OSDN Git Service

pg-rex/syncrep.git
13 years agoMake synchronous replication query cancel/die messages more consistent.
Robert Haas [Fri, 18 Mar 2011 14:20:22 +0000 (10:20 -0400)]
Make synchronous replication query cancel/die messages more consistent.

Per a gripe from Thom Brown about my previous commit in this area,
commit 9a56dc3389b9470031e9ef8e45c95a680982e01a.

13 years agoRemove ancient -X options to pg_dump, pg_dumpall, pg_restore.
Robert Haas [Fri, 18 Mar 2011 13:44:44 +0000 (09:44 -0400)]
Remove ancient -X options to pg_dump, pg_dumpall, pg_restore.

The last version in which these options were documented is now EOL, so
it's time to get rid of them for real.  We now use GNU-style long
options instead.

13 years agoRemove bogus semicolons in recoveryPausesHere.
Robert Haas [Fri, 18 Mar 2011 12:09:09 +0000 (08:09 -0400)]
Remove bogus semicolons in recoveryPausesHere.

Without this, the startup process goes into a tight loop, consuming
100% of one CPU and failing to respond to interrupts.

13 years agoMinor fixes for high availability documentation.
Robert Haas [Thu, 17 Mar 2011 19:46:44 +0000 (15:46 -0400)]
Minor fixes for high availability documentation.

Erik Rijkers and me

13 years agoProtocol documentation for Hot Standby feedback messages.
Robert Haas [Thu, 17 Mar 2011 19:10:55 +0000 (15:10 -0400)]
Protocol documentation for Hot Standby feedback messages.

13 years agoRemove bogus comment.
Robert Haas [Thu, 17 Mar 2011 18:43:14 +0000 (14:43 -0400)]
Remove bogus comment.

13 years agoFix PL/Python memory leak involving array slices
Alvaro Herrera [Thu, 17 Mar 2011 15:28:46 +0000 (12:28 -0300)]
Fix PL/Python memory leak involving array slices

Report and patch from Daniel Popowich, bug #5842
(with some debugging help from Alex Hunsaker)

13 years agoRaise maximum value of several timeout parameters
Peter Eisentraut [Thu, 17 Mar 2011 18:19:51 +0000 (20:19 +0200)]
Raise maximum value of several timeout parameters

The maximum value of deadlock_timeout, max_standby_archive_delay,
max_standby_streaming_delay, log_min_duration_statement, and
log_autovacuum_min_duration was INT_MAX/1000 milliseconds, which is
about 35min, which is too short for some practical uses.  Raise the
maximum value to INT_MAX; the code that uses the parameters already
supports that just fine.

13 years agoAdd pause_at_recovery_target to recovery.conf.sample; improve docs.
Robert Haas [Thu, 17 Mar 2011 18:02:41 +0000 (14:02 -0400)]
Add pause_at_recovery_target to recovery.conf.sample; improve docs.

Fujii Masao, but with the proposed behavior change reverted, and the
rest adjusted accordingly.

13 years agoFix various possible problems with synchronous replication.
Robert Haas [Thu, 17 Mar 2011 17:10:42 +0000 (13:10 -0400)]
Fix various possible problems with synchronous replication.

1. Don't ignore query cancel interrupts.  Instead, if the user asks to
cancel the query after we've already committed it, but before it's on
the standby, just emit a warning and let the COMMIT finish.

2. Don't ignore die interrupts (pg_terminate_backend or fast shutdown).
Instead, emit a warning message and close the connection without
acknowledging the commit.  Other backends will still see the effect of
the commit, but there's no getting around that; it's too late to abort
at this point, and ignoring die interrupts altogether doesn't seem like
a good idea.

3. If synchronous_standby_names becomes empty, wake up all backends
waiting for synchronous replication to complete.  Without this, someone
attempting to shut synchronous replication off could easily wedge the
entire system instead.

4. Avoid depending on the assumption that if a walsender updates
MyProc->syncRepState, we'll see the change even if we read it without
holding the lock.  The window for this appears to be quite narrow (and
probably doesn't exist at all on machines with strong memory ordering)
but protecting against it is practically free, so do that.

5. Remove useless state SYNC_REP_MUST_DISCONNECT, which isn't needed and
doesn't actually do anything.

There's still some further work needed here to make the behavior of fast
shutdown plausible, but that looks complex, so I'm leaving it for a
separate commit.  Review by Fujii Masao.

