Bruce Momjian [Tue, 6 May 2008 21:57:46 +0000 (21:57 +0000)]
Update libpqxx URL in README.
Bruce Momjian [Tue, 6 May 2008 16:14:29 +0000 (16:14 +0000)]
Done:
> * -Add function to report the time of the most recent server reload
Bruce Momjian [Tue, 6 May 2008 00:47:53 +0000 (00:47 +0000)]
Done:
> o -Improve display of enums to show valid enum values
Magnus Hagander [Mon, 5 May 2008 19:31:40 +0000 (19:31 +0000)]
Add more dependencies from libpgport required by
standalone msvc build of libpq.
Hiroshi Saito
Andrew Dunstan [Mon, 5 May 2008 01:38:08 +0000 (01:38 +0000)]
Document extra information provided by psql's \dT+ (not \dt+ as per recent commit).
Andrew Dunstan [Mon, 5 May 2008 01:21:03 +0000 (01:21 +0000)]
Display ACLS using multiple lines for psql's \z. Brendan Jurd.
Andrew Dunstan [Mon, 5 May 2008 00:11:31 +0000 (00:11 +0000)]
Show enum values in psql's \dt+. David Fetter.
Tom Lane [Sun, 4 May 2008 23:19:24 +0000 (23:19 +0000)]
Add timestamp and timestamptz versions of generate_series().
Hitoshi Harada
Tom Lane [Sun, 4 May 2008 21:13:36 +0000 (21:13 +0000)]
Add pg_conf_load_time() function to report when the Postgres configuration
files were last loaded.
George Gensure
Tom Lane [Sun, 4 May 2008 16:42:41 +0000 (16:42 +0000)]
Use new cstring/text conversion functions in some additional places.
These changes assume that the varchar and xml data types are represented
the same as text. (I did not, however, accept the portions of the proposed
patch that wanted to assume bytea is the same as text --- tgl.)
Brendan Jurd
Andrew Dunstan [Sun, 4 May 2008 08:32:21 +0000 (08:32 +0000)]
add ; missing due to my carelessness
Andrew Dunstan [Sun, 4 May 2008 03:46:08 +0000 (03:46 +0000)]
Place statement timeout more appropriately, per gripe from Tom.
Andrew Dunstan [Sat, 3 May 2008 23:32:32 +0000 (23:32 +0000)]
Disable statement timeouts while making or restoring dumps.
Joshua Drake.
Tom Lane [Sat, 3 May 2008 23:19:20 +0000 (23:19 +0000)]
The 8.2 patch that added support for an alias on the target table of
UPDATE/DELETE forgot to teach ruleutils.c to display the alias.
Per bug #4141 from Mathias Seiler.
Tom Lane [Sat, 3 May 2008 02:47:48 +0000 (02:47 +0000)]
Fix plpython to not get totally confused by OUT arguments. (It still doesn't
support multiple OUT arguments, though.)
Hannu Krosing
Andrew Dunstan [Sat, 3 May 2008 00:24:06 +0000 (00:24 +0000)]
Provide for MSVC config equivalents of recently added configure options. Remove
any hardcoding of those options. Along the way, reorder the expression used to
calculate RELSEG_SIZE to make it slightly clearer. For now wal_segsize is only
allowed to have a value of 1 on Windows - we can relax that when we get full
large file support in the backend.
Tom Lane [Sat, 3 May 2008 00:11:36 +0000 (00:11 +0000)]
Support RETURN QUERY EXECUTE in plpgsql.
Pavel Stehule
Tom Lane [Fri, 2 May 2008 22:50:21 +0000 (22:50 +0000)]
Update documentation for psql relation-size-in-\dt+ patch.
Tom Lane [Fri, 2 May 2008 21:26:10 +0000 (21:26 +0000)]
Allow the planner's estimate of the fraction of a cursor's rows that will be
retrieved to be controlled through a GUC variable.
Robert Hell
Tom Lane [Fri, 2 May 2008 19:52:37 +0000 (19:52 +0000)]
Extend yesterday's patch making BLCKSZ and RELSEG_SIZE configurable to also
let XLOG_BLCKSZ and XLOG_SEG_SIZE be set via configure. Per a proposal by
Mark Wong, though I thought it better to call the switches after "wal" rather
than "xlog".
