]> granicus.if.org Git - postgresql/log
postgresql
14 years agoProperly credit Simon for functions pg_last_xlog_replay_location, etc.
Bruce Momjian [Mon, 22 Mar 2010 22:56:52 +0000 (22:56 +0000)]
Properly credit Simon for functions pg_last_xlog_replay_location, etc.

14 years agoFix regression test breakage from recent change in default name of exclusion-constrai...
Tom Lane [Mon, 22 Mar 2010 17:43:28 +0000 (17:43 +0000)]
Fix regression test breakage from recent change in default name of exclusion-constraint index.

14 years agoShorten suffix of automatically created indexes to "_excl" when using
Simon Riggs [Mon, 22 Mar 2010 15:24:11 +0000 (15:24 +0000)]
Shorten suffix of automatically created indexes to "_excl" when using
exclusion constraints, in line with string length of other pre-9.0 suffixes.

14 years agoRe-order createuser and vacuumedb documentation options, for
Bruce Momjian [Mon, 22 Mar 2010 14:56:09 +0000 (14:56 +0000)]
Re-order createuser and vacuumedb documentation options, for
consistency.

Gabrielle (Roth)

14 years agoFix an oversight in join-removal optimization: we have to check not only for
Tom Lane [Mon, 22 Mar 2010 13:57:16 +0000 (13:57 +0000)]
Fix an oversight in join-removal optimization: we have to check not only for
plain Vars that are generated in the inner rel and used above the join, but
also for PlaceHolderVars.  Per report from Oleg K.

14 years agoFixed ECPG regression test to make sure it uses absolute paths for include
Michael Meskes [Mon, 22 Mar 2010 07:41:20 +0000 (07:41 +0000)]
Fixed ECPG regression test to make sure it uses absolute paths for include
files instead of relative ones which break vpath builds.

14 years agoRemove 9.0 release notes xref's to HISTORY.html can be built.
Bruce Momjian [Sun, 21 Mar 2010 22:12:27 +0000 (22:12 +0000)]
Remove 9.0 release notes xref's to HISTORY.html can be built.

Document this restriction.

14 years agoAdding special command line option that is now needed for the one ecpg regression...
Michael Meskes [Sun, 21 Mar 2010 14:26:58 +0000 (14:26 +0000)]
Adding special command line option that is now needed for the one ecpg regression test that was changed.

14 years agoECPG's parser now accepts and handles variables as arguments for the FREE command.
Michael Meskes [Sun, 21 Mar 2010 11:56:45 +0000 (11:56 +0000)]
ECPG's parser now accepts and handles variables as arguments for the FREE command.

Informix allows variables as argument to the embedded SQL command FREE. Given
that we only allow freeing cursors via FREE for compatibility reasons only we
should do the same.

14 years agoAdjusted regression test results to the change I made in debug output for ecpglib.
Michael Meskes [Sun, 21 Mar 2010 11:38:16 +0000 (11:38 +0000)]
Adjusted regression test results to the change I made in debug output for ecpglib.

14 years agoCorrectly name functions in debug output in ecpglib. When the functions were
Michael Meskes [Sun, 21 Mar 2010 11:33:44 +0000 (11:33 +0000)]
Correctly name functions in debug output in ecpglib. When the functions were
refactored the debug output wasn't adjusted.

14 years agoECPG only copied #include statements instead of processing them according to
Michael Meskes [Sun, 21 Mar 2010 10:49:52 +0000 (10:49 +0000)]
ECPG only copied #include statements instead of processing them according to
commandline option "-i". This change fixes this and adds a test case. It also
honors #include_next, although this is probably never used for embedded SQL.

14 years agoAdd more 9.0 release note documentation links.
Bruce Momjian [Sun, 21 Mar 2010 02:24:29 +0000 (02:24 +0000)]
Add more 9.0 release note documentation links.

14 years agoAdd links to documentation sections in 9.0 release notes.
Bruce Momjian [Sun, 21 Mar 2010 00:43:40 +0000 (00:43 +0000)]
Add links to documentation sections in 9.0 release notes.

Also update tagging instructions, and add id tags to a few documentation
sections.

