OSDN Git Service
Tom Lane [Sun, 19 Jun 2011 18:16:41 +0000 (14:16 -0400)]
Add REL9_1_STABLE to the set of branches tracked by git_changelog.
Tom Lane [Sun, 19 Jun 2011 18:00:48 +0000 (14:00 -0400)]
Fix thinko in previous patch to always update pg_class.reltuples/relpages.
I mis-simplified the test where ANALYZE decided if it could get away
without doing anything: under the new regime, that's never allowed. Per
bug #6068 from Jeff Janes. Back-patch to 8.4, just like previous patch.
Peter Eisentraut [Sat, 18 Jun 2011 21:37:30 +0000 (00:37 +0300)]
Capitalization fixes
Peter Eisentraut [Sat, 18 Jun 2011 21:34:32 +0000 (00:34 +0300)]
Make external_pid_file world readable
Magnus Hagander [Sat, 18 Jun 2011 16:26:35 +0000 (18:26 +0200)]
Remove redundant lib*dll.def rules from .gitignore
Since we now have a global rule in the root .gitignore,
there's no need to keep directory-specific ones as well.
Noted by Peter Eisentraut
Tom Lane [Fri, 17 Jun 2011 23:13:03 +0000 (19:13 -0400)]
Don't use "cp -i" in the example WAL archive_command.
This is a dangerous example to provide because on machines with GNU cp,
it will silently do the wrong thing and risk archive corruption. Worse,
during the 9.0 cycle somebody "improved" the discussion by removing the
warning that used to be there about that, and instead leaving the
impression that the command would work as desired on most Unixen.
It doesn't. Try to rectify the damage by providing an example that is safe
most everywhere, and then noting that you can try cp -i if you want but
you'd better test that.
In back-patching this to all supported branches, I also added an example
command for Windows, which wasn't provided before 9.0.
Andrew Dunstan [Fri, 17 Jun 2011 22:22:03 +0000 (18:22 -0400)]
Set FLEXIBLE_ARRAY_MEMBER to empty for MSVC.
Per gripe from Tom Lane. I have tested this with VC 2008, and assume
it will work with earlier versions.
Tom Lane [Fri, 17 Jun 2011 22:19:02 +0000 (18:19 -0400)]
Obtain table locks as soon as practical during pg_dump.
For some reason, when we (I) added table lock acquisition to pg_dump,
we didn't think about making it happen as soon as possible after the
start of the transaction. What with subsequent additions, there was
actually quite a lot going on before we got around to that; which sort
of defeats the purpose. Rearrange the order of calls in dumpSchema()
to close the risk window as much as we easily can. Back-patch to all
supported branches.
Robert Haas [Fri, 17 Jun 2011 18:28:45 +0000 (14:28 -0400)]
Add overflow checks to int4 and int8 versions of generate_series().
The previous code went into an infinite loop after overflow. In fact,
an overflow is not really an error; it just means that the current
value is the last one we need to return. So, just arrange to stop
immediately when overflow is detected.
Back-patch all the way.
Robert Haas [Fri, 17 Jun 2011 17:34:39 +0000 (13:34 -0400)]
Fix crash in CREATE UNLOGGED TABLE.
The code that created the init fork neglected to make sure that the
relation was open at the smgr level before attempting to invoke smgr.
This didn't happen every time; only when the relcache entry was rebuilt
along the way.
Per report from Garick Hamlin.
Robert Haas [Fri, 17 Jun 2011 13:12:19 +0000 (09:12 -0400)]
Fix minor thinko in ProcGlobalShmemSize().
There's no need to add space for startupBufferPinWaitBufId, because
it's part of the PROC_HDR object for which this function already
allocates space.
This has been wrong for a while, but the only consequence is that our
shared memory allocation is increased by 4 bytes, so no back-patch.
Robert Haas [Fri, 17 Jun 2011 04:54:08 +0000 (00:54 -0400)]
Fix typo.
Per Josh Kupershmidt and Tom Lane.
Andrew Dunstan [Fri, 17 Jun 2011 02:30:24 +0000 (22:30 -0400)]
Define FLEXIBLE_ARRAY_MEMBER for MSVC.
Tom Lane [Thu, 16 Jun 2011 21:03:58 +0000 (17:03 -0400)]
Index tuple data arrays using Anum_xxx symbolic constants instead of "i++".
