OSDN Git Service

pg-rex/syncrep.git
19 years agoAdd:
Bruce Momjian [Sat, 30 Jul 2005 02:52:30 +0000 (02:52 +0000)]
Add:

> * Add PQescapeIdentifier() to libpq

19 years agoFix an oversight I introduced on 2003-12-28: find_nots/push_nots should
Tom Lane [Fri, 29 Jul 2005 21:40:02 +0000 (21:40 +0000)]
Fix an oversight I introduced on 2003-12-28: find_nots/push_nots should
continue to recurse after eliminating a NOT-below-a-NOT, since the
contained subexpression will now be part of the top-level AND/OR structure
and so deserves to be simplified.  The real-world impact of this is
probably minimal, since it'd require at least three levels of NOT to make
a difference, but it's still a bug.
Also remove some redundant tests for NULL subexpressions.

19 years agoRemove contrib version of pg_autovacuum --- superseded by integrated
Tom Lane [Fri, 29 Jul 2005 19:38:22 +0000 (19:38 +0000)]
Remove contrib version of pg_autovacuum --- superseded by integrated
version.

19 years agoClean up a number of autovacuum loose ends. Make the stats collector
Tom Lane [Fri, 29 Jul 2005 19:30:09 +0000 (19:30 +0000)]
Clean up a number of autovacuum loose ends.  Make the stats collector
track shared relations in a separate hashtable, so that operations done
from different databases are counted correctly.  Add proper support for
anti-XID-wraparound vacuuming, even in databases that are never connected
to and so have no stats entries.  Miscellaneous other bug fixes.
Alvaro Herrera, some additional fixes by Tom Lane.

19 years agoIt appears we need -DSUNOS4_CC for both solaris and sunos4 templates.
Tom Lane [Fri, 29 Jul 2005 17:00:29 +0000 (17:00 +0000)]
It appears we need -DSUNOS4_CC for both solaris and sunos4 templates.
Per report from Andrew Clark.

19 years agoAdd detail:
Bruce Momjian [Fri, 29 Jul 2005 16:05:28 +0000 (16:05 +0000)]
Add detail:
>
>     If CRC check fails during recovery, remember the page in case
>     a later CRC for that page properly matches.
>

19 years agoMove reindexdb from /contrib to /bin.
Bruce Momjian [Fri, 29 Jul 2005 15:13:11 +0000 (15:13 +0000)]
Move reindexdb from /contrib to /bin.

Euler Taveira de Oliveira

19 years agoUpdate catversion for dbsize changes.
Bruce Momjian [Fri, 29 Jul 2005 15:04:22 +0000 (15:04 +0000)]
Update catversion for dbsize changes.

19 years agoRemove dbsize mention.
Bruce Momjian [Fri, 29 Jul 2005 15:03:51 +0000 (15:03 +0000)]
Remove dbsize mention.

19 years agoMove dbsize functions into the backend. New functions:
Bruce Momjian [Fri, 29 Jul 2005 14:47:04 +0000 (14:47 +0000)]
Move dbsize functions into the backend.  New functions:

pg_tablespace_size
pg_database_size
pg_relation_size
pg_complete_relation_size
pg_size_pretty

Remove /contrib/dbsize.

Dave Page

19 years agoAdd mp3 entry to FAQ.
Bruce Momjian [Fri, 29 Jul 2005 14:05:12 +0000 (14:05 +0000)]
Add mp3 entry to FAQ.

19 years agoI've had this small patch in my local tree for a while. It documents
Bruce Momjian [Fri, 29 Jul 2005 13:00:03 +0000 (13:00 +0000)]
I've had this small patch in my local tree for a while.  It documents
new commands which may make an SPI call fail.

Alvaro Herrera

19 years agoFix typo.
Bruce Momjian [Fri, 29 Jul 2005 12:59:15 +0000 (12:59 +0000)]
Fix typo.

uniware

19 years agoUpdate O_DIRECT comment.
Bruce Momjian [Fri, 29 Jul 2005 03:25:53 +0000 (03:25 +0000)]
Update O_DIRECT comment.