14 years agoMessage tuning
Peter Eisentraut [Sun, 21 Mar 2010 00:17:59 +0000 (00:17 +0000)]
Message tuning

14 years agoFixed ecpg parser to allow more than one C preprocessor command inside a declare...
Michael Meskes [Sat, 20 Mar 2010 18:53:00 +0000 (18:53 +0000)]
Fixed ecpg parser to allow more than one C preprocessor command inside a declare section.

14 years agoIn 9.0 release notes, move libpq/ecpg up into their own sections, like I
Bruce Momjian [Sat, 20 Mar 2010 16:31:09 +0000 (16:31 +0000)]
In 9.0 release notes, move libpq/ecpg up into their own sections, like I
did with server-side languages.

14 years agoMove server-side languages up one section in the 9.0 release notes.
Bruce Momjian [Sat, 20 Mar 2010 14:03:31 +0000 (14:03 +0000)]
Move server-side languages up one section in the 9.0 release notes.

14 years agoFurther corrections of mismatching struct and btree SizeOf macros.
Simon Riggs [Sat, 20 Mar 2010 07:49:48 +0000 (07:49 +0000)]
Further corrections of mismatching struct and btree SizeOf macros.
In this case, correction is to remove now unused fields from struct.
Since these were unused and full of garbage anyway, no version change.

14 years agoCreate 9.0 release notes; keep 9.0-alpha release notes in place.
Bruce Momjian [Sat, 20 Mar 2010 03:30:12 +0000 (03:30 +0000)]
Create 9.0 release notes;  keep 9.0-alpha release notes in place.

14 years agoUpdate sgml release note markup suggestions.
Bruce Momjian [Sat, 20 Mar 2010 03:12:52 +0000 (03:12 +0000)]
Update sgml release note markup suggestions.

14 years agoClear error_context_stack and debug_query_string at the beginning of proc_exit,
Tom Lane [Sat, 20 Mar 2010 00:58:09 +0000 (00:58 +0000)]
Clear error_context_stack and debug_query_string at the beginning of proc_exit,
so that we won't try to attach any context printouts to messages that get
emitted while exiting.  Per report from Dennis Koegel, the context functions
won't necessarily work after we've started shutting down the backend, and it
seems possible that debug_query_string could be pointing at freed storage
as well.  The context information doesn't seem particularly relevant to
such messages anyway, so there's little lost by suppressing it.

Back-patch to all supported branches.  I can only demonstrate a crash with
log_disconnections messages back to 8.1, but the risk seems real in 8.0 and
before anyway.

14 years agoForbid renaming columns of objects whose column names are system-generated.
Robert Haas [Sat, 20 Mar 2010 00:43:42 +0000 (00:43 +0000)]
Forbid renaming columns of objects whose column names are system-generated.

KaiGai Kohei, with adjustments to the comments.

14 years agoModify error context callback functions to not assume that they can fetch
Tom Lane [Fri, 19 Mar 2010 22:54:41 +0000 (22:54 +0000)]
Modify error context callback functions to not assume that they can fetch
catalog entries via SearchSysCache and related operations.  Although, at the
time that these callbacks are called by elog.c, we have not officially aborted
the current transaction, it still seems rather risky to initiate any new
catalog fetches.  In all these cases the needed information is readily
available in the caller and so it's just a matter of a bit of extra notation
to pass it to the callback.

Per crash report from Dennis Koegel.  I've concluded that the real fix for
his problem is to clear the error context stack at entry to proc_exit, but
it still seems like a good idea to make the callbacks a bit less fragile
for other cases.

Backpatch to 8.4.  We could go further back, but the patch doesn't apply
cleanly.  In the absence of proof that this fixes something and isn't just
paranoia, I'm not going to expend the effort.

14 years agoFix oversight in btpo.xact patch; it was in fact installing garbage
Tom Lane [Fri, 19 Mar 2010 20:51:30 +0000 (20:51 +0000)]
Fix oversight in btpo.xact patch; it was in fact installing garbage
in the xact field on replay, due to not writing out all the data in
the wal log struct.

14 years agoClarify docs about database parameter in streaming replication primary_conninfo.
Simon Riggs [Fri, 19 Mar 2010 19:31:06 +0000 (19:31 +0000)]
Clarify docs about database parameter in streaming replication primary_conninfo.
Docs were unclear on whether or not database=replication was required,
nor did they mention the FATAL error this causes if database parameter is
mentioned explicitly, whatever its value.