We had already converted most places to this style, but this patch gets the
last few that were still doing it the old way. The main advantage is that
this exposes a greppable name for each target column, rather than having
to rely on comments (which a couple of places failed to provide anyhow).
Richard Hopkins, additional work by me to clean up update_attstats() too
Peter Eisentraut [Thu, 16 Jun 2011 20:38:46 +0000 (23:38 +0300)]
Avoid compiler warnings due to possibly unused variables
gcc 4.6 complains about these because of the new option
-Wunused-but-set-variable which comes in with -Wall, so cast them to
void, which avoids the warning.
Peter Eisentraut [Thu, 16 Jun 2011 19:39:09 +0000 (22:39 +0300)]
Start using flexible array members
Flexible array members are a C99 feature that avoids "cheating" in the
declaration of variable-length arrays at the end of structs. With
Autoconf support, this should be transparent for older compilers.
We start with one use in gist.h because gcc 4.6 started to raise a
warning there. Over time, it can be expanded to other places in the
source, but they will likely need some review of sizeof and offsetof
usage. The current change in gist.h appears to be safe in this
regard.
Heikki Linnakangas [Thu, 16 Jun 2011 18:12:56 +0000 (21:12 +0300)]
Update README-SSI. Add a section to describe the "dangerous structure" that
SSI is based on, as well as the optimizations about relative commit times
and read-only transactions. Plus a bunch of other misc fixes and
improvements.
Dan Ports
Alvaro Herrera [Thu, 16 Jun 2011 16:10:57 +0000 (12:10 -0400)]
Fix typo
Tom Lane [Thu, 16 Jun 2011 15:43:30 +0000 (11:43 -0400)]
Remove another no-longer-needed inclusion of predicate.h.
Heikki Linnakangas [Thu, 16 Jun 2011 13:16:34 +0000 (16:16 +0300)]
pgindent run of recent SSI changes. Also, remove an unnecessary #include.
Kevin Grittner
Simon Riggs [Thu, 16 Jun 2011 09:19:10 +0000 (10:19 +0100)]
Respect Hot Standby controls while recycling btree index pages.
Btree pages were recycled after VACUUM deletes all records on a
page and then a subsequent VACUUM occurs after the RecentXmin
horizon is reached. Using RecentXmin meant that we did not respond
correctly to the user controls provide to avoid Hot Standby
conflicts and so spurious conflicts could be generated in some
workload combinations. We now reuse pages only when we reach
RecentGlobalXmin, which can be much later in the presence of long
running queries and is also controlled by vacuum_defer_cleanup_age
and hot_standby_feedback.
Noah Misch and Simon Riggs
Tom Lane [Thu, 16 Jun 2011 01:45:23 +0000 (21:45 -0400)]
Use single quotes in preference to double quotes for protecting pathnames.
Per recommendation from Peter. Neither choice is bulletproof, but this
is the existing style and it does help prevent unexpected environment
variable substitution.
Tom Lane [Wed, 15 Jun 2011 23:05:11 +0000 (19:05 -0400)]
Rework parsing of ConstraintAttributeSpec to improve NOT VALID handling.
The initial commit of the ALTER TABLE ADD FOREIGN KEY NOT VALID feature
failed to support labeling such constraints as deferrable. The best fix
for this seems to be to fold NOT VALID into ConstraintAttributeSpec.
That's a bit more general than the documented syntax, but it allows
better-targeted syntax error messages.
In addition, do some mostly-but-not-entirely-cosmetic code review for
the whole NOT VALID patch.
Bruce Momjian [Wed, 15 Jun 2011 21:23:02 +0000 (17:23 -0400)]
In pg_upgrade, clean up code layout in validateDirectoryOption().
Tom Lane [Wed, 15 Jun 2011 18:05:22 +0000 (14:05 -0400)]
Fix failure to account for memory used by tuplestore_putvalues().
This oversight could result in a tuplestore using much more than the
intended amount of memory. It would only happen in a code path that loaded
a tuplestore via tuplestore_putvalues(), and many of those won't emit huge
amounts of data; but cases such as holdable cursors and plpgsql's RETURN
NEXT command could have the problem. The fix ensures that the tuplestore
will switch to write-to-disk mode when it overruns work_mem.
