]> granicus.if.org Git - postgresql/log
postgresql
18 years agoFix bug I introduced while cleaning up pg_dump -t/-n patch.
Bruce Momjian [Tue, 1 Aug 2006 21:05:00 +0000 (21:05 +0000)]
Fix bug I introduced while cleaning up pg_dump -t/-n patch.

18 years agoEditorialize a bit on the recent patch, reverting to the original versions
Alvaro Herrera [Tue, 1 Aug 2006 19:17:18 +0000 (19:17 +0000)]
Editorialize a bit on the recent patch, reverting to the original versions
of the text where appropiate.

18 years agoAdd support for loadable modules to allocated shared memory and
Bruce Momjian [Tue, 1 Aug 2006 19:03:11 +0000 (19:03 +0000)]
Add support for loadable modules to allocated shared memory and
lightweight locks.

Marc Munro

18 years agoAdd pg_dump -X no-data-for-failed-tables option to suppress loading data
Bruce Momjian [Tue, 1 Aug 2006 18:21:44 +0000 (18:21 +0000)]
Add pg_dump -X no-data-for-failed-tables option to suppress loading data
if table creation failed (the table already exists).

Martin Pitt

18 years agoDone:
Bruce Momjian [Tue, 1 Aug 2006 18:07:36 +0000 (18:07 +0000)]
Done:

<  o %Allow pg_dump to use multiple -t and -n switches  [pg_dump]
<
<    http://archives.postgresql.org/pgsql-patches/2006-01/msg00190.php
<
>  o -Allow pg_dump to use multiple -t and -n switches, exclusion
>     ability, and regular expression object matching

18 years agoRemove TODO.detail/pg_dump.
Bruce Momjian [Tue, 1 Aug 2006 18:06:27 +0000 (18:06 +0000)]
Remove TODO.detail/pg_dump.

18 years agoAllow multiple -n (schema) and -t (table) pg_dump options, and add -T
Bruce Momjian [Tue, 1 Aug 2006 18:05:04 +0000 (18:05 +0000)]
Allow multiple -n (schema) and -t (table)  pg_dump options, and add -T
and -N options to exclude objects.  Also support regular expressions for
option object names.

Greg Sabino Mullane

18 years agoRemove extra argument to printf().
Bruce Momjian [Tue, 1 Aug 2006 18:01:36 +0000 (18:01 +0000)]
Remove extra argument to printf().

18 years agohave pg_regress fall back on testing with the canonical results file if an
Andrew Dunstan [Tue, 1 Aug 2006 14:56:29 +0000 (14:56 +0000)]
have pg_regress fall back on testing with the canonical results file if an
alternative test is specified but none succeeds.

18 years agoImprovements to Maintenance and External Projects documentation.
Bruce Momjian [Tue, 1 Aug 2006 00:09:06 +0000 (00:09 +0000)]
Improvements to Maintenance and External Projects documentation.

Joshua D. Drake
Robert Treat

18 years agoAdd:
Bruce Momjian [Mon, 31 Jul 2006 23:04:13 +0000 (23:04 +0000)]
Add:

> * Consider compressed annealing to search for query plans
>
>   This might replace GEQO, http://sixdemonbag.org/Djinni.

18 years agoChange the relation_open protocol so that we obtain lock on a relation
Tom Lane [Mon, 31 Jul 2006 20:09:10 +0000 (20:09 +0000)]
Change the relation_open protocol so that we obtain lock on a relation
(table or index) before trying to open its relcache entry.  This fixes
race conditions in which someone else commits a change to the relation's
catalog entries while we are in process of doing relcache load.  Problems
of that ilk have been reported sporadically for years, but it was not
really practical to fix until recently --- for instance, the recent
addition of WAL-log support for in-place updates helped.

Along the way, remove pg_am.amconcurrent: all AMs are now expected to support
concurrent update.