14 years agoAdd connection messages for streaming replication. log_connections
Simon Riggs [Fri, 19 Mar 2010 19:19:38 +0000 (19:19 +0000)]
Add connection messages for streaming replication. log_connections
was broken for a replication connection and no messages were
displayed on either standby or primary, at any debug level.
Connection messages needed to diagnose session drop/reconnect
events. Use LOG mode for now, discuss lowering in later releases.

14 years agoMinor tweaks on libpqrcv_connect(): ensure conninfo_repl[] is
Simon Riggs [Fri, 19 Mar 2010 17:51:42 +0000 (17:51 +0000)]
Minor tweaks on libpqrcv_connect(): ensure conninfo_repl[] is
correctly sized and expand comment to explain otherwise
undocumented use of replication connection parameter.

14 years agoUpdate XLOG_PAGE_MAGIC to recognise WAL format changes.
Simon Riggs [Fri, 19 Mar 2010 17:42:10 +0000 (17:42 +0000)]
Update XLOG_PAGE_MAGIC to recognise WAL format changes.

14 years agoAdjust comment in .history file to match recovery target specified. Comment
Simon Riggs [Fri, 19 Mar 2010 11:05:15 +0000 (11:05 +0000)]
Adjust comment in .history file to match recovery target specified. Comment
present since 8.0 was never fully meaningful, since two recovery targets
cannot be specified. Refactor recovery target type to make this change
and associated code easier to understand. No change in function.

Bug report arising from internal support question.

14 years agoReset btpo.xact following recovery of btree delete page. Add btpo_xact
Simon Riggs [Fri, 19 Mar 2010 10:41:22 +0000 (10:41 +0000)]
Reset btpo.xact following recovery of btree delete page. Add btpo_xact
field into WAL record and reset it from there, rather than using
FrozenTransactionId which can lead to some corner case bugs.

Problem report and suggested route to a fix from Heikki, details by me.

14 years agoAlso print the libpq error message when lo_create or lo_open fails
Peter Eisentraut [Thu, 18 Mar 2010 20:00:51 +0000 (20:00 +0000)]
Also print the libpq error message when lo_create or lo_open fails

14 years agoPrevent the injection of invalidly encoded strings by PL/Python into PostgreSQL
Peter Eisentraut [Thu, 18 Mar 2010 19:43:03 +0000 (19:43 +0000)]
Prevent the injection of invalidly encoded strings by PL/Python into PostgreSQL
with a few strategically placed pg_verifymbstr calls.

14 years agoMessage style tuning
Peter Eisentraut [Thu, 18 Mar 2010 19:02:46 +0000 (19:02 +0000)]
Message style tuning

14 years agoMention way to get commit details for release notes.
Bruce Momjian [Thu, 18 Mar 2010 16:31:12 +0000 (16:31 +0000)]
Mention way to get commit details for release notes.

14 years agoFix missing parentheses for current_query(), per bug #5378.
Tom Lane [Thu, 18 Mar 2010 15:29:44 +0000 (15:29 +0000)]
Fix missing parentheses for current_query(), per bug #5378.
Also make a couple other minor editorial improvements.

14 years agoUse data-type specific conversion functions also in plpy.execute
Peter Eisentraut [Thu, 18 Mar 2010 13:23:57 +0000 (13:23 +0000)]
Use data-type specific conversion functions also in plpy.execute

In PLy_spi_execute_plan, use the data-type specific Python-to-PostgreSQL
conversion function instead of passing everything through InputFunctionCall
as a string.  The equivalent fix was already done months ago for function
parameters and return values, but this other gateway between Python and
PostgreSQL was apparently forgotten.  As a result, data types that need
special treatment, such as bytea, would misbehave when used with
plpy.execute.

14 years agoAdd restartpoint_command option to recovery.conf. Fix bug in %r handling
Heikki Linnakangas [Thu, 18 Mar 2010 09:17:18 +0000 (09:17 +0000)]
Add restartpoint_command option to recovery.conf. Fix bug in %r handling
in recovery_end_command, it always came out as 0 because InRedo was
cleared before recovery_end_command was executed. Also, always take
ControlFileLock when reading checkpoint location for %r.