13 years agoDocument guc context values, and reference them from the config doc section.
Bruce Momjian [Thu, 17 Mar 2011 04:26:03 +0000 (00:26 -0400)]
Document guc context values, and reference them from the config doc section.

Tom Lane

13 years agoUse correct PATH separator for Cygwin in pg_regress.c.
Andrew Dunstan [Thu, 17 Mar 2011 04:06:52 +0000 (00:06 -0400)]
Use correct PATH separator for Cygwin in pg_regress.c.

This has been broken for years, and I'm not sure why it has not been
noticed before, but now a very modern Cygwin breaks on it, and the fix
is clearly correct. Backpatching to all live branches.

13 years agoTab completion for \pset format and \pset linestyle.
Robert Haas [Wed, 16 Mar 2011 21:46:15 +0000 (17:46 -0400)]
Tab completion for \pset format and \pset linestyle.

Pavel Stehule

13 years agoCorrections to CREATE FOREIGN TABLE documentation.
Robert Haas [Wed, 16 Mar 2011 12:11:24 +0000 (08:11 -0400)]
Corrections to CREATE FOREIGN TABLE documentation.

Omit incorrect = sign, and properly document server_name parameter.

Shigeru Hanada

13 years agoImprove handling of unknown-type literals in UNION/INTERSECT/EXCEPT.
Tom Lane [Wed, 16 Mar 2011 01:51:36 +0000 (21:51 -0400)]
Improve handling of unknown-type literals in UNION/INTERSECT/EXCEPT.

This patch causes unknown-type Consts to be coerced to the resolved output
type of the set operation at parse time.  Formerly such Consts were left
alone until late in the planning stage.  The disadvantage of that approach
is that it disables some optimizations, because the planner sees the set-op
leaf query as having different output column types than the overall set-op.
We saw an example of that in a recent performance gripe from Claudio
Freire.

Fixing such a Const requires scribbling on the leaf query in
transformSetOperationTree, but that should be all right since if the leaf
query's semantics depended on that output column, it would already have
resolved the unknown to something else.

Most of the bulk of this patch is a simple adjustment of
transformSetOperationTree's API so that upper levels can get at the
TargetEntry containing a Const to be replaced: it now returns a list of
TargetEntries, instead of just the bare expressions.

13 years agoRemove unused copy_dir() function from pg_upgrade.
Bruce Momjian [Wed, 16 Mar 2011 00:52:35 +0000 (20:52 -0400)]
Remove unused copy_dir() function from pg_upgrade.

13 years agoAdjust doc wording about the casting exceptions for composite types.
Bruce Momjian [Wed, 16 Mar 2011 00:47:31 +0000 (20:47 -0400)]
Adjust doc wording about the casting exceptions for composite types.

13 years agoAdd hint suggesting the use of EXECUTE CREATE TABLE ... AS for EXECUTE INTO.
Bruce Momjian [Wed, 16 Mar 2011 00:18:35 +0000 (20:18 -0400)]
Add hint suggesting the use of EXECUTE CREATE TABLE ... AS for EXECUTE INTO.

13 years agoFix spelling error in fuzzystrmatch docs.
Bruce Momjian [Tue, 15 Mar 2011 23:50:02 +0000 (19:50 -0400)]
Fix spelling error in fuzzystrmatch docs.

13 years agoWording adjustment for restart_after_crash entry
Bruce Momjian [Tue, 15 Mar 2011 16:43:39 +0000 (12:43 -0400)]
Wording adjustment for restart_after_crash entry

Specifically, mention that "restart" is disabled by this parameter.

13 years agoAdd database comments to template0 and postgres databases, and improve
Bruce Momjian [Tue, 15 Mar 2011 15:26:20 +0000 (11:26 -0400)]
Add database comments to template0 and postgres databases, and improve
the comments on the template1 database.  No catalog version bump because
they are just comments.

13 years agoMinor sync rep documentation improvements.
Robert Haas [Tue, 15 Mar 2011 15:25:04 +0000 (11:25 -0400)]
Minor sync rep documentation improvements.

- Make the name of the ID tag for the GUC entry match the GUC name.
- Clarify that synchronous_replication waits for xlog flush, not receipt.
- Mention that synchronous_replication won't wait if max_wal_senders=0.

13 years agoRemove 13 keywords that are used only for ROLE options.
Robert Haas [Tue, 15 Mar 2011 14:22:58 +0000 (10:22 -0400)]
Remove 13 keywords that are used only for ROLE options.

Review by Tom Lane.