18 years agoAdded missing ECPGget_sqlca() call
Michael Meskes [Mon, 31 Jul 2006 13:26:46 +0000 (13:26 +0000)]
Added missing ECPGget_sqlca() call

18 years agoFirst small patches needed for regression tests
Michael Meskes [Mon, 31 Jul 2006 10:15:31 +0000 (10:15 +0000)]
First small patches needed for regression tests

18 years agoFix memory allocation bug in pgbench.
Tatsuo Ishii [Mon, 31 Jul 2006 05:15:14 +0000 (05:15 +0000)]
Fix memory allocation bug in pgbench.
patches submitted by ITAGAKI Takahiro.

18 years agoChange the bootstrap sequence so that toast tables for system catalogs are
Tom Lane [Mon, 31 Jul 2006 01:16:38 +0000 (01:16 +0000)]
Change the bootstrap sequence so that toast tables for system catalogs are
created in the bootstrap phase proper, rather than added after-the-fact
by initdb.  This is cleaner than before because it allows us to retire the
undocumented ALTER TABLE ... CREATE TOAST TABLE command, but the real reason
I'm doing it is so that toast tables of shared catalogs will now have
predetermined OIDs.  This will allow a reasonably clean solution to the
problem of locking tables before we load their relcache entries, to appear
in a forthcoming patch.

18 years agoFix confusion between COPY FROM and COPY TO, per Gavin Sharry and Arul Shaji.
Alvaro Herrera [Mon, 31 Jul 2006 01:10:56 +0000 (01:10 +0000)]
Fix confusion between COPY FROM and COPY TO, per Gavin Sharry and Arul Shaji.

18 years agoFix a couple of comments.
Tom Lane [Sun, 30 Jul 2006 20:17:11 +0000 (20:17 +0000)]
Fix a couple of comments.

18 years agoReplaced $Header$ by $§PosgreSQL$
Michael Meskes [Sun, 30 Jul 2006 16:28:58 +0000 (16:28 +0000)]
Replaced $Header$ by $§PosgreSQL$

18 years agoAdded missing $Header$ entries.
Michael Meskes [Sun, 30 Jul 2006 10:24:10 +0000 (10:24 +0000)]
Added missing $Header$ entries.

18 years agoModify snapshot definition so that lazy vacuums are ignored by other
Alvaro Herrera [Sun, 30 Jul 2006 02:07:18 +0000 (02:07 +0000)]
Modify snapshot definition so that lazy vacuums are ignored by other
vacuums.  This allows a OLTP-like system with big tables to continue
regular vacuuming on small-but-frequently-updated tables while the
big tables are being vacuumed.

Original patch from Hannu Krossing, rewritten by Tom Lane and updated
by me.

18 years agoAdd:
Bruce Momjian [Sun, 30 Jul 2006 01:48:40 +0000 (01:48 +0000)]
Add:

> * Consider detoasting keys before sorting

18 years agoFix WIN32 wait() return value macros to be accurate, particularly
Bruce Momjian [Sun, 30 Jul 2006 01:45:21 +0000 (01:45 +0000)]
Fix WIN32 wait() return value macros to be accurate, particularly
because they are used for testing the return value from system().
(WIN32 doesn't overlay the return code with other failure conditions
like Unix does, so they are just simple macros.)

Fix regression checks to properly handle diff failures on Win32 using
the new macros.

18 years agoprevent multiplexing Windows kernel event objects we listen for across various socket...
Andrew Dunstan [Sat, 29 Jul 2006 19:55:18 +0000 (19:55 +0000)]
prevent multiplexing Windows kernel event objects we listen for across various sockets - should fix the occasional stats test regression failures we see.

18 years agoAdd entry:
Bruce Momjian [Sat, 29 Jul 2006 18:08:36 +0000 (18:08 +0000)]
Add entry:

> * Add column to pg_stat_activity that shows the progress of long-running
>   commands like CREATE INDEX and VACUUM

18 years agoDon't use #include <crtdefs.h> for MSVC <= 1400.
Bruce Momjian [Sat, 29 Jul 2006 17:35:07 +0000 (17:35 +0000)]
Don't use #include <crtdefs.h> for MSVC <= 1400.