Peter Eisentraut [Fri, 2 May 2008 14:16:24 +0000 (14:16 +0000)]
Make the AWK default value also take effect if $AWK is an empty string.
Heikki Linnakangas [Fri, 2 May 2008 10:16:16 +0000 (10:16 +0000)]
Show relation size in verbose (e.g. \d+) output.
Dickson S. Guedes
Peter Eisentraut [Fri, 2 May 2008 09:27:51 +0000 (09:27 +0000)]
One-letter backslash commands now need a space before the first argument.
This has been the only documented and encouraged syntax for a long time, and
with extension facilities such as aliases being proposed, it is a good time to
clean up the legacy syntax a bit.
Author: Bernd Helmle <mailings@oopsware.de>
Tom Lane [Fri, 2 May 2008 03:41:46 +0000 (03:41 +0000)]
Sigh ... pg_config.h.win32 needs to define BLCKSZ and RELSEG_SIZE now.
Tom Lane [Fri, 2 May 2008 01:08:27 +0000 (01:08 +0000)]
Remove the recently added USE_SEGMENTED_FILES option, and indeed remove all
support for a nonsegmented mode from md.c. Per recent discussions, there
doesn't seem to be much value in a "never segment" option as opposed to
segmenting with a suitably large segment size. So instead provide a
configure-time switch to set the desired segment size in units of gigabytes.
While at it, expose a configure switch for BLCKSZ as well.
Zdenek Kotala
Tom Lane [Thu, 1 May 2008 20:05:08 +0000 (20:05 +0000)]
Add SGT (Singapore time) to the default list of known timezone
abbreviations. It doesn't conflict with any other abbreviation
so there seems no reason not to include it. Per a recent gripe.
Tom Lane [Thu, 1 May 2008 19:55:40 +0000 (19:55 +0000)]
Make the minimum allowed value of work_mem be 64KB always, rather than having
it vary with BLCKSZ as before. This agrees with what the documentation says,
and avoids a regression test problem when BLCKSZ is larger than default.
Per recent discussion.
Tom Lane [Wed, 30 Apr 2008 23:28:32 +0000 (23:28 +0000)]
Fix nodeTidscan.c to not trigger an error if the block number portion of
a user-supplied TID is out of range for the relation. This is needed to
preserve compatibility with our pre-8.3 behavior, and it is sensible anyway
since if the query were implemented by brute force rather than optimized
into a TidScan, the behavior for a non-existent TID would be zero rows out,
never an error. Per gripe from Gurjeet Singh.
Alvaro Herrera [Wed, 30 Apr 2008 12:38:49 +0000 (12:38 +0000)]
Add example showing how to remove a password from a role.
Andreas Scherbaum
Tom Lane [Tue, 29 Apr 2008 20:44:49 +0000 (20:44 +0000)]
Some minor further cleanup around A_Const. Don't attach a typecast in
makeFloatConst, and avoid "manual" construction of A_Const nodes in grammar
productions, in favor of using makeXXXConst subroutines.
Alvaro Herrera [Tue, 29 Apr 2008 19:37:04 +0000 (19:37 +0000)]
Fix REASSIGN OWNED so that it works on procedural languages too.
The capability for changing language owners is new in 8.3, so that's how
far back this needs to be backpatched.
Per bug #4132 by Kirill Simonov.
Alvaro Herrera [Tue, 29 Apr 2008 14:59:17 +0000 (14:59 +0000)]
Remove typename from A_Const.
Brendan Jurd, minor editorialization by me.
Alvaro Herrera [Tue, 29 Apr 2008 13:00:22 +0000 (13:00 +0000)]
Fix volatility marking for the generate_series and generate_subscripts families
of functions.
Tom Lane [Mon, 28 Apr 2008 23:48:16 +0000 (23:48 +0000)]
Increase the statement_timeout value used in the prepared_xacts regression
test. We have seen some buildfarm failures that seem to be due to this
limit being unexpectedly exceeded when the machine is under load.