13 years agoRemove special case allowing parameters to ident auth for initdb
Magnus Hagander [Mon, 14 Mar 2011 18:46:52 +0000 (19:46 +0100)]
Remove special case allowing parameters to ident auth for initdb

This was required in pre-8.4 versions to allow the specification of
"ident sameuser", but sameuser is no longer required. It could be extended
to allow all parameters in the future, but should then apply to all
methods and not just ident.

13 years agoAdjust regression test to avoid platform-dependent failure.
Tom Lane [Mon, 14 Mar 2011 16:52:03 +0000 (12:52 -0400)]
Adjust regression test to avoid platform-dependent failure.

We have a test that verifies that max(anyarray) will cope if the array
column elements aren't all the same array type.  However, it's now possible
for that to produce a collation-related error message instead of the
expected one, if the first two column elements happen to be of the same
type and it's one that expects to be given collation info.  Tweak the test
to ensure this doesn't happen.  Per buildfarm member pika.

13 years agoRemove duplicate time-based macros recently added.
Bruce Momjian [Mon, 14 Mar 2011 14:40:14 +0000 (10:40 -0400)]
Remove duplicate time-based macros recently added.

13 years agoSimplify list traversal logic in add_path().
Tom Lane [Sun, 13 Mar 2011 16:57:14 +0000 (12:57 -0400)]
Simplify list traversal logic in add_path().

Its mechanism for recovering after deleting the current list cell was
a bit klugy.  Borrow the technique used in other places.

13 years agoMake all comparisons done for/with statistics use the default collation.
Tom Lane [Sat, 12 Mar 2011 21:30:36 +0000 (16:30 -0500)]
Make all comparisons done for/with statistics use the default collation.

While this will give wrong answers when estimating selectivity for a
comparison operator that's using a non-default collation, the estimation
error probably won't be large; and anyway the former approach created
estimation errors of its own by trying to use a histogram that might have
been computed with some other collation.  So we'll adopt this simplified
approach for now and perhaps improve it sometime in the future.

This patch incorporates changes from Andres Freund to make sure that
selfuncs.c passes a valid collation OID to any datatype-specific function
it calls, in case that function wants collation information.  Said OID will
now always be DEFAULT_COLLATION_OID, but at least we won't get errors.

13 years agoUse "backend process" rather than "backend server", where appropriate.
Bruce Momjian [Sat, 12 Mar 2011 14:38:56 +0000 (09:38 -0500)]
Use "backend process" rather than "backend server", where appropriate.

13 years agoUse macros for time-based constants, rather than constants.
Bruce Momjian [Sat, 12 Mar 2011 14:31:18 +0000 (09:31 -0500)]
Use macros for time-based constants, rather than constants.

13 years agoAdd test case for collation mismatch in recursive query
Peter Eisentraut [Sat, 12 Mar 2011 08:07:23 +0000 (10:07 +0200)]
Add test case for collation mismatch in recursive query

This isn't very important by itself, but was left on my list of things
without test coverage for the collation feature.

13 years agoOn further reflection, we'd better do the same in int.c.
Tom Lane [Sat, 12 Mar 2011 00:04:02 +0000 (19:04 -0500)]
On further reflection, we'd better do the same in int.c.

We previously heard of the same problem in int24div(), so there's not a
good reason to suppose the problem is confined to cases involving int8.

13 years agoPut in some more safeguards against executing a division-by-zero.
Tom Lane [Fri, 11 Mar 2011 23:18:55 +0000 (18:18 -0500)]
Put in some more safeguards against executing a division-by-zero.

Add dummy returns before every potential division-by-zero in int8.c,
because apparently further "improvements" in gcc's optimizer have
enabled it to break functions that weren't broken before.

Aurelien Jarno, via Martin Pitt

13 years agoSplit CollateClause into separate raw and analyzed node types.
Tom Lane [Fri, 11 Mar 2011 21:27:51 +0000 (16:27 -0500)]
Split CollateClause into separate raw and analyzed node types.

CollateClause is now used only in raw grammar output, and CollateExpr after
parse analysis.  This is for clarity and to avoid carrying collation names
in post-analysis parse trees: that's both wasteful and possibly misleading,
since the collation's name could be changed while the parsetree still
exists.

Also, clean up assorted infelicities and omissions in processing of the
node type.

13 years agoIn docs, rename "backwards compatibility" to "backward compatibility"
Bruce Momjian [Fri, 11 Mar 2011 19:33:10 +0000 (14:33 -0500)]
In docs, rename "backwards compatibility" to "backward compatibility"
for consistency.