19 years agoDone:
Bruce Momjian [Fri, 29 Jul 2005 03:23:00 +0000 (03:23 +0000)]
Done:

< * Consider use of open/fcntl(O_DIRECT) to minimize OS caching,
<   especially for WAL writes
> * -Consider use of open/fcntl(O_DIRECT) to minimize OS caching,
>   for WAL writes

19 years agoUse O_DIRECT if available when using O_SYNC for wal_sync_method.
Bruce Momjian [Fri, 29 Jul 2005 03:22:33 +0000 (03:22 +0000)]
Use O_DIRECT if available when using O_SYNC for wal_sync_method.

Also, write multiple WAL buffers out in one write() operation.

ITAGAKI Takahiro

---------------------------------------------------------------------------

> If we disable writeback-cache and use open_sync, the per-page writing
> behavior in WAL module will show up as bad result. O_DIRECT is similar
> to O_DSYNC (at least on linux), so that the benefit of it will disappear
> behind the slow disk revolution.
>
> In the current source, WAL is written as:
>     for (i = 0; i < N; i++) { write(&buffers[i], BLCKSZ); }
> Is this intentional? Can we rewrite it as follows?
>    write(&buffers[0], N * BLCKSZ);
>
> In order to achieve it, I wrote a 'gather-write' patch (xlog.gw.diff).
> Aside from this, I'll also send the fixed direct io patch (xlog.dio.diff).
> These two patches are independent, so they can be applied either or both.
>
>
> I tested them on my machine and the results as follows. It shows that
> direct-io and gather-write is the best choice when writeback-cache is off.
> Are these two patches worth trying if they are used together?
>
>
>             | writeback | fsync= | fdata | open_ | fsync_ | open_
> patch       | cache     |  false |  sync |  sync | direct | direct
> ------------+-----------+--------+-------+-------+--------+---------
> direct io   | off       |  124.2 | 105.7 |  48.3 |   48.3 |  48.2
> direct io   | on        |  129.1 | 112.3 | 114.1 |  142.9 | 144.5
> gather-write| off       |  124.3 | 108.7 | 105.4 |  (N/A) | (N/A)
> both        | off       |  131.5 | 115.5 | 114.4 |  145.4 | 145.2
>
> - 20runs * pgbench -s 100 -c 50 -t 200
>    - with tuning (wal_buffers=64, commit_delay=500, checkpoint_segments=8)
> - using 2 ATA disks:
>    - hda(reiserfs) includes system and wal.
>    - hdc(jfs) includes database files. writeback-cache is always on.
>
> ---
> ITAGAKI Takahiro

19 years agoThank you for applying patch --- regexp_replace.
Bruce Momjian [Fri, 29 Jul 2005 03:17:55 +0000 (03:17 +0000)]
Thank you for applying patch --- regexp_replace.

An attached patch is a small additional improvement.

This patch use appendStringInfoText instead of appendStringInfoString.
There is an overhead of PG_TEXT_GET_STR when appendStringInfoString is
executed by text type. This can be reduced by appendStringInfoText.

Atsushi Ogawa

19 years agoMake use of new list primitives list_append_unique and list_concat_unique
Tom Lane [Thu, 28 Jul 2005 22:27:02 +0000 (22:27 +0000)]
Make use of new list primitives list_append_unique and list_concat_unique
where applicable.

19 years agoAdd tab completion syntax for PREPARE 2-phase commit syntax.
Bruce Momjian [Thu, 28 Jul 2005 22:14:30 +0000 (22:14 +0000)]
Add tab completion syntax for PREPARE 2-phase commit syntax.

There's no tab-completion for "PREPARE TRANSACTION 'xxx'" since that
would be ambigous with "PREPARE xxx AS xx".