Hiroshi Saito

18 years agoAdd:
Bruce Momjian [Sat, 29 Jul 2006 15:56:32 +0000 (15:56 +0000)]
Add:

> * Add support for SQL-standard GENERATED/IDENTITY columns
>
>   http://archives.postgresql.org/pgsql-hackers/2006-07/msg00543.php

18 years agoAdd #include postgres_fe.h in two win32-specific files, because there is
Bruce Momjian [Sat, 29 Jul 2006 15:22:27 +0000 (15:22 +0000)]
Add #include postgres_fe.h in two win32-specific files, because there is
pg-specific code in those files, or might be some day.

18 years agoAdjust initialization sequence for timezone_abbreviations so that
Tom Lane [Sat, 29 Jul 2006 03:02:56 +0000 (03:02 +0000)]
Adjust initialization sequence for timezone_abbreviations so that
it's handled just about like timezone; in particular, don't try
to read anything during InitializeGUCOptions.  Should solve current
startup failure on Windows, and avoid wasted cycles if a nondefault
setting is specified in postgresql.conf too.  Possibly we need to
think about a more general solution for handling 'expensive to set'
GUC options.

18 years agoChange the delta val from 0 and 10000 to -5000 and 5000 per recent
Tatsuo Ishii [Fri, 28 Jul 2006 22:58:26 +0000 (22:58 +0000)]
Change the delta val from 0 and 10000 to -5000 and 5000 per recent
discussion in hackers list.  Also enhance predefined benchmark
scenarios to reflect the scaling factor parameter flexibly.

18 years agoRemove update TODO.detail item.
Bruce Momjian [Fri, 28 Jul 2006 19:28:08 +0000 (19:28 +0000)]
Remove update TODO.detail item.

18 years agoUpdates per Tom:
Bruce Momjian [Fri, 28 Jul 2006 19:00:45 +0000 (19:00 +0000)]
Updates per Tom:

> * Allow DISTINCT to work in multiple-argument aggregate calls
<  o Allow UPDATE to handle complex aggregates [update]?

18 years agoSQL2003-standard statistical aggregates, by Sergey Koposov. I've added only
Tom Lane [Fri, 28 Jul 2006 18:33:04 +0000 (18:33 +0000)]
SQL2003-standard statistical aggregates, by Sergey Koposov.  I've added only
the float8 versions of the aggregates, which is all that the standard requires.
Sergey's original patch also provided versions using numeric arithmetic,
but given the size and slowness of the code, I doubt we ought to include
those in core.

18 years agoFix table title.
Tom Lane [Fri, 28 Jul 2006 16:21:57 +0000 (16:21 +0000)]
Fix table title.

18 years agoMake it clearer that not every Postgres character set can be used as a
Tom Lane [Fri, 28 Jul 2006 15:33:17 +0000 (15:33 +0000)]
Make it clearer that not every Postgres character set can be used as a
server-side character set.

18 years agoUpdate Tatsuo:
Bruce Momjian [Fri, 28 Jul 2006 12:24:07 +0000 (12:24 +0000)]
Update Tatsuo:

< * Tatsuo is Tatsuo Ishii <t-ishii@sra.co.jp> of Software Research Assoc.
> * Tatsuo is Tatsuo Ishii <ishii@sraoss.co.jp> of SRA OSS, Inc. Japan

18 years agoVariables should be free'ed only once.
Michael Meskes [Fri, 28 Jul 2006 11:49:36 +0000 (11:49 +0000)]
Variables should be free'ed only once.

18 years agoRemoved superfluous comment
Michael Meskes [Fri, 28 Jul 2006 10:12:56 +0000 (10:12 +0000)]
Removed superfluous comment

18 years agoAnd one more:
Michael Meskes [Fri, 28 Jul 2006 10:10:42 +0000 (10:10 +0000)]
And one more:
Connection identifier has to be unique

18 years agoAdded more SoC changes by Joachim Wieland <joe@mcknight.de>:
Michael Meskes [Fri, 28 Jul 2006 09:08:01 +0000 (09:08 +0000)]
Added more SoC changes by Joachim Wieland <joe@mcknight.de>:
- SHOW statement puts result into a variable
- COPY TO STDOUT works

18 years agoAdd URL for UPDATE/SET for multiple values.
Bruce Momjian [Thu, 27 Jul 2006 22:05:49 +0000 (22:05 +0000)]
Add URL for UPDATE/SET for multiple values.