The potential overrun was finite, because we would still count the space
used by the tuple pointer array, so the tuplestore code would eventually
flip into write-to-disk mode anyway. When storing wide tuples we would
go far past the expected work_mem usage before that happened; but this
may account for the lack of prior reports.
Back-patch to 8.4, where tuplestore_putvalues was introduced.
Per bug #6061 from Yann Delorme.
Tom Lane [Wed, 15 Jun 2011 17:15:05 +0000 (13:15 -0400)]
Fix oversights in pg_basebackup's -z (compression) option.
The short-form -z switch didn't work, for lack of telling getopt_long
about it; and even if specified long-form, it failed to do anything,
because the various tests elsewhere in the file would take
Z_DEFAULT_COMPRESSION (which is -1) as meaning "don't compress".
Per bug #6060 from Shigehiro Honda, though I editorialized on his patch
a bit.
Heikki Linnakangas [Wed, 15 Jun 2011 10:35:28 +0000 (13:35 +0300)]
The rolled-back flag on serializable xacts was pointless and redundant with
the marked-for-death flag. It was only set for a fleeting moment while a
transaction was being cleaned up at rollback. All the places that checked
for the rolled-back flag should also check the marked-for-death flag, as
both flags mean that the transaction will roll back. I also renamed the
marked-for-death into "doomed", which is a lot shorter name.
Heikki Linnakangas [Wed, 15 Jun 2011 08:43:05 +0000 (11:43 +0300)]
Make non-MVCC snapshots exempt from predicate locking. Scans with non-MVCC
snapshots, like in REINDEX, are basically non-transactional operations. The
DDL operation itself might participate in SSI, but there's separate
functions for that.
Kevin Grittner and Dan Ports, with some changes by me.
Peter Eisentraut [Tue, 14 Jun 2011 20:45:00 +0000 (23:45 +0300)]
Allow psql \d tab completion to complete all relation kinds
This matches what \d actually accepts.
Peter Eisentraut [Tue, 14 Jun 2011 20:44:01 +0000 (23:44 +0300)]
Document that psql's \d also works with composite types
This has always been true, it was just never documented.
Bruce Momjian [Tue, 14 Jun 2011 20:46:52 +0000 (16:46 -0400)]
In pg_upgrade, document that link mode has to have data directories on
the same file system, and that authentication should lock out normal
users.
Per suggestsion from #postgresql irc channel.
Backpatch to 9.1.
Tom Lane [Tue, 14 Jun 2011 16:00:37 +0000 (12:00 -0400)]
Fix assorted issues with build and install paths containing spaces.
Apparently there is no buildfarm critter exercising this case after all,
because it fails in several places. With this patch, build, install,
check-world, and installcheck-world pass for me on OS X.
Peter Eisentraut [Tue, 14 Jun 2011 20:00:43 +0000 (23:00 +0300)]
Remove unused variable
The variable became obsolete in commit
68739ba856c52e6721d6cffec21f1bf0327a9a7b, but only gcc 4.6 shows the
warning.
Peter Eisentraut [Tue, 14 Jun 2011 19:53:02 +0000 (22:53 +0300)]
Add comment that attributes.is_nullable was removed from SQL standard
We don't have to remove the column if no one is bothered, but it's
useful to comment on it in case someone looks for it in newer
standards versions.
Bruce Momjian [Tue, 14 Jun 2011 18:53:35 +0000 (14:53 -0400)]
In pg_upgrade, check there are no prepared transactions.
Alvaro Herrera [Tue, 14 Jun 2011 17:48:23 +0000 (13:48 -0400)]
Fix grammatical mistake introduced by previous commit
Per note from Tom
Alvaro Herrera [Tue, 14 Jun 2011 15:20:52 +0000 (11:20 -0400)]
Mention DROP TABLE as well as ALTER TABLE NO INHERIT
... when talking about how good they are in replacement of bulk DELETE
in partitioned setups.
The original wording was a bit confusing.
Per an observation from David Wheeler.
Robert Haas [Tue, 14 Jun 2011 15:05:54 +0000 (11:05 -0400)]
Improve accuracy of ON_ERROR_STOP documentation.
Per a gripe from Tom Lane.
Heikki Linnakangas [Tue, 14 Jun 2011 12:16:36 +0000 (15:16 +0300)]
Oops, forgot to change the order of entries in 2PC callback arrays when I
renumbered the resource managers. This should fix the buildfarm..