Bruce Momjian [Mon, 28 Apr 2008 23:08:33 +0000 (23:08 +0000)]
Update comment for 8.4:
> #A hyphen, "-", marks changes that will appear in the upcoming 8.4 release.#
Tom Lane [Mon, 28 Apr 2008 22:47:03 +0000 (22:47 +0000)]
Fix broken configure test for libxslt: it was probing for xsltLibxmlVersion,
which is a global variable not a function, and so the probe failed on machines
where the linker makes a distinction (cf. Red Hat bug #444317). Probe for
an actual function instead.
Alvaro Herrera [Mon, 28 Apr 2008 14:57:35 +0000 (14:57 +0000)]
Bump the catversion for the previous patch.
Alvaro Herrera [Mon, 28 Apr 2008 14:48:58 +0000 (14:48 +0000)]
Add generate_subscripts, a series-generation function which generates an
array's subscripts.
Pavel Stehule, some editorialization by me.
Alvaro Herrera [Sun, 27 Apr 2008 04:33:27 +0000 (04:33 +0000)]
Add SGML ID attributes to the arrays subsections, and a few index entries for
them.
Tom Lane [Sat, 26 Apr 2008 22:47:40 +0000 (22:47 +0000)]
Code review for recent patch to terminate online backup during shutdown:
do CancelBackup at a sane place, fix some oversights in the state transitions,
allow only superusers to connect while we are waiting for backup mode to end.
Bruce Momjian [Fri, 25 Apr 2008 21:38:46 +0000 (21:38 +0000)]
Add URL for:
* Allow adding/renaming/removing enumerated values to an existing
enumerated data type
>
> http://archives.postgresql.org/pgsql-hackers/2008-04/msg01718.php
>
Bruce Momjian [Fri, 25 Apr 2008 14:34:26 +0000 (14:34 +0000)]
Update wording from David Fetter.
Bruce Momjian [Fri, 25 Apr 2008 14:28:51 +0000 (14:28 +0000)]
Update to remove passive wording from FAQ, David Fetter
Bruce Momjian [Thu, 24 Apr 2008 20:57:34 +0000 (20:57 +0000)]
Update TODO item:
> * Allow adding/renaming/removing enumerated values to an existing
> enumerated data type
Tom Lane [Thu, 24 Apr 2008 20:46:49 +0000 (20:46 +0000)]
Remove transformAlterTableStmt's kluge to replace ColumnDef.is_not_null
flags by separate AT_SetNotNull subcommands. That was always ugly and
inefficient, and it's now clear that it was merely a partial workaround
for the bug just identified in ATExecAddColumn. This is just code
beautification not a bug fix, so no back-patch.
Brendan Jurd, with some trivial additional cleanup by me.
Bruce Momjian [Thu, 24 Apr 2008 20:29:16 +0000 (20:29 +0000)]
Update:
< * Allow adding enumerated values to an existing enumerated data
> * Allow adding/removing enumerated values to an existing enumerated data
Tom Lane [Thu, 24 Apr 2008 20:17:50 +0000 (20:17 +0000)]
Fix ALTER TABLE ADD COLUMN ... PRIMARY KEY so that the new column is correctly
checked to see if it's been initialized to all non-nulls. The implicit NOT
NULL constraint was not being checked during the ALTER (in fact, not even if
there was an explicit NOT NULL too), because ATExecAddColumn neglected to
set the flag needed to make the test happen. This has been broken since
the capability was first added, in 8.0.
Brendan Jurd, per a report from Kaloyan Iliev.
Bruce Momjian [Thu, 24 Apr 2008 19:58:32 +0000 (19:58 +0000)]
Add embedded usage mention to FAQ, per Greg Smith.