18 years agoAdd for UPDATE SET for multiple values:
Bruce Momjian [Thu, 27 Jul 2006 22:01:34 +0000 (22:01 +0000)]
Add for UPDATE SET for multiple values:

>
>           A subselect can also be used as the value source.

18 years agoFix incorrect function return code in cube_cmp and cube_dim;
Tom Lane [Thu, 27 Jul 2006 21:55:09 +0000 (21:55 +0000)]
Fix incorrect function return code in cube_cmp and cube_dim;
update regression expected files to what I think is correct.

18 years agorevert last change - broke non-AIX platforms.
Andrew Dunstan [Thu, 27 Jul 2006 19:58:39 +0000 (19:58 +0000)]
revert last change - broke non-AIX platforms.

18 years agoAggregate functions now support multiple input arguments. I also took
Tom Lane [Thu, 27 Jul 2006 19:52:07 +0000 (19:52 +0000)]
Aggregate functions now support multiple input arguments.  I also took
the opportunity to treat COUNT(*) as a zero-argument aggregate instead
of the old hack that equated it to COUNT(1); this is materially cleaner
(no more weird ANYOID cases) and ought to be at least a tiny bit faster.
Original patch by Sergey Koposov; review, documentation, simple regression
tests, pg_dump and psql support by moi.

18 years agoanother try at keeping AIX/ppc happy on cube test.
Andrew Dunstan [Thu, 27 Jul 2006 18:38:54 +0000 (18:38 +0000)]
another try at keeping AIX/ppc happy on cube test.

18 years agofix most regression tests for new cube code.
Andrew Dunstan [Thu, 27 Jul 2006 16:12:36 +0000 (16:12 +0000)]
fix most regression tests for new cube code.

18 years agoOriginal coding of pg_regress.c made the results and log directories
Tom Lane [Thu, 27 Jul 2006 15:37:19 +0000 (15:37 +0000)]
Original coding of pg_regress.c made the results and log directories
with restrictive permissions, which was not the behavior of the shell
script and doesn't seem very desirable.  Use the umask setting instead.

18 years agoAllow LDAP lookups from pg_service.conf.
Bruce Momjian [Thu, 27 Jul 2006 13:20:24 +0000 (13:20 +0000)]
Allow LDAP lookups from pg_service.conf.

Albe Laurenz

18 years agoAllow units to be specified with configuration settings.
Peter Eisentraut [Thu, 27 Jul 2006 08:30:41 +0000 (08:30 +0000)]
Allow units to be specified with configuration settings.

18 years agoCode review for bigint-LIMIT patch. Fix missed planner dependency,
Tom Lane [Wed, 26 Jul 2006 19:31:51 +0000 (19:31 +0000)]
Code review for bigint-LIMIT patch.  Fix missed planner dependency,
eliminate unnecessary code, force initdb because stored rules change
(limit nodes are now supposed to be int8 not int4 expressions).
Update comments and error messages, which still all said 'integer'.

18 years agoReorder items:
Bruce Momjian [Wed, 26 Jul 2006 17:23:02 +0000 (17:23 +0000)]
Reorder items:

<    o Allow point-in-time recovery to archive partially filled
<      write-ahead logs? [pitr]
>    o Add command to archive partially filled write-ahead logs? [pitr]
<      of a disk failure. This could be triggered by a user command or
<      a timer.
>      of a disk failure.
<      recovery.  A function call to do this would also be useful.
>      recovery.
>    o Add reporting of the current WAL file and offset, perhaps as
>      part of partial log file archiving
>
>      The offset allows parts of a WAL file to be archived using
>      an external program.
>
<    o Add reporting of the current WAL file and offset, perhaps as
<      part of partial log file archiving
<
<      The offset allows parts of a WAL file to be archived using
<      an external program.

18 years agoWork around bug in strxfmt() but in MS VS2005.
Bruce Momjian [Wed, 26 Jul 2006 17:17:28 +0000 (17:17 +0000)]
Work around bug in strxfmt() but in MS VS2005.