The recovery_end_command bug and the missing locking was present in 8.4
as well, that part of this patch will be backported separately.

14 years agoAdd vacuum_defer_cleanup_age to postgresql.conf.sample.
Simon Riggs [Thu, 18 Mar 2010 07:01:01 +0000 (07:01 +0000)]
Add vacuum_defer_cleanup_age to postgresql.conf.sample.

14 years agoDon't link PL/Python against LOCALMODLIBS
Peter Eisentraut [Wed, 17 Mar 2010 22:02:44 +0000 (22:02 +0000)]
Don't link PL/Python against LOCALMODLIBS

This variable is apparently only for Python internally.  In newer releases
of Python this variable pulls in more and more libraries that users are
less likely to have, leading to potential build failures.

14 years agoMessage style tuning
Peter Eisentraut [Wed, 17 Mar 2010 21:31:17 +0000 (21:31 +0000)]
Message style tuning

14 years agoFix peculiar, untranslatable message concatenation attempt
Peter Eisentraut [Wed, 17 Mar 2010 20:58:38 +0000 (20:58 +0000)]
Fix peculiar, untranslatable message concatenation attempt

14 years agoMake typography consistent
Peter Eisentraut [Wed, 17 Mar 2010 20:53:19 +0000 (20:53 +0000)]
Make typography consistent

14 years agoTypo fixes.
Magnus Hagander [Wed, 17 Mar 2010 18:03:55 +0000 (18:03 +0000)]
Typo fixes.

Fujii Masao

14 years agoUpdate broken and permanently moved links
Peter Eisentraut [Wed, 17 Mar 2010 17:12:31 +0000 (17:12 +0000)]
Update broken and permanently moved links

14 years agoPass incompletely-transformed aggregate argument lists as separate parameters
Tom Lane [Wed, 17 Mar 2010 16:52:38 +0000 (16:52 +0000)]
Pass incompletely-transformed aggregate argument lists as separate parameters
to transformAggregateCall, instead of abusing fields in Aggref to carry them
temporarily.  No change in functionality but hopefully the code is a bit
clearer now.  Per gripe from Gokulakannan Somasundaram.

14 years agoFix incorrect example in CREATE INDEX reference page, per Josh Kupershmidt.
Tom Lane [Wed, 17 Mar 2010 15:55:50 +0000 (15:55 +0000)]
Fix incorrect example in CREATE INDEX reference page, per Josh Kupershmidt.
Also fix and uncomment an old example of creating a GIST index, and make
a couple of other minor editorial adjustments.

14 years agoFix typo in streaming replication protocol docs.
Simon Riggs [Wed, 17 Mar 2010 00:28:15 +0000 (00:28 +0000)]
Fix typo in streaming replication protocol docs.

14 years agoFix typo in math function docs, spotted by Heikki.
Simon Riggs [Tue, 16 Mar 2010 16:03:24 +0000 (16:03 +0000)]
Fix typo in math function docs, spotted by Heikki.

14 years agoThrow a nicer error message if a standby server attempts to connect while
Heikki Linnakangas [Tue, 16 Mar 2010 09:09:55 +0000 (09:09 +0000)]
Throw a nicer error message if a standby server attempts to connect while
the master is still in recovery. We don't support cascading slaves yet.

Patch by Fujii Masao, with slightly changed wording.

14 years agoDocument that trig functions accept/return values in radians.
Simon Riggs [Mon, 15 Mar 2010 19:03:31 +0000 (19:03 +0000)]
Document that trig functions accept/return values in radians.

14 years agoRemove incorrect comment from GetWriteRecPtr(): the return value is always
Simon Riggs [Mon, 15 Mar 2010 18:49:17 +0000 (18:49 +0000)]
Remove incorrect comment from GetWriteRecPtr(): the return value is always
correct, as described in comments at start of xlog.c

14 years agoUpdate oidjoins regression test for 9.0.
Tom Lane [Sun, 14 Mar 2010 04:17:54 +0000 (04:17 +0000)]
Update oidjoins regression test for 9.0.