Heikki Linnakangas [Tue, 14 Jun 2011 09:32:45 +0000 (12:32 +0300)]
Renumber 2PC resource managers so that compared to 9.0, predicate lock rmgr
is added to the end, and existing resource managers keep their old ids.
We're not going to guarantee on-disk compatibility for 2PC state files over
major releases, but it seems better to avoid changing the ids them anyway.
It will help anyone who might want to write external tools to inspect the
state files to work with files from different versions, if nothing else.
Per complaint from Tom Lane.
Peter Eisentraut [Tue, 14 Jun 2011 04:34:00 +0000 (07:34 +0300)]
Move parse2.pl to parse.pl
We have a SCM, so we don't need to keep old versions of files around.
Bruce Momjian [Tue, 14 Jun 2011 00:54:45 +0000 (20:54 -0400)]
Document that bits to the right of the CIDR mask "should" be zero, not
"must".
Alvaro Herrera [Mon, 13 Jun 2011 21:50:30 +0000 (17:50 -0400)]
Fix aboriginal copy-paste mistake in error message
Spotted by Jaime Casanova
Alvaro Herrera [Mon, 13 Jun 2011 21:12:26 +0000 (17:12 -0400)]
Expand warnings on locks acquired by CREATE INDEX CONCURRENTLY
The previous wording wasn't explicit enough, which could misled readers
into thinking that the locks acquired are more restricted in nature than
they really are. The resulting optimism can be damaging to morale when
confronted with reality, as has been observed in the field.
Greg Smith
Robert Haas [Mon, 13 Jun 2011 17:02:54 +0000 (13:02 -0400)]
Remove parentheses from mention of current_schemas function.
This is more consistent with what we do elsewhere, and hopefully avoids
creating the perception that current_schemas takes no arguments.
As suggested by Brendan Jurd
Robert Haas [Mon, 13 Jun 2011 16:59:04 +0000 (12:59 -0400)]
Clarify that NATURAL without matching columns is like CROSS JOIN.
As suggested by Grzegorz Szpetkowski.
Robert Haas [Mon, 13 Jun 2011 16:37:49 +0000 (12:37 -0400)]
Add doc cross-reference to search_path discussion of current_schemas().
Brendan Jurd
Robert Haas [Mon, 13 Jun 2011 16:23:42 +0000 (12:23 -0400)]
Document GUC context for synchronous_standby_names/hot_standby_feedback.
Fujii Masao
Robert Haas [Mon, 13 Jun 2011 14:23:14 +0000 (10:23 -0400)]
In ALTER EXTENSION docs, document left_type/right_type.
Noted by Daniele Varrazzo.
Robert Haas [Mon, 13 Jun 2011 14:14:02 +0000 (10:14 -0400)]
Markup fix.
Fujii Masao
Robert Haas [Mon, 13 Jun 2011 12:28:41 +0000 (08:28 -0400)]
More foreign table documentation improvements.
Shigeru Hanada, with some additional wordsmithing by me
Heikki Linnakangas [Sun, 12 Jun 2011 19:49:33 +0000 (22:49 +0300)]
Remove now-unnecessary casts.
Kevin Grittner
Robert Haas [Sun, 12 Jun 2011 04:18:43 +0000 (00:18 -0400)]
Corrections to ALTER FOREIGN TABLE documentation.
Shigeru Hanada, with a minor grammar correction.
Robert Haas [Sun, 12 Jun 2011 04:07:04 +0000 (00:07 -0400)]
Code cleanup for InitProcGlobal.
The old code creates three separate arrays when only one is needed,
using two different shmem allocation functions for no obvious reason.
It also strangely splits up the initialization of AuxilaryProcs
between the top and bottom of the function to no evident purpose.
Review by Tom Lane.
Robert Haas [Sun, 12 Jun 2011 03:52:44 +0000 (23:52 -0400)]
Tab completion improvements for COMMENT.
These pertain to object types introduced in PostgreSQL 9.1, so back-patch.
Josh Kupershmidt, with some kibitzing by me.
Tom Lane [Sat, 11 Jun 2011 21:46:49 +0000 (17:46 -0400)]
Stamp HEAD as 9.2devel.