Bruce Momjian [Thu, 24 Apr 2008 18:23:05 +0000 (18:23 +0000)]
Add to TODO:
> * Allow prepared transactions with temporary tables created and dropped
> in the same transaction, and when an ON COMMIT DELETE ROWS temporary
> table is accessed
>
> http://archives.postgresql.org/pgsql-hackers/2008-03/msg00047.php
Bruce Momjian [Thu, 24 Apr 2008 17:32:44 +0000 (17:32 +0000)]
Add TODO item:
>
> o Impove COPY performance
>
> http://archives.postgresql.org/pgsql-hackers/2008-02/msg00954.php
Bruce Momjian [Thu, 24 Apr 2008 17:06:13 +0000 (17:06 +0000)]
Add URL for:
* Allow index scans to return matching index keys, not just the matching
heap locations
>
> http://archives.postgresql.org/pgsql-hackers/2008-04/msg01657.php
>
Bruce Momjian [Thu, 24 Apr 2008 16:42:39 +0000 (16:42 +0000)]
Add to TODO:
> * Allow index scans to return matching index keys, not just the matching
> heap locations
Bruce Momjian [Thu, 24 Apr 2008 16:14:25 +0000 (16:14 +0000)]
Add to TODO:
<
> * Allow adding enumerated values to an existing enumerated data
> type
Bruce Momjian [Thu, 24 Apr 2008 16:11:31 +0000 (16:11 +0000)]
Remove, item done, description unclear:
< * Allow index scans to return matching index keys
<
< http://archives.postgresql.org/pgsql-hackers/2007-03/msg01079.php
Bruce Momjian [Thu, 24 Apr 2008 15:54:00 +0000 (15:54 +0000)]
Update TODO description for GIT:
* Consider smaller indexes that record a range of values per heap page,
rather than having one index entry for every heap row
Magnus Hagander [Thu, 24 Apr 2008 14:23:43 +0000 (14:23 +0000)]
Update error messages, per notes from Tom.
Laurenz Albe
Bruce Momjian [Thu, 24 Apr 2008 02:24:29 +0000 (02:24 +0000)]
Add GIT URL:
> http://archives.postgresql.org/pgsql-hackers/2008-04/msg01589.php
Bruce Momjian [Thu, 24 Apr 2008 00:39:50 +0000 (00:39 +0000)]
Add to TODO:
<
> o Add prompt escape to display the client and server versions
Bruce Momjian [Wed, 23 Apr 2008 22:38:47 +0000 (22:38 +0000)]
Add URL for MERGE:
> http://archives.postgresql.org/pgsql-hackers/2008-04/msg01475.php
Magnus Hagander [Wed, 23 Apr 2008 13:44:59 +0000 (13:44 +0000)]
Prevent shutdown in normal mode if online backup is running, and
have pg_ctl warn about this.
Cancel running online backups (by renaming the backup_label file,
thus rendering the backup useless) when shutting down in fast mode.
Laurenz Albe
Teodor Sigaev [Tue, 22 Apr 2008 17:52:43 +0000 (17:52 +0000)]
Fix using too many LWLocks bug, reported by Craig Ringer
<craig@postnewspapers.com.au>.
It was my mistake, I missed limitation of number of held locks, now GIN doesn't
use continiuous locks, but still hold buffers pinned to prevent interference
with vacuum's deletion algorithm.
Backpatch is needed.
Magnus Hagander [Tue, 22 Apr 2008 13:06:57 +0000 (13:06 +0000)]
Don't pull in pthreads header files in this file because we never use them.
Bruce Momjian [Tue, 22 Apr 2008 10:30:32 +0000 (10:30 +0000)]
FAQ2txt now only controls the main FAQ, not FAQ_DEV.
Magnus Hagander [Tue, 22 Apr 2008 09:26:32 +0000 (09:26 +0000)]
Replace developer FAQ with a reference to the wiki, which is where
it now lives (per discussion). Leave the other FAQs alone for now.
Tom Lane [Tue, 22 Apr 2008 01:34:34 +0000 (01:34 +0000)]
Issue explicit error messages for attempts to use "shell" operators in
ordinary expressions. This probably doesn't catch every single case
where you might get "cache lookup failed for function 0" for use of a
shell operator, but it will catch most. Per bug #4120 from Pedro Gimeno.
This patch incidentally folds make_op_expr() into its sole remaining
caller --- the alternative was to give it yet more arguments, which
didn't seem an improvement.
Tom Lane [Mon, 21 Apr 2008 20:54:15 +0000 (20:54 +0000)]
Fix convert_IN_to_join to properly handle the case where the subselect's
output is not of the same type that's needed for the IN comparison (ie,
where the parser inserted an implicit coercion above the subselect result).
We should record the coerced expression, not just a raw Var referencing
the subselect output, as the quantity that needs to be unique-ified if
we choose to implement the IN as Unique followed by a plain join.