14 years agoAdd some more structure and bits of information to PL/Python documentation
Peter Eisentraut [Sat, 13 Mar 2010 20:55:05 +0000 (20:55 +0000)]
Add some more structure and bits of information to PL/Python documentation

14 years agoRevert all keepalive moves, reevaluate.
Bruce Momjian [Sat, 13 Mar 2010 16:56:37 +0000 (16:56 +0000)]
Revert all keepalive moves, reevaluate.

14 years agoMove more tcp keepalive macros to be consistent.
Bruce Momjian [Sat, 13 Mar 2010 16:40:38 +0000 (16:40 +0000)]
Move more tcp keepalive macros to be consistent.

14 years agoMove TCP keepalive macro definition
Bruce Momjian [Sat, 13 Mar 2010 15:35:46 +0000 (15:35 +0000)]
Move TCP keepalive macro definition

Jaime Casanova

14 years agoAdd libpq warning message if the .pgpass-retrieved password fails.
Bruce Momjian [Sat, 13 Mar 2010 14:55:57 +0000 (14:55 +0000)]
Add libpq warning message if the .pgpass-retrieved password fails.

Add ERRCODE_INVALID_PASSWORD sqlstate error code.

14 years agoUpdate shared memory configuration information for Linux
Peter Eisentraut [Sat, 13 Mar 2010 11:00:19 +0000 (11:00 +0000)]
Update shared memory configuration information for Linux

In particular, the assertion that shmall is sufficiently sized by default
is slowly becoming untrue.

14 years agoDocument what user name email symbols are supported by tsearch.
Bruce Momjian [Sat, 13 Mar 2010 03:09:04 +0000 (03:09 +0000)]
Document what user name email symbols are supported by tsearch.

14 years agoAllow underscores in tsearch email addressses, per RFC 5322 and report
Bruce Momjian [Sat, 13 Mar 2010 00:41:58 +0000 (00:41 +0000)]
Allow underscores in tsearch email addressses, per RFC 5322 and report
by Dan O'Hara.

Patch by Teodor Sigaev

14 years agoSmall adjustment by pgindent.
Bruce Momjian [Sat, 13 Mar 2010 00:40:43 +0000 (00:40 +0000)]
Small adjustment by pgindent.

14 years agoRe-add comma, oops.
Bruce Momjian [Fri, 12 Mar 2010 23:21:48 +0000 (23:21 +0000)]
Re-add comma, oops.

14 years agoAdd some logging code for unexpected cases in pgstat.c, particularly being
Tom Lane [Fri, 12 Mar 2010 22:19:19 +0000 (22:19 +0000)]
Add some logging code for unexpected cases in pgstat.c, particularly being
unable to read a stats file for reasons other than ENOENT, and having to reset
last_statrequest because it's later than current time in the collector.
Not clear if this will shed any light on the "pgstat wait timeout" business,
but it seems like a good idea in general.

In passing, do some message-style-police work on recently-added
pgstat_reset_shared_counters code.

14 years agoFix a commented-out block of code that pgindent didn't understand
Tom Lane [Fri, 12 Mar 2010 21:40:36 +0000 (21:40 +0000)]
Fix a commented-out block of code that pgindent didn't understand
properly; it's been making that comment uglier with each run.

14 years agoRemove unnecessary comma in documentation.
Bruce Momjian [Fri, 12 Mar 2010 17:46:07 +0000 (17:46 +0000)]
Remove unnecessary comma in documentation.

14 years agoAdd missing reset of need_initialization in reloptions code.
Tom Lane [Thu, 11 Mar 2010 21:47:19 +0000 (21:47 +0000)]
Add missing reset of need_initialization in reloptions code.

This resulted in useless extra work during every call of parseRelOptions,
but no bad effects other than that.  Noted by Alvaro.

14 years agoModify psql's \d printout to fold exclusion constraints in with regular
Tom Lane [Thu, 11 Mar 2010 21:29:32 +0000 (21:29 +0000)]
Modify psql's \d printout to fold exclusion constraints in with regular
indexes, rather than printing them twice.  Per my gripe when the exclusion
constraint feature was committed.