Bruce Momjian [Sat, 11 Jun 2011 14:00:28 +0000 (10:00 -0400)]
Add C comment mentioning pg_stat_activity.procpid should have been
called 'pid'.
Tom Lane [Fri, 10 Jun 2011 21:03:03 +0000 (17:03 -0400)]
Work around gcc 4.6.0 bug that breaks WAL replay.
ReadRecord's habit of using both direct references to tmpRecPtr and
references to *RecPtr (which is pointing at tmpRecPtr) triggers an
optimization bug in gcc 4.6.0, which apparently has forgotten about
aliasing rules. Avoid the compiler bug, and make the code more readable
to boot, by getting rid of the direct references. Improve the comments
while at it.
Back-patch to all supported versions, in case they get built with 4.6.0.
Tom Lane, with some cosmetic suggestions from Alex Hunsaker
Heikki Linnakangas [Fri, 10 Jun 2011 20:15:05 +0000 (23:15 +0300)]
Fix locking while setting flags in MySerializableXact.
Even if a flag is modified only by the backend owning the transaction, it's
not safe to modify it without a lock. Another backend might be setting or
clearing a different flag in the flags field concurrently, and that
operation might be lost because setting or clearing a bit in a word is not
atomic.
Make did-write flag a simple backend-private boolean variable, because it
was only set or tested in the owning backend (except when committing a
prepared transaction, but it's not worthwhile to optimize for the case of a
read-only prepared transaction). This also eliminates the need to add
locking where that flag is set.
Also, set the did-write flag when doing DDL operations like DROP TABLE or
TRUNCATE -- that was missed earlier.
Alvaro Herrera [Fri, 10 Jun 2011 19:27:38 +0000 (15:27 -0400)]
Add comment about pg_ctl stop
Alvaro Herrera [Fri, 10 Jun 2011 17:43:02 +0000 (13:43 -0400)]
Use "transient" files for blind writes, take 2
"Blind writes" are a mechanism to push buffers down to disk when
evicting them; since they may belong to different databases than the one
a backend is connected to, the backend does not necessarily have a
relation to link them to, and thus no way to blow them away. We were
keeping those files open indefinitely, which would cause a problem if
the underlying table was deleted, because the operating system would not
be able to reclaim the disk space used by those files.
To fix, have bufmgr mark such files as transient to smgr; the lower
layer is allowed to close the file descriptor when the current
transaction ends. We must be careful to have any other access of the
file to remove the transient markings, to prevent unnecessary expensive
system calls when evicting buffers belonging to our own database (which
files we're likely to require again soon.)
This commit fixes a bug in the previous one, which neglected to cleanly
handle the LRU ring that fd.c uses to manage open files, and caused an
unacceptable failure just before beta2 and was thus reverted.
Alvaro Herrera [Fri, 10 Jun 2011 17:37:06 +0000 (13:37 -0400)]
Use a constant sprintf format to silence compiler warning
Heikki Linnakangas [Fri, 10 Jun 2011 14:22:00 +0000 (17:22 +0300)]
Small comment fixes and enhancements.
Bruce Momjian [Fri, 10 Jun 2011 00:51:10 +0000 (20:51 -0400)]
Mention "pg_ctl stop" in pgindent README instructions.
Tom Lane [Thu, 9 Jun 2011 23:40:42 +0000 (19:40 -0400)]
Tag 9.1beta2.
Bruce Momjian [Thu, 9 Jun 2011 22:47:37 +0000 (18:47 -0400)]
Ecpg doc cleanups, as reported by Marc Cousin.
Tom Lane [Thu, 9 Jun 2011 22:44:05 +0000 (18:44 -0400)]
Update 9.1 release notes to reflect commits through today.
Also do some desultory copy-editing on the notes.
Alvaro Herrera [Thu, 9 Jun 2011 20:41:44 +0000 (16:41 -0400)]
Revert "Use "transient" files for blind writes"
This reverts commit
54d9e8c6c19cbefa8fb42ed3442a0a5327590ed3, which
caused a failure on the buildfarm. Not a good thing to have just before
a beta release.
Alvaro Herrera [Thu, 9 Jun 2011 17:41:12 +0000 (13:41 -0400)]
Use "transient" files for blind writes
"Blind writes" are a mechanism to push buffers down to disk when
evicting them; since they may belong to different databases than the one
a backend is connected to, the backend does not necessarily have a
relation to link them to, and thus no way to blow them away. We were
keeping those files open indefinitely, which would cause a problem if
the underlying table was deleted, because the operating system would not
be able to reclaim the disk space used by those files.