13 years agoCreate an explicit concept of collations that work for any encoding.
Tom Lane [Fri, 11 Mar 2011 18:20:11 +0000 (13:20 -0500)]
Create an explicit concept of collations that work for any encoding.

Use collencoding = -1 to represent such a collation in pg_collation.
We need this to make the "default" entry work sanely, and a later
patch will fix the C/POSIX entries to be represented this way instead
of duplicating them across all encodings.  All lookup operations now
search first for an entry that's database-encoding-specific, and then
for the same name with collencoding = -1.

Also some incidental code cleanup in collationcmds.c and pg_collation.c.

13 years agoRemove duplicate indexterm to silence openjade wrning.
Tom Lane [Fri, 11 Mar 2011 16:41:14 +0000 (11:41 -0500)]
Remove duplicate indexterm to silence openjade wrning.

13 years agoUpdate documentation on FreeBSD write cache control.
Bruce Momjian [Fri, 11 Mar 2011 16:36:42 +0000 (11:36 -0500)]
Update documentation on FreeBSD write cache control.

13 years agoDocument that to_char('FM') removes only trailing zeros.
Bruce Momjian [Fri, 11 Mar 2011 15:48:30 +0000 (10:48 -0500)]
Document that to_char('FM') removes only trailing zeros.

13 years agoDocument how listen_addresses can do only IPv4 or IPv6.
Bruce Momjian [Fri, 11 Mar 2011 15:31:25 +0000 (10:31 -0500)]
Document how listen_addresses can do only IPv4 or IPv6.

13 years agoIn ecpg preprocessor, don't try to look up constants in the test for
Heikki Linnakangas [Fri, 11 Mar 2011 13:44:40 +0000 (15:44 +0200)]
In ecpg preprocessor, don't try to look up constants in the test for
variable hiding. A constant is not a variable. It worked in most cases by
accident, because we add constants to the global list of variables (why?),
but float constants like 1.23 were interpreted as struct field references,
and not found.

Backpatch to 9.0, where the test for variable hiding was added.

13 years agoAdds index entries for session_user and pg_describe_object
Bruce Momjian [Fri, 11 Mar 2011 15:23:59 +0000 (10:23 -0500)]
Adds index entries for session_user and pg_describe_object
Removes extraneous closing parenthesis from pg_describe_object
Puts pg_describe_object and has_sequence_privilege in correct
alphabetical position in function listing

Thom Brown

13 years agoIn plpsql docs, use RAISE rather than undefined log function.
Bruce Momjian [Fri, 11 Mar 2011 13:47:44 +0000 (08:47 -0500)]
In plpsql docs, use RAISE rather than undefined log function.

13 years agoImprove extract(day) documentation with interval values.
Bruce Momjian [Fri, 11 Mar 2011 13:17:49 +0000 (08:17 -0500)]
Improve extract(day) documentation with interval values.

13 years agoFix "unparenthesized" mention in vacuum docs.
Bruce Momjian [Fri, 11 Mar 2011 12:57:53 +0000 (07:57 -0500)]
Fix "unparenthesized" mention in vacuum docs.

13 years agoUpdate kernel docs for Solaris 10.
Bruce Momjian [Fri, 11 Mar 2011 12:53:34 +0000 (07:53 -0500)]
Update kernel docs for Solaris 10.

Josh Berkus

13 years agoDocument that pg_dump --clean might generate some harmless errors on restore.
Bruce Momjian [Fri, 11 Mar 2011 12:05:38 +0000 (07:05 -0500)]
Document that pg_dump --clean might generate some harmless errors on restore.

13 years agoUpdate C comment about O_DIRECT and fsync().
Bruce Momjian [Fri, 11 Mar 2011 11:46:44 +0000 (06:46 -0500)]
Update C comment about O_DIRECT and fsync().

13 years agoDocument that libpq's PQgetResult() should be called after a fatal error
Bruce Momjian [Fri, 11 Mar 2011 10:55:13 +0000 (05:55 -0500)]
Document that libpq's PQgetResult() should be called after a fatal error
to fully process errors.

13 years agoDocument that the parenthesized VACUUM syntax is deprecated, not the
Bruce Momjian [Fri, 11 Mar 2011 10:34:09 +0000 (05:34 -0500)]
Document that the parenthesized VACUUM syntax is deprecated, not the
FREEZE functionality.