14 years agoSync timezone code with tzcode 2010c from the Olson group. This fixes some
Tom Lane [Thu, 11 Mar 2010 18:43:24 +0000 (18:43 +0000)]
Sync timezone code with tzcode 2010c from the Olson group.  This fixes some
corner cases that come up in certain timezones (apparently, only those with
lots and lots of distinct TZ transition rules, as far as I can gather from
a quick scan of their archives).  Per suggestion from Jeevan Chalke.

Back-patch to 8.4.  Possibly we need to push this into earlier releases
as well, but I'm hesitant to update them to the 64-bit tzcode without
more thought and testing.

14 years agoFix bug in KnownAssignedXidsMany(). I saw this when looking at the
Heikki Linnakangas [Thu, 11 Mar 2010 09:26:59 +0000 (09:26 +0000)]
Fix bug in KnownAssignedXidsMany(). I saw this when looking at the
assertion failure reported by Erik Rijkers, but this alone doesn't explain
the failure.

14 years agoFix comment which was apparently copy-pasted from another function.
Heikki Linnakangas [Thu, 11 Mar 2010 09:10:25 +0000 (09:10 +0000)]
Fix comment which was apparently copy-pasted from another function.

14 years agoSimplify a couple of pg_dump and psql \d queries about index constraints
Tom Lane [Thu, 11 Mar 2010 04:36:43 +0000 (04:36 +0000)]
Simplify a couple of pg_dump and psql \d queries about index constraints
by joining to pg_constraint.conindid, instead of the former technique of
joining indirectly through pg_depend.  This is much more straightforward
and probably faster as well.  I had originally desisted from changing these
queries when conindid was added because I was worried about losing
performance, but if we join on conrelid as well as conindid then the index
on conrelid can be used when pg_constraint is large.

14 years agoUpdate comment for pg_constraint.conindid to mention that it's used for
Tom Lane [Thu, 11 Mar 2010 03:36:22 +0000 (03:36 +0000)]
Update comment for pg_constraint.conindid to mention that it's used for
exclusion constraints.  Not sure how we managed to update the comment for
it in catalogs.sgml but miss this one.

14 years agoFix incorrect comment about permissions checking being done in utility.c.
Robert Haas [Wed, 10 Mar 2010 19:48:39 +0000 (19:48 +0000)]
Fix incorrect comment about permissions checking being done in utility.c.

Noted while reviewing a patch from KaiGai Kohei.

14 years agopg_start_backup() can use a share lock to lock ControlFileLock
Itagaki Takahiro [Wed, 10 Mar 2010 02:04:48 +0000 (02:04 +0000)]
pg_start_backup() can use a share lock to lock ControlFileLock
instead of an exclusive lock.

The change is almost for code cleanup. Since there seems to be no
performance benefits from it, backports should not be needed.

Fujii Masao

14 years agoPreliminary release notes for releases 8.4.3, 8.3.10, 8.2.16, 8.1.20, 8.0.24,
Tom Lane [Wed, 10 Mar 2010 01:58:11 +0000 (01:58 +0000)]
Preliminary release notes for releases 8.4.3, 8.3.10, 8.2.16, 8.1.20, 8.0.24,
7.4.28.

14 years agoUse SvROK(sv) rather than directly checking SvTYPE(sv) == SVt_RV in plperl.
Tom Lane [Tue, 9 Mar 2010 22:34:38 +0000 (22:34 +0000)]
Use SvROK(sv) rather than directly checking SvTYPE(sv) == SVt_RV in plperl.
The latter is considered unwarranted chumminess with the implementation,
and can lead to crashes with recent Perl versions.

Report and fix by Tim Bunce.  Back-patch to all versions containing the
questionable coding pattern.

14 years agoUpdate time zone data files to tzdata release 2010d: DST law changes in Fiji,
Alvaro Herrera [Tue, 9 Mar 2010 14:27:28 +0000 (14:27 +0000)]
Update time zone data files to tzdata release 2010d: DST law changes in Fiji,
Samoa, Chile; corrections to recent changes in Paraguay and Bangladesh.

14 years agoecpg now adds a unique counter to its varchar struct definitions to make these defini...
Michael Meskes [Tue, 9 Mar 2010 11:09:46 +0000 (11:09 +0000)]
ecpg now adds a unique counter to its varchar struct definitions to make these definitions unique, too. It used to use the linenumber but in the rare case of two definitions in one line this was not unique.