William ZHANG

18 years agoWhen a GUC string variable is not set, print the empty string (in SHOW etc.),
Peter Eisentraut [Wed, 26 Jul 2006 11:39:47 +0000 (11:39 +0000)]
When a GUC string variable is not set, print the empty string (in SHOW etc.),
not "unset".  An "unset" state doesn't really exist; all variables behave
like an empty string value if the string being pointed to has not been
initialized.

18 years agoConvert effective_cache_size to an integer, for better integration with
Peter Eisentraut [Wed, 26 Jul 2006 11:35:56 +0000 (11:35 +0000)]
Convert effective_cache_size to an integer, for better integration with
upcoming units feature.

18 years agoNew features contributed by Tomoaki Sato.
Tatsuo Ishii [Wed, 26 Jul 2006 07:24:50 +0000 (07:24 +0000)]
New features contributed by Tomoaki Sato.

- predefined variable "tps"
  The value of variable tps is taken from the scaling factor
  specified by -s option.

- -D option
  Variable values can be defined by -D option.

- \set command now allows arithmetic calculations.

18 years agoDone:
Bruce Momjian [Wed, 26 Jul 2006 00:35:01 +0000 (00:35 +0000)]
Done:

> * -Change LIMIT/OFFSET and FETCH/MOVE to use int8

18 years agoChange LIMIT/OFFSET to use int8
Bruce Momjian [Wed, 26 Jul 2006 00:34:48 +0000 (00:34 +0000)]
Change LIMIT/OFFSET to use int8

Dhanaraj M

18 years ago/contrib/cube improvements:
Bruce Momjian [Tue, 25 Jul 2006 23:23:45 +0000 (23:23 +0000)]
/contrib/cube improvements:

Update the calling convention for all external facing functions. By
external facing, I mean all functions that are directly referenced in
cube.sql. Prior to my update, all functions used the older V0 calling
convention. They now use V1.

New Functions:

cube(float[]), which makes a zero volume cube from a float array

cube(float[], float[]), which allows the user to create a cube from
two float arrays; one for the upper right and one for the lower left
coordinate.

cube_subset(cube, int4[]), to allow you to reorder or choose a subset of
dimensions from a cube, using index values specified in the array.

Joshua Reich

18 years agoModify btree to delete known-dead index entries without an actual VACUUM.
Tom Lane [Tue, 25 Jul 2006 19:13:00 +0000 (19:13 +0000)]
Modify btree to delete known-dead index entries without an actual VACUUM.
When we are about to split an index page to do an insertion, first look
to see if any entries marked LP_DELETE exist on the page, and if so remove
them to try to make enough space for the desired insert.  This should reduce
index bloat in heavily-updated tables, although of course you still need
VACUUM eventually to clean up the heap.

Junji Teramoto

18 years agoAdd reason for WAL offset reporting:
Bruce Momjian [Tue, 25 Jul 2006 16:23:42 +0000 (16:23 +0000)]
Add reason for WAL offset reporting:

>
>      The offset allows parts of a WAL file to be archived using
>      an external program.

18 years agoAdd offset mention:
Bruce Momjian [Tue, 25 Jul 2006 15:27:24 +0000 (15:27 +0000)]
Add offset mention:

<    o Add reporting of the current WAL file, perhaps as part of
<      partial log file archiving
>    o Add reporting of the current WAL file and offset, perhaps as
>      part of partial log file archiving

18 years agoUpdate PITR:
Bruce Momjian [Tue, 25 Jul 2006 15:21:48 +0000 (15:21 +0000)]
Update PITR:

<      write-ahead logs [pitr]
>      write-ahead logs? [pitr]
<      recovery.
>      recovery.  A function call to do this would also be useful.

18 years agoAdd missing @OVERRIDE to 'India' config file, per Michael Fuhr.
Tom Lane [Tue, 25 Jul 2006 13:49:21 +0000 (13:49 +0000)]
Add missing @OVERRIDE to 'India' config file, per Michael Fuhr.
Add $PostgreSQL$ lines for CVS identity, too.