13 years agoAdd comment to tools/git-external-diff.
Bruce Momjian [Fri, 11 Mar 2011 10:06:31 +0000 (05:06 -0500)]
Add comment to tools/git-external-diff.

13 years agoChange "TIP" to "HINT" to match backend message style, in pg_ctl -m fast
Bruce Momjian [Fri, 11 Mar 2011 10:02:57 +0000 (05:02 -0500)]
Change "TIP" to "HINT" to match backend message style, in pg_ctl -m fast
suggestion.

13 years agoRemove pg_dump -X options that are only in 9.1 and not needed for
Bruce Momjian [Fri, 11 Mar 2011 03:58:18 +0000 (22:58 -0500)]
Remove pg_dump -X options that are only in 9.1 and not needed for
backward compatibility.

13 years agoAdd C comment that new new pg_dump -X options are to be created.
Bruce Momjian [Fri, 11 Mar 2011 03:36:14 +0000 (22:36 -0500)]
Add C comment that new new pg_dump -X options are to be created.

13 years agoWhen a smart pg_ctl shutdown fails, mention -m fast as a tip.
Bruce Momjian [Fri, 11 Mar 2011 03:04:00 +0000 (22:04 -0500)]
When a smart pg_ctl shutdown fails, mention -m fast as a tip.

13 years agoModify pg_test_fsync to match the behavior of git head in regards to
Bruce Momjian [Fri, 11 Mar 2011 01:25:41 +0000 (20:25 -0500)]
Modify pg_test_fsync to match the behavior of git head in regards to
O_DIRECT behavior.

13 years agoClarify C comment that O_SYNC/O_FSYNC are really the same settting, as
Bruce Momjian [Fri, 11 Mar 2011 01:02:52 +0000 (20:02 -0500)]
Clarify C comment that O_SYNC/O_FSYNC are really the same settting, as
opposed to O_DSYNC.

13 years agoReference doc "examples" section for pg_dump options that often need
Bruce Momjian [Fri, 11 Mar 2011 00:40:23 +0000 (19:40 -0500)]
Reference doc "examples" section for pg_dump options that often need
complex quoting, e.g. -t and -n.

13 years agoMore wording improvements for disallowed SQL function commands.
Bruce Momjian [Thu, 10 Mar 2011 23:50:44 +0000 (18:50 -0500)]
More wording improvements for disallowed SQL function commands.

13 years agoImprove SQL function disallowed command wording.
Bruce Momjian [Thu, 10 Mar 2011 23:47:44 +0000 (18:47 -0500)]
Improve SQL function disallowed command wording.

13 years agoClarify what commands are not allowed in SQL functions.
Bruce Momjian [Thu, 10 Mar 2011 23:46:21 +0000 (18:46 -0500)]
Clarify what commands are not allowed in SQL functions.

13 years agoAdd proper git-external-diff script to src/tools.
Bruce Momjian [Thu, 10 Mar 2011 23:06:13 +0000 (18:06 -0500)]
Add proper git-external-diff script to src/tools.

13 years agoRevert addition of third argument to format_type().
Tom Lane [Thu, 10 Mar 2011 22:30:18 +0000 (17:30 -0500)]
Revert addition of third argument to format_type().

Including collation in the behavior of that function promotes a world view
we do not want.  Moreover, it was producing the wrong behavior for pg_dump
anyway: what we want is to dump a COLLATE clause on attributes whose
attcollation is different from the underlying type, and likewise for
domains, and the function cannot do that for us.  Doing it the hard way
in pg_dump is a bit more tedious but produces more correct output.

In passing, fix initdb so that the initial entry in pg_collation is
properly pinned.  It was droppable before :-(

13 years agoMake error handling of synchronous_standby_names consistent.
Robert Haas [Thu, 10 Mar 2011 21:24:52 +0000 (16:24 -0500)]
Make error handling of synchronous_standby_names consistent.

It's not a good idea to kill the postmaster just because someone muffs
this, and it's not consistent with what we do for other, similar GUCs.

Fujii Masao, with a bit more hacking by me

13 years agoMore synchronous replication typo fixes.
Robert Haas [Thu, 10 Mar 2011 20:56:18 +0000 (15:56 -0500)]
More synchronous replication typo fixes.

Fujii Masao

13 years agoMore synchronous replication tweaks.
Robert Haas [Thu, 10 Mar 2011 20:38:39 +0000 (15:38 -0500)]
More synchronous replication tweaks.