14 years agoAdd checks for valid encoding of query strings passed to SPI by plperl.
Andrew Dunstan [Tue, 9 Mar 2010 02:48:33 +0000 (02:48 +0000)]
Add checks for valid encoding of query strings passed to SPI by plperl.

14 years agoReturn proper exit code (3) from psql when ON_ERROR_STOP=on and
Bruce Momjian [Mon, 8 Mar 2010 23:03:00 +0000 (23:03 +0000)]
Return proper exit code (3) from psql when ON_ERROR_STOP=on and
--single-transaction are both used and the failure happens in commit,
e.g. failed deferred trigger.  Also properly free BEGIN/COMMIT result
structures from --single-transaction.

Per report from Dominic Bevacqua

14 years agoAdd missing space in example.
Magnus Hagander [Mon, 8 Mar 2010 12:38:37 +0000 (12:38 +0000)]
Add missing space in example.

Tim Landscheidt

14 years agoBetter test the content of the SQLSTATE string in ecpglib than the pointer.
Michael Meskes [Mon, 8 Mar 2010 12:15:24 +0000 (12:15 +0000)]
Better test the content of the SQLSTATE string in ecpglib than the pointer.

14 years agoRequire hostname to be set when using GSSAPI authentication. Without it,
Magnus Hagander [Mon, 8 Mar 2010 10:01:12 +0000 (10:01 +0000)]
Require hostname to be set when using GSSAPI authentication. Without it,
the GSSAPI libraries crash.

Noted by Zdenek Kotala

14 years agoDisallow gssapi authentication on local connections, since it
Magnus Hagander [Mon, 8 Mar 2010 09:57:26 +0000 (09:57 +0000)]
Disallow gssapi authentication on local connections, since it
requires a hostname to function.

Noted by Zdenek Kotala

14 years agoUpdate time zone data files to tzdata release 2010c: DST law changes in
Tom Lane [Mon, 8 Mar 2010 01:18:09 +0000 (01:18 +0000)]
Update time zone data files to tzdata release 2010c: DST law changes in
Bangladesh, Mexico, Paraguay.

14 years agoRevert patch for --psqlrc=FILENAME in psql.
Magnus Hagander [Sun, 7 Mar 2010 17:02:34 +0000 (17:02 +0000)]
Revert patch for --psqlrc=FILENAME in psql.

14 years agoFix warning messages in restrict_and_check_grant() to include the column name
Tom Lane [Sat, 6 Mar 2010 23:10:42 +0000 (23:10 +0000)]
Fix warning messages in restrict_and_check_grant() to include the column name
when warning about column-level privileges.  This is more useful than before
and makes the apparent duplication complained of by Piyush Newe not so
duplicate.  Also fix lack of quote marks in a related message text.

Back-patch to 8.4, where column-level privileges were introduced.

Stephen Frost

14 years agoAdd --psqlrc=FILENAME parameter to psql, to process an explicitly named
Magnus Hagander [Sat, 6 Mar 2010 15:28:09 +0000 (15:28 +0000)]
Add --psqlrc=FILENAME parameter to psql, to process an explicitly named
file instead of ~/.psqlrc on startup.

14 years agoWhen reading pg_hba.conf and similar files, do not treat @file as an inclusion
Tom Lane [Sat, 6 Mar 2010 00:45:49 +0000 (00:45 +0000)]
When reading pg_hba.conf and similar files, do not treat @file as an inclusion
unless (1) the @ isn't quoted and (2) the filename isn't empty.  This guards
against unexpectedly treating usernames or other strings in "flat files"
as inclusion requests, as seen in a recent trouble report from Ed L.
The empty-filename case would be guaranteed to misbehave anyway, because our
subsequent path-munging behavior results in trying to read the directory
containing the current input file.

I think this might finally explain the report at
http://archives.postgresql.org/pgsql-bugs/2004-05/msg00132.php
of a crash after printing "authentication file token too long, skipping",
since I was able to duplicate that message (though not a crash) on a
platform where stdio doesn't refuse to read directories.  We never got
far in investigating that problem, but now I'm suspicious that the trigger
condition was an @ in the flat password file.