To fix, have bufmgr mark such files as transient to smgr; the lower
layer is allowed to close the file descriptor when the current
transaction ends. We must be careful to have any other access of the
file to remove the transient markings, to prevent unnecessary expensive
system calls when evicting buffers belonging to our own database (which
files we're likely to require again soon.)
Peter Eisentraut [Thu, 9 Jun 2011 20:01:31 +0000 (23:01 +0300)]
Translation updates for 9.1beta2
Heikki Linnakangas [Thu, 9 Jun 2011 17:30:59 +0000 (20:30 +0300)]
Fix the truncation logic of the OldSerXid SLRU mechanism. We can't pass
SimpleLruTruncate() a page number that's "in the future", because it will
issue a warning and refuse to truncate anything. Instead, we leave behind
the latest segment. If the slru is not needed before XID wrap-around, the
segment will appear as new again, and not be cleaned up until it gets old
enough again. That's a bit unpleasant, but better than not cleaning up
anything.
Also, fix broken calculation to check and warn if the span of the OldSerXid
SLRU is getting too large to fit in the 64k SLRU pages that we have
available. It was not XID wraparound aware.
Kevin Grittner and me.
Bruce Momjian [Thu, 9 Jun 2011 18:32:50 +0000 (14:32 -0400)]
Pgindent run before 9.1 beta2.
Bruce Momjian [Thu, 9 Jun 2011 18:01:49 +0000 (14:01 -0400)]
Update typedef list for upcoming pgindent run.
Magnus Hagander [Thu, 9 Jun 2011 16:21:38 +0000 (18:21 +0200)]
Use the correct eventlog severity for error
Magnus Hagander [Thu, 9 Jun 2011 16:18:45 +0000 (18:18 +0200)]
Support silent mode for service registrations on win32
Using -s when registering a service will now suppress
the application eventlog entries stating that the service
is starting and started.
MauMau
Magnus Hagander [Thu, 9 Jun 2011 16:11:47 +0000 (18:11 +0200)]
Add gitignore for mingw/cygwin build outputs
Noted by Radosław Smogura
Heikki Linnakangas [Thu, 9 Jun 2011 09:10:14 +0000 (12:10 +0300)]
Mark the SLRU page as dirty when setting an entry in pg_serial. In the
passing, fix an incorrect comment.
Peter Eisentraut [Thu, 9 Jun 2011 04:24:14 +0000 (07:24 +0300)]
Fix documentation of information_schema.element_types
The documentation of the columns collection_type_identifier and
dtd_identifier was wrong. This effectively reverts commits
8e1ccad51901e83916dae297cd9afa450957a36c and
57352df66d3a0885899d39c04c067e63c7c0ba30 and updates the name
array_type_identifier (the name in SQL:1999) to
collection_type_identifier.
closes bug #5926
Tom Lane [Wed, 8 Jun 2011 19:24:27 +0000 (15:24 -0400)]
Make citext's equality and hashing functions collation-insensitive.
This is an ugly hack to get around the fact that significant parts of the
core backend assume they don't need to worry about passing collation to
equality and hashing functions. That's true for the core string datatypes,
but citext should ideally have equality behavior that depends on the
specified collation's LC_CTYPE. However, there's no chance of fixing the
core before 9.2, so we'll have to live with this compromise arrangement for
now. Per bug #6053 from Regina Obe.
The code changes in this commit should be reverted in full once the core
code is up to speed, but be careful about reverting the docs changes:
I fixed a number of obsolete statements while at it.
Peter Eisentraut [Wed, 8 Jun 2011 17:51:25 +0000 (20:51 +0300)]
Reorder pg_ctl promote after pg_ctl status
Since start/stop/restart/reload/status is a kind of standard command
set, it seems odd to insert the special-purpose "promote" in between
the closely related "restart" and "reload". So put it after "status"
in code and documentation.
Put the documentation of the -U option in some sensible place.
Rewrite the synopsis sentence in help and documentation to make it
less of a growing mouthful.
Tom Lane [Wed, 8 Jun 2011 16:52:12 +0000 (12:52 -0400)]
Allow domains over arrays to match ANYARRAY parameters again.
