o Change pg_dump so that a comment on the dumped database is
>    applied to the loaded database, even if the database has a
>    different name.  This will require new backend syntax, perhaps
>    COMMENT ON CURRENT DATABASE.

16 years agoUpdate wording:
Bruce Momjian [Mon, 19 Nov 2007 23:48:46 +0000 (23:48 +0000)]
Update wording:

<  o Allow COMMENT ON dbname to work when loading into a database
<    with a different name, perhaps using COMMENT ON CURRENT
<           DATABASE
>  o Prevent COMMENT ON dbname from issuing a warning when loading
>    into a database with a different name, perhaps using COMMENT ON
>    CURRENT DATABASE

16 years agoUpdate comment.
Bruce Momjian [Mon, 19 Nov 2007 23:48:33 +0000 (23:48 +0000)]
Update comment.

16 years agoAdd suggested syntax for COMMENT ON DATABASE:
Bruce Momjian [Mon, 19 Nov 2007 19:27:03 +0000 (19:27 +0000)]
Add suggested syntax for COMMENT ON DATABASE:

<    with a different name
>    with a different name, perhaps using COMMENT ON CURRENT
>           DATABASE
>

16 years agoAdd comment about COMMENT ON DATABASE failing for differently-named
Bruce Momjian [Mon, 19 Nov 2007 18:44:06 +0000 (18:44 +0000)]
Add comment about COMMENT ON DATABASE failing for differently-named
databases.

16 years agoAdd:
Bruce Momjian [Mon, 19 Nov 2007 18:42:11 +0000 (18:42 +0000)]
Add:

>  o Allow COMMENT ON dbname to work when loading into a database
>    with a different name

16 years agoImprove description of implicit-text-casting compatibility issue.
Tom Lane [Mon, 19 Nov 2007 03:47:14 +0000 (03:47 +0000)]
Improve description of implicit-text-casting compatibility issue.
A bunch of other desultory copy-editing, too.

16 years agoMinor improvement in documentation of temp_tablespaces: avoid grammatical
Tom Lane [Mon, 19 Nov 2007 02:26:10 +0000 (02:26 +0000)]
Minor improvement in documentation of temp_tablespaces: avoid grammatical
ambiguity and say what the default value is.

16 years agoRemove no-longer-accurate claim that REINDEX won't invalidate cached
Tom Lane [Sun, 18 Nov 2007 18:42:03 +0000 (18:42 +0000)]
Remove no-longer-accurate claim that REINDEX won't invalidate cached
plans.

16 years agoWordsmith improvements to first half of release notes, move TEXT cast
Bruce Momjian [Sat, 17 Nov 2007 21:13:42 +0000 (21:13 +0000)]
Wordsmith improvements to first half of release notes, move TEXT cast
item to #1 imcompatibility slot.

16 years agoSimplify TEXT cast working; spelling fixes to release notes.
Bruce Momjian [Sat, 17 Nov 2007 15:17:16 +0000 (15:17 +0000)]
Simplify TEXT cast working;  spelling fixes to release notes.

16 years agoImprove GIN index build's tracking of memory usage by using
Tom Lane [Fri, 16 Nov 2007 21:55:59 +0000 (21:55 +0000)]
Improve GIN index build's tracking of memory usage by using
GetMemoryChunkSpace, not just the palloc request size.  This brings the
allocatedMemory counter close enough to reality (as measured by
MemoryContextStats printouts) that I think we can get rid of the arbitrary
factor-of-2 adjustment that was put into the code initially.  Given the
sensitivity of GIN build to work memory size, not using as much of work
memory as we're allowed to seems a pretty bad idea.

16 years agoGIN index build's allocatedMemory counter needs to be long, not uint32.
Tom Lane [Fri, 16 Nov 2007 21:50:06 +0000 (21:50 +0000)]
GIN index build's allocatedMemory counter needs to be long, not uint32.
Else, in a 64-bit machine with maintenance_work_mem set to above 4Gb,
the counter overflows and we never recognize having reached the
maintenance_work_mem limit.  I believe this explains out-of-memory
failure recently reported by Sean Davis.

This is a bug, so backpatch to 8.2.