As of 8.3 this error was causing crashes, as seen in bug #4113 from Javier
Hernandez, because the executor was being told to hash or sort the raw
subselect output column using operators appropriate to the coerced type.
In prior versions there was no crash because the executor chose the
hash or sort operators for itself based on the column type it saw.
However, that's still not really right, because what's unique for one data
type might not be unique for another. In corner cases we could get multiple
outputs of a row that should appear only once, as demonstrated by the
regression test case included in this commit.
However, this patch doesn't apply cleanly to 8.2 or before, and the code
involved has shifted enough over time that I'm hesitant to try to back-patch.
Given the lack of complaints from the field about such corner cases, I think
the bug may not be important enough to risk breaking other things with a
back-patch.
Magnus Hagander [Mon, 21 Apr 2008 18:37:28 +0000 (18:37 +0000)]
Clean up float4byval and float8byval handling by dealing with them completely
from inside the build script.
Magnus Hagander [Mon, 21 Apr 2008 11:06:13 +0000 (11:06 +0000)]
Fix typo, noted by Stefan Kaltenbrunner.
Magnus Hagander [Mon, 21 Apr 2008 10:01:32 +0000 (10:01 +0000)]
Make float4byval and float8byval configurable on MSVC.
Magnus Hagander [Mon, 21 Apr 2008 09:44:47 +0000 (09:44 +0000)]
Add link to major version release notes at the top of the minor
version ones, to make it clear to users just browsing the notes
that there are a lot more changes available from whatever version
they are at than what's in the minor version release notes.
Tom Lane [Mon, 21 Apr 2008 03:49:45 +0000 (03:49 +0000)]
Fix a couple of places in execMain that erroneously assumed that SELECT FOR
UPDATE/SHARE couldn't occur as a subquery in a query with a non-SELECT
top-level operation. Symptoms included outright failure (as in report from
Mark Mielke) and silently neglecting to take the requested row locks.
Back-patch to 8.3, because the visible failure in the INSERT ... SELECT case
is a regression from 8.2. I'm a bit hesitant to back-patch further given the
lack of field complaints.
Tom Lane [Mon, 21 Apr 2008 02:04:09 +0000 (02:04 +0000)]
Add FLOAT4PASSBYVAL/FLOAT8PASSBYVAL to pg_config.h.win32, as a stopgap
measure to get the Windows buildfarm members working again. I don't
know if it's worth exposing these as configurables, or exactly how to
do it in the MSVC build system ...
Tom Lane [Mon, 21 Apr 2008 01:11:43 +0000 (01:11 +0000)]
Make earthdistance use version-0 calling convention if not USE_FLOAT8_BYVAL,
and version-1 if USE_FLOAT8_BYVAL. This might seem a bit pointless, but the
idea is to have at least one regression test that will fail if we ever
accidentally break version-0 functions that return float8. However, they're
already broken, or at least hopelessly unportable, in the USE_FLOAT8_BYVAL
case.
Per a recent suggestion from Greg Stark.
Tom Lane [Mon, 21 Apr 2008 00:26:47 +0000 (00:26 +0000)]
Allow float8, int8, and related datatypes to be passed by value on machines
where Datum is 8 bytes wide. Since this will break old-style C functions
(those still using version 0 calling convention) that have arguments or
results of these types, provide a configure option to disable it and retain
the old pass-by-reference behavior. Likewise, provide a configure option
to disable the recently-committed float4 pass-by-value change.
Zoltan Boszormenyi, plus configurability stuff by me.
Teodor Sigaev [Sun, 20 Apr 2008 09:17:57 +0000 (09:17 +0000)]
Fix broken compare function for tsquery_ops. Per Tom's report.
I never understood why initial authors GiST in pgsql choose so
stgrange signature for 'same' method:
bool *sameFn(Datum a, Datum b, bool* result)
instead of simple, logical
bool sameFn(Datum a, Datum b)
This change will break any existing GiST extension, so we still live with
it and will live.
Tom Lane [Sun, 20 Apr 2008 01:05:52 +0000 (01:05 +0000)]
Convert earthdistance's only C function to v1 call convention,
to future-proof it against pass-by-value float8.