SyncRepRequested() must check not only the value of the
synchronous_replication GUC but also whether max_wal_senders > 0.
Otherwise, we might end up waiting for sync rep even when there's no
possibility of a standby ever managing to connect.  There are some
existing cross-checks to prevent this, but they're not quite sufficient:
the user can start the server with max_wal_senders=0,
synchronous_standby_names='', and synchronous_replication=off and then
subsequent make synchronous_standby_names not empty using pg_ctl reload,
and then SET synchronous_standby=on, leading to an indefinite hang.

Along the way, rename the global variable for the synchronous_replication
GUC to match the name of the GUC itself, for clarity.

Report by Fujii Masao, though I didn't use his patch.

13 years agoRemove obsolete comment.
Robert Haas [Thu, 10 Mar 2011 20:00:20 +0000 (15:00 -0500)]
Remove obsolete comment.

In earlier versions of the sync rep patch, waiters removed themselves from
the queue, but now walsender removes them before doing the wakeup.

Report by Fujii Masao.

13 years agoMinor sync rep corrections.
Robert Haas [Thu, 10 Mar 2011 19:57:02 +0000 (14:57 -0500)]
Minor sync rep corrections.

Fujii Masao, with a bit of additional wordsmithing by me.

13 years agoEmit a LOG message when pausing at the recovery target.
Robert Haas [Thu, 10 Mar 2011 19:37:14 +0000 (14:37 -0500)]
Emit a LOG message when pausing at the recovery target.

Fujii Masao

13 years agoAdd missing index terms for recovery control functions.
Robert Haas [Thu, 10 Mar 2011 19:29:53 +0000 (14:29 -0500)]
Add missing index terms for recovery control functions.

Fujii Masao

13 years agoUse '=' when documenting long options.
Bruce Momjian [Thu, 10 Mar 2011 14:09:35 +0000 (09:09 -0500)]
Use '=' when documenting long options.

13 years agoReplication README updates.
Robert Haas [Thu, 10 Mar 2011 13:59:59 +0000 (08:59 -0500)]
Replication README updates.

Fujii Masao

13 years agoFix bugs in the isolation tester flex rules.
Heikki Linnakangas [Thu, 10 Mar 2011 07:01:27 +0000 (09:01 +0200)]
Fix bugs in the isolation tester flex rules.

Tom Lane pointed out that it was giving a warning: "-s option given but
default rule can be matched". That was because there was no rule to handle
newline in a quoted string. I made that throw an error.

Also, line number tracking was broken, giving incorrect line number on
error. Fixed that too.

13 years agoCleanup copyright years and file names in the header comments of some files.
Itagaki Takahiro [Thu, 10 Mar 2011 06:05:33 +0000 (15:05 +0900)]
Cleanup copyright years and file names in the header comments of some files.

13 years agoreplication/repl_gram.h needs to be cleaned too ...
Tom Lane [Thu, 10 Mar 2011 05:12:38 +0000 (00:12 -0500)]
replication/repl_gram.h needs to be cleaned too ...

13 years agoFix some oversights in distprep and maintainer-clean targets.
Tom Lane [Thu, 10 Mar 2011 05:03:26 +0000 (00:03 -0500)]
Fix some oversights in distprep and maintainer-clean targets.

At least two recent commits have apparently imagined that a comment in
a Makefile stating that something would be included in the distribution
tarball was sufficient to make it so.  They hadn't bothered to hook
into the upper maintainer-clean targets either.  Per bug #5923 from
Charles Johnson, in which it emerged that the 9.1alpha4 tarballs are
short a few files that should be there.

13 years agoMention gcc version in C comment.
Bruce Momjian [Thu, 10 Mar 2011 04:40:39 +0000 (23:40 -0500)]
Mention gcc version in C comment.

13 years agoRemove collation information from TypeName, where it does not belong.
Tom Lane [Thu, 10 Mar 2011 03:38:52 +0000 (22:38 -0500)]
Remove collation information from TypeName, where it does not belong.

The initial collations patch treated a COLLATE spec as part of a TypeName,
following what can only be described as brain fade on the part of the SQL
committee.  It's a lot more reasonable to treat COLLATE as a syntactically
separate object, so that it can be added in only the productions where it
actually belongs, rather than needing to reject it in a boatload of places
where it doesn't belong (something the original patch mostly failed to do).
In addition this change lets us meet the spec's requirement to allow
COLLATE anywhere in the clauses of a ColumnDef, and it avoids unfriendly
behavior for constructs such as "foo::type COLLATE collation".