18 years agoDone:
Bruce Momjian [Tue, 25 Jul 2006 04:11:24 +0000 (04:11 +0000)]
Done:

>  o -Allow customization of the known set of TZ names (generalize the

18 years agoRemove hard-wired lists of timezone abbreviations in favor of providing
Tom Lane [Tue, 25 Jul 2006 03:51:23 +0000 (03:51 +0000)]
Remove hard-wired lists of timezone abbreviations in favor of providing
configuration files that can be altered by a DBA.  The australian_timezones
GUC setting disappears, replaced by a timezone_abbreviations setting (set this
to 'Australia' to get the effect of australian_timezones).  The list of zone
names defined by default has undergone a bit of cleanup, too.  Documentation
still needs some work --- in particular, should we fix Table B-4, or just get
rid of it?  Joachim Wieland, with some editorializing by moi.

18 years agoUse correct ifdef test for cygwin, namely __CYGWIN__ (note underscores).
Andrew Dunstan [Tue, 25 Jul 2006 01:37:42 +0000 (01:37 +0000)]
Use correct ifdef test for cygwin, namely __CYGWIN__ (note underscores).

18 years agoMake the order of operations in PostmasterMain a bit saner ... some
Tom Lane [Tue, 25 Jul 2006 01:23:34 +0000 (01:23 +0000)]
Make the order of operations in PostmasterMain a bit saner ... some
recent patches had added stuff in rather random spots.

18 years agoFix spectacular misspellings of procedural language names
Peter Eisentraut [Mon, 24 Jul 2006 16:55:59 +0000 (16:55 +0000)]
Fix spectacular misspellings of procedural language names

18 years agoDTrace support, with a small initial set of probes
Peter Eisentraut [Mon, 24 Jul 2006 16:32:45 +0000 (16:32 +0000)]
DTrace support, with a small initial set of probes

by Robert Lor

18 years agoMark postgresql.conf entries that require server restart; some minor
Peter Eisentraut [Mon, 24 Jul 2006 10:44:40 +0000 (10:44 +0000)]
Mark postgresql.conf entries that require server restart; some minor
editing and reformatting.

18 years agoRemove dubious and redundant (we think) setting of libdir in PATH for non temp-instal...
Andrew Dunstan [Mon, 24 Jul 2006 01:50:22 +0000 (01:50 +0000)]
Remove dubious and redundant (we think) setting of libdir in PATH for non temp-install case.

18 years agoAdd a fudge factor to genericcostestimate() to prevent the planner from
Tom Lane [Mon, 24 Jul 2006 01:19:48 +0000 (01:19 +0000)]
Add a fudge factor to genericcostestimate() to prevent the planner from
thinking that indexes of different sizes are equally attractive.  Per
gripe from Jim Nasby.  (I remain unconvinced that there's such a problem
in existing releases, but CVS HEAD definitely has got a problem because
of its new count-only-leaf-pages approach to indexscan costing.)

18 years agoConvert the lock manager to use the new dynahash.c support for partitioned
Tom Lane [Sun, 23 Jul 2006 23:08:46 +0000 (23:08 +0000)]
Convert the lock manager to use the new dynahash.c support for partitioned
hash tables, instead of the previous kluge involving multiple hash tables.
This partially undoes my patch of last December.

18 years agoFix oversight in sizing of shared buffer lookup hashtable. Because
Tom Lane [Sun, 23 Jul 2006 18:34:45 +0000 (18:34 +0000)]
Fix oversight in sizing of shared buffer lookup hashtable.  Because
BufferAlloc tries to insert a new mapping entry before deleting the old one
for a buffer, we have a transient need for more than NBuffers entries ---
one more in 8.1, and as many as NUM_BUFFER_PARTITIONS more in CVS HEAD.
In theory this could lead to an "out of shared memory" failure if shmem
had already been completely claimed by the time the extra entries were
needed.

18 years agoSplit the buffer mapping table into multiple separately lockable
Tom Lane [Sun, 23 Jul 2006 03:07:58 +0000 (03:07 +0000)]
Split the buffer mapping table into multiple separately lockable
partitions, as per discussion.  Passes functionality checks, but
I don't have any performance data yet.