16 years agoFix inconsistent message wording, and split off a couple of duplicated strings.
Alvaro Herrera [Fri, 16 Nov 2007 21:47:32 +0000 (21:47 +0000)]
Fix inconsistent message wording, and split off a couple of duplicated strings.

16 years agoRepair still another bug in the btree page split WAL reduction patch:
Tom Lane [Fri, 16 Nov 2007 19:53:50 +0000 (19:53 +0000)]
Repair still another bug in the btree page split WAL reduction patch:
it failed for splits of non-leaf pages because in such pages the first
data key on a page is suppressed, and so we can't just copy the first
key from the right page to reconstitute the left page's high key.
Problem found by Koichi Suzuki, patch by Heikki.

16 years agoFix a few mentions of "Postgres" that slipped into the release notes.
Bruce Momjian [Fri, 16 Nov 2007 17:22:21 +0000 (17:22 +0000)]
Fix a few mentions of "Postgres" that slipped into the release notes.

16 years agoAdd release note that "Postgres" is now documented as an acceptable
Bruce Momjian [Fri, 16 Nov 2007 16:12:14 +0000 (16:12 +0000)]
Add release note that "Postgres" is now documented as an acceptable
alias.

16 years agoSimplify wording.
Bruce Momjian [Fri, 16 Nov 2007 15:53:57 +0000 (15:53 +0000)]
Simplify wording.

16 years agoAdd FAQ items about "Postgres" alias usage, remove old reference.
Bruce Momjian [Fri, 16 Nov 2007 15:50:31 +0000 (15:50 +0000)]
Add FAQ items about "Postgres" alias usage, remove old reference.

16 years agoFix tsvectorout() and tsqueryout() to escape backslesh, add test of that.
Teodor Sigaev [Fri, 16 Nov 2007 15:05:59 +0000 (15:05 +0000)]
Fix tsvectorout() and tsqueryout() to escape backslesh, add test of that.
Patch by Bruce Momjian <bruce@momjian.us>

Backpatch is needed, but it's impossible to apply it directly

16 years agoFix markup that doesn't work in HISTORY generation.
Tom Lane [Fri, 16 Nov 2007 06:28:07 +0000 (06:28 +0000)]
Fix markup that doesn't work in HISTORY generation.

16 years agoRe-do configure with autoconf 2.59
Tom Lane [Fri, 16 Nov 2007 04:59:25 +0000 (04:59 +0000)]
Re-do configure with autoconf 2.59

16 years agoupdate files for beta3
Marc G. Fournier [Fri, 16 Nov 2007 04:29:45 +0000 (04:29 +0000)]
update files for beta3

16 years agoAdd a couple of notes pointing out that GIN index build time is very
Tom Lane [Fri, 16 Nov 2007 03:23:07 +0000 (03:23 +0000)]
Add a couple of notes pointing out that GIN index build time is very
sensitive to maintenance_work_mem (something I just learned the hard
way).

16 years agoSmall comment spacing improvement.
Bruce Momjian [Fri, 16 Nov 2007 01:51:22 +0000 (01:51 +0000)]
Small comment spacing improvement.

16 years agoBetter guard token used by pgindent.
Bruce Momjian [Fri, 16 Nov 2007 01:25:15 +0000 (01:25 +0000)]
Better guard token used by pgindent.

16 years agoRun pgindent on remaining files now that LOOPBYTE is a usable macro.
Bruce Momjian [Fri, 16 Nov 2007 01:12:24 +0000 (01:12 +0000)]
Run pgindent on remaining files now that LOOPBYTE is a usable macro.

16 years agoCleanup for new else/comment handling.
Bruce Momjian [Fri, 16 Nov 2007 01:11:04 +0000 (01:11 +0000)]
Cleanup for new else/comment handling.

16 years agoFix stupid typo in recently-added code :-(
Tom Lane [Fri, 16 Nov 2007 00:57:55 +0000 (00:57 +0000)]
Fix stupid typo in recently-added code :-(

16 years agoAdd snb_ru_init(internal) to list of stub functions in tsearch2
Tom Lane [Fri, 16 Nov 2007 00:34:54 +0000 (00:34 +0000)]
Add snb_ru_init(internal) to list of stub functions in tsearch2
compatibility module.  Needed to support loading of 8.1-era tsearch2
configuration data.