To do this, pull collation information out of TypeName and put it in
ColumnDef instead, thus reverting most of the collation-related changes in
parse_type.c's API.  I made one additional structural change, which was to
use a ColumnDef as an intermediate node in AT_AlterColumnType AlterTableCmd
nodes.  This provides enough room to get rid of the "transform" wart in
AlterTableCmd too, since the ColumnDef can carry the USING expression
easily enough.

Also fix some other minor bugs that have crept in in the same areas,
like failure to copy recently-added fields of ColumnDef in copyfuncs.c.

While at it, document the formerly secret ability to specify a collation
in ALTER TABLE ALTER COLUMN TYPE, ALTER TYPE ADD ATTRIBUTE, and
ALTER TYPE ALTER ATTRIBUTE TYPE; and correct some misstatements about
what the default collation selection will be when COLLATE is omitted.

BTW, the three-parameter form of format_type() should go away too,
since it just contributes to the confusion in this area; but I'll do
that in a separate patch.

13 years agoDocument that server single-user mode does not do checkpoints or other
Bruce Momjian [Thu, 10 Mar 2011 01:19:59 +0000 (20:19 -0500)]
Document that server single-user mode does not do checkpoints or other
background processing.

13 years agoAdjust the permissions required for COMMENT ON ROLE.
Tom Lane [Wed, 9 Mar 2011 16:28:20 +0000 (11:28 -0500)]
Adjust the permissions required for COMMENT ON ROLE.

Formerly, any member of a role could change the role's comment, as of
course could superusers; but holders of CREATEROLE privilege could not,
unless they were also members.  This led to the odd situation that a
CREATEROLE holder could create a role but then could not comment on it.
It also seems a bit dubious to let an unprivileged user change his own
comment, let alone those of group roles he belongs to.  So, change the
rule to be "you must be superuser to comment on a superuser role, or
hold CREATEROLE to comment on non-superuser roles".  This is the same
as the privilege check for creating/dropping roles, and thus fits much
better with the rule for other object types, namely that only the owner
of an object can comment on it.

In passing, clean up the documentation for COMMENT a little bit.

Per complaint from Owen Jacobson and subsequent discussion.

13 years agoRemove '=' from initdb switch syntax.
Bruce Momjian [Wed, 9 Mar 2011 14:48:05 +0000 (09:48 -0500)]
Remove '=' from initdb switch syntax.

13 years agoImprove wording of initdb and pg_controldata manual pages.
Bruce Momjian [Wed, 9 Mar 2011 14:18:44 +0000 (09:18 -0500)]
Improve wording of initdb and pg_controldata manual pages.

gabrielle <gorthx@gmail.com>

13 years agoRemove 's' from recovery_target_timeline's' from the release note.
Itagaki Takahiro [Wed, 9 Mar 2011 11:00:11 +0000 (20:00 +0900)]
Remove 's' from recovery_target_timeline's' from the release note.

13 years agosynchronous_standby_names is a string parameter.
Itagaki Takahiro [Wed, 9 Mar 2011 10:49:16 +0000 (19:49 +0900)]
synchronous_standby_names is a string parameter.

13 years agoMake alpha release notes more consistent as regards periods.
Robert Haas [Wed, 9 Mar 2011 05:06:02 +0000 (00:06 -0500)]
Make alpha release notes more consistent as regards periods.

13 years agoUpdate alpha release notes for latest commits.
Robert Haas [Wed, 9 Mar 2011 04:59:52 +0000 (23:59 -0500)]
Update alpha release notes for latest commits.

13 years agoA bit more editing for collation documentation.
Tom Lane [Wed, 9 Mar 2011 03:50:15 +0000 (22:50 -0500)]
A bit more editing for collation documentation.

13 years agoFix file descriptor leaks in pg_upgrade in failure code paths.
Bruce Momjian [Wed, 9 Mar 2011 02:35:42 +0000 (21:35 -0500)]
Fix file descriptor leaks in pg_upgrade in failure code paths.

13 years agoCreate "replication and recovery" section in alpha release notes.
Robert Haas [Wed, 9 Mar 2011 01:55:29 +0000 (20:55 -0500)]
Create "replication and recovery" section in alpha release notes.

13 years agoAssorted editing for collation documentation.
Tom Lane [Tue, 8 Mar 2011 22:10:34 +0000 (17:10 -0500)]
Assorted editing for collation documentation.

I made a pass over this to familiarize myself with the feature, and found
some things that could be improved.