18 years agoAdd support to dynahash.c for partitioning shared hashtables according
Tom Lane [Sat, 22 Jul 2006 23:04:39 +0000 (23:04 +0000)]
Add support to dynahash.c for partitioning shared hashtables according
to the low-order bits of the entry hash value.  Also make some incidental
cleanups in the dynahash API, such as not exporting the hash header
structs to the world.

18 years agoHmm, seems --disable-spinlocks has been broken for awhile and nobody
Tom Lane [Sat, 22 Jul 2006 21:04:40 +0000 (21:04 +0000)]
Hmm, seems --disable-spinlocks has been broken for awhile and nobody
noticed.  Fix SpinlockSemas() to report the correct count considering
that PG 8.1 adds a spinlock to each shared-buffer header.

18 years agoIn the recent changes to make the planner account better for cache
Tom Lane [Sat, 22 Jul 2006 15:41:56 +0000 (15:41 +0000)]
In the recent changes to make the planner account better for cache
effects in a nestloop inner indexscan, I had only dealt with plain index
scans and the index portion of bitmap scans.  But there will be cache
benefits for the heap accesses of bitmap scans too, so fix
cost_bitmap_heap_scan() to account for that.

18 years agoAdd libdir to PATH for Cygwin as well as WIN32 - should fix buildfarm eel.
Andrew Dunstan [Sat, 22 Jul 2006 14:05:20 +0000 (14:05 +0000)]
Add libdir to PATH for Cygwin as well as WIN32 - should fix buildfarm eel.

18 years agoFor automatic dependency tracking, replace our sed hackery with the GCC-
Peter Eisentraut [Fri, 21 Jul 2006 22:37:37 +0000 (22:37 +0000)]
For automatic dependency tracking, replace our sed hackery with the GCC-
-built-in mechanism through the -MP flag.  Adjust the file extensions to
look more like Automake practice.  This frees up the .d suffix for use by
DTrace.

18 years agoAdd the full set of comparison functions for type TID, including a btree
Tom Lane [Fri, 21 Jul 2006 20:51:33 +0000 (20:51 +0000)]
Add the full set of comparison functions for type TID, including a btree
opclass.  This is not so much because anyone's likely to create an index
on TID, as that sorting TIDs can be useful.  Also added max and min
aggregates while at it, so that one can investigate the clusteredness of
a table with queries like SELECT min(ctid), max(ctid) FROM tab WHERE ...
Greg Stark and Tom Lane

18 years agoAh, I finally realize why Magnus wanted to add a --bindir option to
Tom Lane [Fri, 21 Jul 2006 00:24:04 +0000 (00:24 +0000)]
Ah, I finally realize why Magnus wanted to add a --bindir option to
pg_regress: there's no other way to cope with testing a relocated
installation.  Seems better to call it --psqldir though, since the
only thing we need to find in that case is psql.  It'd be better if
we could use find_other_exec, but that's not happening unless we are
willing to install pg_regress alongside psql, which seems unlikely
to happen.

18 years agoUpdate documentation on GRANT SCEMA USAGE, again.
Bruce Momjian [Thu, 20 Jul 2006 18:00:03 +0000 (18:00 +0000)]
Update documentation on GRANT SCEMA USAGE, again.

18 years agoAs a stopgap to get the Windows buildfarm members running again, hot-wire
Tom Lane [Thu, 20 Jul 2006 16:25:30 +0000 (16:25 +0000)]
As a stopgap to get the Windows buildfarm members running again, hot-wire
the check on diff's exit status to check for literally 0 or 1.  Someone
should look into why WIFEXITED/WEXITSTATUS don't work for this, but I've
spent more than enough time on it already.

18 years agoMake pg_config location overridable for pgxs builds.
Peter Eisentraut [Thu, 20 Jul 2006 09:30:18 +0000 (09:30 +0000)]
Make pg_config location overridable for pgxs builds.

18 years agoPrint out diff status code when we think there's a hard failure.
Tom Lane [Thu, 20 Jul 2006 03:30:58 +0000 (03:30 +0000)]
Print out diff status code when we think there's a hard failure.
May help in debugging behavior on Windows.