This use-case was broken in commit
529cb267a6843a6a8190c86b75d091771d99d6a9
of 2010-10-21, in which I commented "For the moment, we just forbid such
matching. We might later wish to insert an automatic downcast to the
underlying array type, but such a change should also change matching of
domains to ANYELEMENT for consistency". We still lack consensus about what
to do with ANYELEMENT; but not matching ANYARRAY is a clear loss of
functionality compared to prior releases, so let's go ahead and make that
happen. Per complaint from Regina Obe and extensive subsequent discussion.
Heikki Linnakangas [Wed, 8 Jun 2011 10:47:21 +0000 (13:47 +0300)]
Make DDL operations play nicely with Serializable Snapshot Isolation.
Truncating or dropping a table is treated like deletion of all tuples, and
check for conflicts accordingly. If a table is clustered or rewritten by
ALTER TABLE, all predicate locks on the heap are promoted to relation-level
locks, because the tuple or page ids of any existing tuples will change and
won't be valid after rewriting the table. Arguably ALTER TABLE should be
treated like a mass-UPDATE of every row, but if you e.g change the datatype
of a column, you could also argue that it's just a change to the physical
layout, not a logical change. Reindexing promotes all locks on the index to
relation-level lock on the heap.
Kevin Grittner, with a lot of cosmetic changes by me.
Robert Haas [Wed, 8 Jun 2011 02:12:44 +0000 (22:12 -0400)]
Complain politely about access temp/unlogged tables during recovery.
This has never been supported, but we previously let md.c issue the
complaint for us at whatever point we tried to examine the backing file.
Now we print a nicer error message.
Per bug #6041, reported by Emanuel, and extensive discussion with Tom
Lane over where to put the check.
Alvaro Herrera [Wed, 8 Jun 2011 01:39:43 +0000 (21:39 -0400)]
Revert psql bits to display NOT VALID for FKs
These are superseded by pg_get_constraintdef's ability to display the
same when appropriate, which is a better place to do it anyway.
Heikki Linnakangas [Tue, 7 Jun 2011 06:54:24 +0000 (09:54 +0300)]
Make ascii-art in comments pgindent-safe, and some other formatting changes.
Kevin Grittner
Tom Lane [Tue, 7 Jun 2011 04:08:31 +0000 (00:08 -0400)]
Fix rewriter to cope (more or less) with CTEs in the query being rewritten.
Since the original implementation of CTEs only allowed them in SELECT
queries, the rule rewriter did not expect to find any CTEs in statements
being rewritten by ON INSERT/UPDATE/DELETE rules. We had dealt with this
to some extent but the code was still several bricks shy of a load, as
illustrated in bug #6051 from Jehan-Guillaume de Rorthais.
In particular, we have to be able to copy CTEs from the original query's
cteList into that of a rule action, in case the rule action references the
CTE (which it pretty much always will). This also implies we were doing
things in the wrong order in RewriteQuery: we have to recursively rewrite
the CTE queries before expanding the main query, so that we have the
rewritten queries available to copy.
There are unpleasant limitations yet to resolve here, but at least we now
throw understandable FEATURE_NOT_SUPPORTED errors for them instead of just
failing with bizarre implementation-dependent errors. In particular, we
can't handle propagating the same CTE into multiple post-rewrite queries
(because then the CTE would be evaluated multiple times), and we can't cope
with conflicts between CTE names in the original query and in the rule
actions.
Tom Lane [Mon, 6 Jun 2011 02:30:04 +0000 (22:30 -0400)]
Reset reindex-in-progress state before reverifying an exclusion constraint.
This avoids an Assert failure when we try to use ordinary index fetches
while checking for exclusion conflicts. Per report from Noah Misch.
No need for back-patch because the Assert wasn't there before 9.1.
Andrew Dunstan [Sat, 4 Jun 2011 23:32:10 +0000 (19:32 -0400)]
Allow building with perl 5.14.
Patch from Alex Hunsaker.