13 years agoFix parallel make when running make install before make all
Peter Eisentraut [Tue, 8 Mar 2011 21:49:59 +0000 (23:49 +0200)]
Fix parallel make when running make install before make all

In addition to the

all-foo-recurse: all-bar-recurse

dependencies that constraint the order of the rule execution, we need

install-foo-recurse: install-bar-recurse

dependencies in case one runs make install without a make all first,
as some people apparently do.

13 years agoAdd missing keywords to gram.y's unreserved_keywords list.
Tom Lane [Tue, 8 Mar 2011 21:43:02 +0000 (16:43 -0500)]
Add missing keywords to gram.y's unreserved_keywords list.

We really need an automated check for this ... and did VALIDATE really
need to become a keyword at all, rather than picking some other syntax
using existing keywords?

13 years agoIgnore files built by coverage builds
Peter Eisentraut [Tue, 8 Mar 2011 20:10:00 +0000 (22:10 +0200)]
Ignore files built by coverage builds

13 years agoFix overly strict assertion in SummarizeOldestCommittedSxact(). There's a
Heikki Linnakangas [Tue, 8 Mar 2011 19:01:29 +0000 (21:01 +0200)]
Fix overly strict assertion in SummarizeOldestCommittedSxact(). There's a
race condition where SummarizeOldestCommittedSxact() is called even though
another backend already cleared out all finished sxact entries. That's OK,
RegisterSerializableTransactionInt() can just retry getting a news xact
slot from the available-list when that happens.

Reported by YAMAMOTO Takashi, bug #5918.

13 years agoDon't throw a warning if vacuum sees PD_ALL_VISIBLE flag set on a page that
Heikki Linnakangas [Tue, 8 Mar 2011 18:13:52 +0000 (20:13 +0200)]
Don't throw a warning if vacuum sees PD_ALL_VISIBLE flag set on a page that
contains newly-inserted tuples that according to our OldestXmin are not
yet visible to everyone. The value returned by GetOldestXmin() is conservative,
and it can move backwards on repeated calls, so if we see that contradiction
between the PD_ALL_VISIBLE flag and status of tuples on the page, we have to
assume it's because an earlier vacuum calculated a higher OldestXmin value,
and all the tuples really are visible to everyone.

We have received several reports of this bug, with the "PD_ALL_VISIBLE flag
was incorrectly set in relation ..." warning appearing in logs. We were
finally able to hunt it down with David Gould's help to run extra diagnostics
in an environment where this happened frequently.

Also reword the warning, per Robert Haas' suggestion, to not imply that the
PD_ALL_VISIBLE flag is necessarily at fault, as it might also be a symptom
of corruption on a tuple header.

Backpatch to 8.4, where the PD_ALL_VISIBLE flag was introduced.

13 years agoAdjust CHAR() doc mention of pattern matching issues for trailing
Bruce Momjian [Tue, 8 Mar 2011 17:54:00 +0000 (12:54 -0500)]
Adjust CHAR() doc mention of pattern matching issues for trailing
spaces.

13 years agoDocument that char() ignores spaces in non-pattern comparisons, not in
Bruce Momjian [Tue, 8 Mar 2011 16:03:02 +0000 (11:03 -0500)]
Document that char() ignores spaces in non-pattern comparisons, not in
pattern comparisons such as LIKE and regex.

13 years agoAdded new version of ecpg's parser test script which was written by Andy Colson ...
Michael Meskes [Tue, 8 Mar 2011 10:27:32 +0000 (11:27 +0100)]
Added new version of ecpg's parser test script which was written by Andy Colson <andy@squeakycode.net>.

13 years agoTruncate predicate lock manager's SLRU lazily at checkpoint. That's safer
Heikki Linnakangas [Tue, 8 Mar 2011 10:07:29 +0000 (12:07 +0200)]
Truncate predicate lock manager's SLRU lazily at checkpoint. That's safer
than doing it aggressively whenever the tail-XID pointer is advanced, because
this way we don't need to do it while holding SerializableXactHashLock.

This also fixes bug #5915 spotted by YAMAMOTO Takashi, and removes an
obsolete comment spotted by Kevin Grittner.

13 years agoFix behavior when raising plpy.Fatal()
Peter Eisentraut [Mon, 7 Mar 2011 21:01:54 +0000 (23:01 +0200)]
Fix behavior when raising plpy.Fatal()

It should cause a elog(FATAL) error, and it fact it was simply causing
a elog(ERROR).

Jan Urbański