Back-patch to all active branches since the problem can be demonstrated in all
branches except HEAD.  The test case, creating a user named "@", doesn't cause
a problem in HEAD since we got rid of the flat password file.  Nonetheless it
seems like a good idea to not consider quoted @ as a file inclusion spec,
so I changed HEAD too.

14 years agoIn case the connection magically disappears libecpg only returns an internal
Michael Meskes [Fri, 5 Mar 2010 13:57:09 +0000 (13:57 +0000)]
In case the connection magically disappears libecpg only returns an internal
error sqlstate. This change makes it return a correct value..

14 years agoFix IsBinaryCoercible to not confuse a cast using in/out functions
Heikki Linnakangas [Thu, 4 Mar 2010 09:39:53 +0000 (09:39 +0000)]
Fix IsBinaryCoercible to not confuse a cast using in/out functions
with binary compatibility.

Backpatch to 8.4 where INOUT casts were introduced.

14 years agoAdd C comment about DDL changes possibly causing pg_dump errors.
Bruce Momjian [Wed, 3 Mar 2010 23:38:44 +0000 (23:38 +0000)]
Add C comment about DDL changes possibly causing pg_dump errors.

14 years agoDocument that "Q" is ignored by to_date and to_timestamp. Add C comment
Bruce Momjian [Wed, 3 Mar 2010 22:28:42 +0000 (22:28 +0000)]
Document that "Q" is ignored by to_date and to_timestamp.  Add C comment
about the behavior.

Document that quotes in to_date, to_timestamp, to_number skip input
characters.

14 years agoFix a couple of places that would loop forever if attempts to read a stdio file
Tom Lane [Wed, 3 Mar 2010 20:31:09 +0000 (20:31 +0000)]
Fix a couple of places that would loop forever if attempts to read a stdio file
set ferror() but never set feof().  This is known to be the case for recent
glibc when trying to read a directory as a file, and might be true for other
platforms/cases too.  Per report from Ed L.  (There is more that we ought to
do about his report, but this is one easily identifiable issue.)

14 years agoFix pg_dump of ACLs of foreign servers. The command to grant/revoke
Heikki Linnakangas [Wed, 3 Mar 2010 20:10:48 +0000 (20:10 +0000)]
Fix pg_dump of ACLs of foreign servers. The command to grant/revoke
privileges of foreign servers is "GRANT ... ON *FOREIGN* SERVER ...".

14 years agoMake contrib/xml2 use core xml.c's error handler, when available (that is,
Tom Lane [Wed, 3 Mar 2010 19:10:22 +0000 (19:10 +0000)]
Make contrib/xml2 use core xml.c's error handler, when available (that is,
in versions >= 8.3).  The core code is more robust and efficient than what
was there before, and this also reduces risks involved in swapping different
libxml error handler settings.

Before 8.3, there is still some risk of problems if add-on modules such as
Perl invoke libxml without setting their own error handler.  Given the lack
of reports I'm not sure there's a risk in practice, so I didn't take the
step of actually duplicating the core code into older contrib/xml2 branches.
Instead I just tweaked the existing code to ensure it didn't leave a dangling
pointer to short-lived memory when throwing an error.

14 years agoExport xml.c's libxml-error-handling support so that contrib/xml2 can use it
Tom Lane [Wed, 3 Mar 2010 17:29:45 +0000 (17:29 +0000)]
Export xml.c's libxml-error-handling support so that contrib/xml2 can use it
too, instead of duplicating the functionality (badly).

I renamed xml_init to pg_xml_init, because the former seemed just a bit too
generic to be safe as a global symbol.  I considered likewise renaming
xml_ereport to pg_xml_ereport, but felt that the reference to ereport probably
made it sufficiently PG-centric already.

14 years agoMake iconv work like other optional libraries for MSVC.
Andrew Dunstan [Wed, 3 Mar 2010 03:29:37 +0000 (03:29 +0000)]
Make iconv work like other optional libraries for MSVC.

14 years agoFix SGML markup.
Bruce Momjian [Wed, 3 Mar 2010 03:23:12 +0000 (03:23 +0000)]
Fix SGML markup.