Peter Eisentraut [Sat, 4 Jun 2011 20:12:27 +0000 (23:12 +0300)]
Fix documentation reference to "above" example
found by Thom Brown
Peter Eisentraut [Sat, 4 Jun 2011 19:29:26 +0000 (22:29 +0300)]
More ECPG documentation fixes
Marc Cousin
Peter Eisentraut [Sat, 4 Jun 2011 19:11:20 +0000 (22:11 +0300)]
ECPG documentation fix
Marc Cousin, Satoshi Nagayasu
Tom Lane [Sat, 4 Jun 2011 19:48:17 +0000 (15:48 -0400)]
Expose the "*VALUES*" alias that we generate for a stand-alone VALUES list.
We were trying to make that strictly an internal implementation detail,
but it turns out that it's exposed anyway when dumping a view defined
like
CREATE VIEW test_view AS VALUES (1), (2), (3) ORDER BY 1;
This comes out as
CREATE VIEW ... ORDER BY "*VALUES*".column1;
which fails to parse when reloading the dump.
Hacking ruleutils.c to suppress the column qualification looks like it'd
be a risky business, so instead promote the RTE alias to full-fledged
usability.
Per bug #6049 from Dylan Adams. Back-patch to all supported branches.
Alvaro Herrera [Thu, 2 Jun 2011 16:39:53 +0000 (12:39 -0400)]
Fix pg_get_constraintdef to cope with NOT VALID constraints
This case was missed when NOT VALID constraints were first introduced in
commit
722bf7017bbe796decc79c1fde03e7a83dae9ada by Simon Riggs on
2011-02-08. Among other things, it causes pg_dump to omit the NOT VALID
flag when dumping such constraints, which may cause them to fail to
load afterwards, if they contained values failing the constraint.
Per report from Thom Brown.
Tom Lane [Fri, 3 Jun 2011 19:38:12 +0000 (15:38 -0400)]
Fix failure to check whether a rowtype's component types are sortable.
The existence of a btree opclass accepting composite types caused us to
assume that every composite type is sortable. This isn't true of course;
we need to check if the column types are all sortable. There was logic
for this for the case of array comparison (ie, check that the element
type is sortable), but we missed the point for rowtypes. Per Teodor's
report of an ANALYZE failure for an unsortable composite type.
Rather than just add some more ad-hoc logic for this, I moved knowledge of
the issue into typcache.c. The typcache will now only report out array_eq,
record_cmp, and friends as usable operators if the array or composite type
will work with those functions.
Unfortunately we don't have enough info to do this for anonymous RECORD
types; in that case, just assume it will work, and take the runtime failure
as before if it doesn't.
This patch might be a candidate for back-patching at some point, but
given the lack of complaints from the field, I'd rather just test it in
HEAD for now.
Note: most of the places touched in this patch will need further work
when we get around to supporting hashing of record types.
Peter Eisentraut [Fri, 3 Jun 2011 19:23:50 +0000 (22:23 +0300)]
Truncate id to <=44 characters
This is the original DocBook SGML limit, but apparently most
installations have changed it or ignore it, which is why few people
have run into this problem.
pointed out by Brendan Jurd
Heikki Linnakangas [Fri, 3 Jun 2011 09:45:42 +0000 (12:45 +0300)]
SSI comment fixes and enhancements. Notably, document that the conflict-out
flag actually means that the transaction has a conflict out to a transaction
that committed before the flagged transaction.
Kevin Grittner
Tom Lane [Fri, 3 Jun 2011 02:24:08 +0000 (22:24 -0400)]
Need to list getpeereid.c in .gitignore, too ...
Bruce Momjian [Thu, 2 Jun 2011 23:32:41 +0000 (19:32 -0400)]
Use proper SGML doc entities rather than angle-brackets.
Marco Nenciarini
Tom Lane [Thu, 2 Jun 2011 22:37:57 +0000 (18:37 -0400)]
Handle domains when checking for recursive inclusion of composite types.
We need this now because we allow domains over arrays, and we'll probably
allow domains over composites pretty soon, which makes the problem even
more obvious.
Although domains over arrays also exist in previous versions, this does not
need to be back-patched, because the coding used in older versions
successfully "looked through" domains over arrays. The problem is exposed
by not treating a domain as having a typelem.
Problem identified by Noah Misch, though I did not use his patch, since
it would require additional work to handle domains over composites that
way. This approach is more future-proof.
Tom Lane [Thu, 2 Jun 2011 21:27:30 +0000 (17:27 -0400)]
Looks like we can't declare getpeereid on Windows anyway.
... for lack of the uid_t and gid_t typedefs. Per buildfarm.