18 years agoSuppress unused-variable compiler warning, per Andrew Dunstan.
Tom Lane [Thu, 20 Jul 2006 02:15:17 +0000 (02:15 +0000)]
Suppress unused-variable compiler warning, per Andrew Dunstan.

18 years agoFix pg_regress.c to report tests in a parallel group when they finish,
Tom Lane [Thu, 20 Jul 2006 02:10:00 +0000 (02:10 +0000)]
Fix pg_regress.c to report tests in a parallel group when they finish,
not when they're started.  This mimics a subtle point of the behavior
of the old shell script, and gives better feedback when watching the
tests.

18 years agoMake pg_regress.c get paths from pg_config_paths.h, instead of -D
Tom Lane [Thu, 20 Jul 2006 01:16:57 +0000 (01:16 +0000)]
Make pg_regress.c get paths from pg_config_paths.h, instead of -D
switches passed from the Makefile.  This looks like it will fix
problem with virtual vs real paths under msys.

18 years agoDon't try to truncate multixact SLRU files in checkpoints done during xlog
Tom Lane [Thu, 20 Jul 2006 00:46:42 +0000 (00:46 +0000)]
Don't try to truncate multixact SLRU files in checkpoints done during xlog
recovery.  In the first place, it doesn't work because slru's
latest_page_number isn't set up yet (this is why we've been hearing reports
of strange "apparent wraparound" log messages during crash recovery, but
only from people who'd managed to advance their next-mxact counters some
considerable distance from 0).  In the second place, it seems a bit unwise
to be throwing away data during crash recovery anwyway.  This latter
consideration convinces me to just disable truncation during recovery,
rather than computing latest_page_number and pushing ahead.

18 years agoUpdate GRANT USAGE on schema permission description.
Bruce Momjian [Wed, 19 Jul 2006 18:42:31 +0000 (18:42 +0000)]
Update GRANT USAGE on schema permission description.

18 years agopgcrypto merge cleanup:
Neil Conway [Wed, 19 Jul 2006 17:05:50 +0000 (17:05 +0000)]
pgcrypto merge cleanup:

  - Few README fixes
  - Keep imath Id string, put $PostgreSQL$ separately.

Patch from Marko Kreen.

18 years agoAdjust spawn_process() to avoid unnecessary overhead processes: we can
Tom Lane [Wed, 19 Jul 2006 17:02:59 +0000 (17:02 +0000)]
Adjust spawn_process() to avoid unnecessary overhead processes: we can
just exec instead of creating a subprocess.  This reduces process usage
from four processes per parallel test to two.  I have no idea whether
a comparable optimization is possible or useful in the Windows port.

18 years agoUpdate WIN32 quoting roles documentation.
Bruce Momjian [Wed, 19 Jul 2006 17:01:24 +0000 (17:01 +0000)]
Update WIN32 quoting roles documentation.

18 years agoAdjust pg_regress to print out the exact string given to system() when
Tom Lane [Wed, 19 Jul 2006 16:23:17 +0000 (16:23 +0000)]
Adjust pg_regress to print out the exact string given to system() when
'make install' or 'initdb' fails.  Also minor simplification of fgets()
usage --- fgets guarantees a trailing null anyway.

18 years agoAdd WIN32 compile suggestion to pginclude README.
Bruce Momjian [Wed, 19 Jul 2006 15:06:06 +0000 (15:06 +0000)]
Add WIN32 compile suggestion to pginclude README.

18 years agokill() is declared in <signal.h> per Single Unix Spec.
Tom Lane [Wed, 19 Jul 2006 05:21:57 +0000 (05:21 +0000)]
kill() is declared in <signal.h> per Single Unix Spec.

18 years agoTweak command quoting for Windows (I'd forgotten about SYSTEMQUOTE).
Tom Lane [Wed, 19 Jul 2006 04:50:57 +0000 (04:50 +0000)]
Tweak command quoting for Windows (I'd forgotten about SYSTEMQUOTE).

18 years agoRemove unnecessary inclusion of libpq into pg_regress --- overly
Tom Lane [Wed, 19 Jul 2006 04:02:31 +0000 (04:02 +0000)]
Remove unnecessary inclusion of libpq into pg_regress --- overly
enthusiastic copy and paste ...