Tom Lane [Sat, 19 Apr 2008 22:55:03 +0000 (22:55 +0000)]
seg_size() has to be V1 calling convention, too.
Bruce Momjian [Sat, 19 Apr 2008 12:52:51 +0000 (12:52 +0000)]
Update doc script comments.
Bruce Momjian [Sat, 19 Apr 2008 12:42:35 +0000 (12:42 +0000)]
Add script FAQ2txt to convert HTML files to txt. Add comment to TODO2html.
Bruce Momjian [Sat, 19 Apr 2008 12:34:54 +0000 (12:34 +0000)]
Add TODO2html tool to convert TODO to HTML.
Bruce Momjian [Sat, 19 Apr 2008 12:33:06 +0000 (12:33 +0000)]
Update TODO using new script.
Bruce Momjian [Sat, 19 Apr 2008 12:19:41 +0000 (12:19 +0000)]
Remove TODO.detail directory. All URLs now in TODO file as references.
Bruce Momjian [Sat, 19 Apr 2008 12:16:21 +0000 (12:16 +0000)]
Remove TODO.detail references; instead add TODO URLs pointing to the
archives. We have been using URLs for a while for new items.
Alvaro Herrera [Fri, 18 Apr 2008 21:11:35 +0000 (21:11 +0000)]
Change the float4-returning functions in contrib/seg to fmgr v1 calling
conventions.
I also changed seg_in and seg_out, which was probably unnecessary, but
it can't harm.
Tom Lane [Fri, 18 Apr 2008 20:51:17 +0000 (20:51 +0000)]
Fix typo.
Alvaro Herrera [Fri, 18 Apr 2008 18:43:09 +0000 (18:43 +0000)]
Modify the float4 datatype to be pass-by-val. Along the way, remove the last
uses of the long-deprecated float32 in contrib/seg; the definitions themselves
are still there, but no longer used. fmgr/README updated to match.
I added a CREATE FUNCTION to account for existing seg_center() code in seg.c
too, and some tests for it and the neighbor functions. At the same time,
remove checks for NULL which are not needed (because the functions are declared
STRICT).
I had to do some adjustments to contrib's btree_gist too. The choices for
representation there are not ideal for changing the underlying types :-(
Original patch by Zoltan Boszormenyi, with some adjustments by me.
Tom Lane [Fri, 18 Apr 2008 17:05:45 +0000 (17:05 +0000)]
Fix rmtree() so that it keeps going after failure to remove any individual
file; the idea is that we should clean up as much as we can, even if there's
some problem removing one file. Make the error messages a bit less misleading,
too. In passing, const-ify function arguments.
Heikki Linnakangas [Fri, 18 Apr 2008 06:48:38 +0000 (06:48 +0000)]
Fix two race conditions between the pending unlink mechanism that was put in
place to prevent reusing relation OIDs before next checkpoint, and DROP
DATABASE. First, if a database was dropped, bgwriter would still try to unlink
the files that the rmtree() call by the DROP DATABASE command has already
deleted, or is just about to delete. Second, if a database is dropped, and
another database is created with the same OID, bgwriter would in the worst
case delete a relation in the new database that happened to get the same OID
as a dropped relation in the old database.
To fix these race conditions:
- make rmtree() ignore ENOENT errors. This fixes the 1st race condition.
- make ForgetDatabaseFsyncRequests forget unlink requests as well.
- force checkpoint on in dropdb on all platforms
Since ForgetDatabaseFsyncRequests() is asynchronous, the 2nd change isn't
enough on its own to fix the problem of dropping and creating a database with
same OID, but forcing a checkpoint on DROP DATABASE makes it sufficient.
Per Tom Lane's bug report and proposal. Backpatch to 8.3.
Tom Lane [Fri, 18 Apr 2008 01:42:17 +0000 (01:42 +0000)]
Cause EXPLAIN's VERBOSE option to print the target list (output column list)
of each plan node, instead of its former behavior of dumping the internal
representation of the plan tree. The latter display is still available for
those who really want it (see debug_print_plan), but uses for it are certainly
few and and far between. Per discussion.
This patch also removes the explain_pretty_print GUC, which is obsoleted
by the change.