16 years agoModify LOOPBYTE/LOOPBIT macros to be more logical; rather than have the
Bruce Momjian [Fri, 16 Nov 2007 00:13:02 +0000 (00:13 +0000)]
Modify LOOPBYTE/LOOPBIT macros to be more logical;  rather than have the
for() body passed as a parameter, make the macros act as simple headers
to code blocks.

This allows pgindent to be run on these files.

16 years agoFix pgindent to properly handle 'else' and single-line comments on the
Bruce Momjian [Thu, 15 Nov 2007 23:23:44 +0000 (23:23 +0000)]
Fix pgindent to properly handle 'else' and single-line comments on the
same line;  previous fix was only partial.  Re-run pgindent on files
that need it.

16 years agoRe-run pgindent with updated list of typedefs. (Updated README should
Bruce Momjian [Thu, 15 Nov 2007 22:25:18 +0000 (22:25 +0000)]
Re-run pgindent with updated list of typedefs.  (Updated README should
avoid this problem in the future.)

16 years agoBeef up README instructions, again.
Bruce Momjian [Thu, 15 Nov 2007 22:15:46 +0000 (22:15 +0000)]
Beef up README instructions, again.

16 years agoCUpdate README to suggest 'gmake distclean'. Add library typedefs.
Bruce Momjian [Thu, 15 Nov 2007 22:12:09 +0000 (22:12 +0000)]
CUpdate README to suggest 'gmake distclean'.  Add library typedefs.

16 years agoUpdate pgtools README to be clearer about typdefs.
Bruce Momjian [Thu, 15 Nov 2007 22:09:07 +0000 (22:09 +0000)]
Update pgtools README to be clearer about typdefs.

16 years agoUpdate pgindent with current typedefs.
Bruce Momjian [Thu, 15 Nov 2007 22:06:07 +0000 (22:06 +0000)]
Update pgindent with current typedefs.

16 years agoWhen logging the recovery.conf parameters, show them quoted as they would
Peter Eisentraut [Thu, 15 Nov 2007 22:02:12 +0000 (22:02 +0000)]
When logging the recovery.conf parameters, show them quoted as they would
appear in the configuration file.

16 years agoAdd blank lines to pgindent.
Bruce Momjian [Thu, 15 Nov 2007 21:52:39 +0000 (21:52 +0000)]
Add blank lines to pgindent.

16 years agoUse ftruncate() not truncate() in mdunlink. Seems Windows doesn't
Tom Lane [Thu, 15 Nov 2007 21:49:47 +0000 (21:49 +0000)]
Use ftruncate() not truncate() in mdunlink.  Seems Windows doesn't
support the latter.

16 years agoUpdate timezone data files to release 2007i of the zic database.
Tom Lane [Thu, 15 Nov 2007 21:21:14 +0000 (21:21 +0000)]
Update timezone data files to release 2007i of the zic database.

16 years agopgindent run for 8.3.
Bruce Momjian [Thu, 15 Nov 2007 21:14:46 +0000 (21:14 +0000)]
pgindent run for 8.3.

16 years agoTranslation updates
Peter Eisentraut [Thu, 15 Nov 2007 20:38:15 +0000 (20:38 +0000)]
Translation updates

16 years agoPrevent re-use of a deleted relation's relfilenode until after the next
Tom Lane [Thu, 15 Nov 2007 20:36:40 +0000 (20:36 +0000)]
Prevent re-use of a deleted relation's relfilenode until after the next
checkpoint.  This guards against an unlikely data-loss scenario in which
we re-use the relfilenode, then crash, then replay the deletion and
recreation of the file.  Even then we'd be OK if all insertions into the
new relation had been WAL-logged ... but that's not guaranteed given all
the no-WAL-logging optimizations that have recently been added.

Patch by Heikki Linnakangas, per a discussion last month.

16 years agoUpdate config.guess and config.sub
Peter Eisentraut [Thu, 15 Nov 2007 20:21:05 +0000 (20:21 +0000)]
Update config.guess and config.sub