Alvaro Herrera [Thu, 17 Apr 2008 21:37:28 +0000 (21:37 +0000)]
Clean up a few places where Datums were being treated as pointers (and vice
versa) without going through DatumGetPointer.
Gavin Sherry, with Feng Tian.
Tom Lane [Thu, 17 Apr 2008 21:22:14 +0000 (21:22 +0000)]
Fix a couple of oversights associated with the "physical tlist" optimization:
we had several code paths where a physical tlist could be used for the input
to a Sort node, which is a dumb idea because any unneeded table columns will
increase the volume of data the sort has to push around.
(Unfortunately the easy-looking fix of calling disuse_physical_tlist during
make_sort_xxx doesn't work because in most cases we're already committed to
the current input tlist --- it's been marked with sort column numbers, or
we've built grouping column numbers using it, etc. The tlist has to be
selected properly at the calling level before we start constructing sort-col
information. This is easy enough to do, we were just failing to take the
point into consideration.)
Back-patch to 8.3. I believe the problem probably exists clear back to 7.4
when the physical tlist optimization was added, but I'm afraid to back-patch
further than 8.3 without a great deal more study than I want to put into it.
The code in this area has drifted a lot over time. The real-world importance
of these code paths is uncertain anyway --- I think in many cases we'd
probably prefer hash-based methods.
Bruce Momjian [Thu, 17 Apr 2008 20:57:07 +0000 (20:57 +0000)]
TODO Done:
> * -Allow administrators to safely terminate individual sessions
Bruce Momjian [Thu, 17 Apr 2008 20:56:41 +0000 (20:56 +0000)]
Re-enable pg_terminate_backend() using SIGTERM. SIGTERM testing still
needed.
Tom Lane [Thu, 17 Apr 2008 18:30:18 +0000 (18:30 +0000)]
Add some code to EXPLAIN to show the targetlist (ie, output columns)
of each plan node. For the moment this is debug support only and is
not enabled unless EXPLAIN_PRINT_TLISTS is defined at build time.
Later I'll see about the idea of letting EXPLAIN VERBOSE do it.
Tom Lane [Wed, 16 Apr 2008 23:59:40 +0000 (23:59 +0000)]
Repair two places where SIGTERM exit could leave shared memory state
corrupted. (Neither is very important if SIGTERM is used to shut down the
whole database cluster together, but there's a problem if someone tries to
SIGTERM individual backends.) To do this, introduce new infrastructure
macros PG_ENSURE_ERROR_CLEANUP/PG_END_ENSURE_ERROR_CLEANUP that take care
of transiently pushing an on_shmem_exit cleanup hook. Also use this method
for createdb cleanup --- that wasn't a shared-memory-corruption problem,
but SIGTERM abort of createdb could leave orphaned files lying around.
Backpatch as far as 8.2. The shmem corruption cases don't exist in 8.1,
and the createdb usage doesn't seem important enough to risk backpatching
further.
Andrew Dunstan [Wed, 16 Apr 2008 22:16:00 +0000 (22:16 +0000)]
Fix MinGW warnings re formats and unused variables. per ITAGAKI Takahiro
Bruce Momjian [Wed, 16 Apr 2008 21:03:08 +0000 (21:03 +0000)]
Ignore blank lines in typedef file.
Bruce Momjian [Wed, 16 Apr 2008 20:57:18 +0000 (20:57 +0000)]
Update most recent release to 8.3.1.
Tom Lane [Wed, 16 Apr 2008 18:23:04 +0000 (18:23 +0000)]
Fix LOAD_CRIT_INDEX() macro to take out AccessShareLock on the system index
it is trying to build a relcache entry for. This is an oversight in my 8.2
patch that tried to ensure we always took a lock on a relation before trying
to build its relcache entry. The implication is that if someone committed a
reindex of a critical system index at about the same time that some other
backend were starting up without a valid pg_internal.init file, the second one
might PANIC due to not seeing any valid version of the index's pg_class row.
Improbable case, but definitely not impossible.
Bruce Momjian [Wed, 16 Apr 2008 18:18:00 +0000 (18:18 +0000)]
Fix comment typo.
Bryce Nesbitt