Heikki Linnakangas

19 years agoFix a bunch of bad interactions between partial indexes and the new
Tom Lane [Thu, 28 Jul 2005 20:26:22 +0000 (20:26 +0000)]
Fix a bunch of bad interactions between partial indexes and the new
planning logic for bitmap indexscans.  Partial indexes create corner
cases in which a scan might be done with no explicit index qual conditions,
and the code wasn't handling those cases nicely.  Also be a little
tenser about eliminating redundant clauses in the generated plan.
Per report from Dmitry Karasik.

19 years agoCode cleanup.
Bruce Momjian [Thu, 28 Jul 2005 15:30:55 +0000 (15:30 +0000)]
Code cleanup.

19 years agoRefactor exec_cast_value() and exec_simple_cast_value(): since they do
Neil Conway [Thu, 28 Jul 2005 07:51:13 +0000 (07:51 +0000)]
Refactor exec_cast_value() and exec_simple_cast_value(): since they do
not ever write through the `isnull' parameter, it does not need to be
an out parameter. Therefore it can be declared a "bool" rather than a
"bool *".

19 years agoMark a static array "const" to move a few bytes from the "data" segment
Neil Conway [Thu, 28 Jul 2005 07:38:33 +0000 (07:38 +0000)]
Mark a static array "const" to move a few bytes from the "data" segment
to the "text" segment. It would be possible to mark the elements of the
array "const" as well, but this would require multiple API changes and
does not seem to be worth the notational inconvenience.

19 years agoRemove no-longer-needed hack to add pgsleep.c to regress.so.
Tom Lane [Thu, 28 Jul 2005 04:32:32 +0000 (04:32 +0000)]
Remove no-longer-needed hack to add pgsleep.c to regress.so.

19 years agoPut libpgport into OBJS instead of LIBS, so that it gets included
Tom Lane [Thu, 28 Jul 2005 04:31:30 +0000 (04:31 +0000)]
Put libpgport into OBJS instead of LIBS, so that it gets included
into .def and .exp files automatically on Windows, AIX, and the like.
An additional benefit is that changes in libpgport files correctly
propagate to force rebuild of the backend executable.  This is my
reworking of Rocco Altier's idea, and if it breaks anything it's
definitely my fault.

19 years agoFix a whole bunch of #includes that were either wrong or redundant.
Tom Lane [Thu, 28 Jul 2005 04:03:14 +0000 (04:03 +0000)]
Fix a whole bunch of #includes that were either wrong or redundant.
The first rule of portability for us is 'thou shalt have no other gods
before c.h', and a whole lot of these files were either not including
c.h at all, or including random system headers beforehand, either of
which sins can mess up largefile support nicely.  Once you have
included c.h, there is no need to re-include what it includes, either.

19 years agoRemove some long-dead support for libpgtcl with Tk.
Tom Lane [Thu, 28 Jul 2005 03:15:52 +0000 (03:15 +0000)]
Remove some long-dead support for libpgtcl with Tk.

19 years agoFix breakage introduced in plpgsql CONTINUE patch.
Tom Lane [Thu, 28 Jul 2005 00:26:30 +0000 (00:26 +0000)]
Fix breakage introduced in plpgsql CONTINUE patch.
Per Kevin McArthur.

19 years agoAdd some missing SGML markup.
Neil Conway [Wed, 27 Jul 2005 13:42:01 +0000 (13:42 +0000)]
Add some missing SGML markup.

19 years agoFix a few macro definitions to ensure that unary minus is enclosed in
Neil Conway [Wed, 27 Jul 2005 12:44:10 +0000 (12:44 +0000)]
Fix a few macro definitions to ensure that unary minus is enclosed in
parentheses. This avoids possible operator precedence problems, and
is consistent with most of the macro definitions in the tree.

19 years agoRemove MMCacheLock -- it is no longer used. Per ITAGAKI Takahiro.
Neil Conway [Wed, 27 Jul 2005 08:05:36 +0000 (08:05 +0000)]
Remove MMCacheLock -- it is no longer used. Per ITAGAKI Takahiro.

19 years agoAvoid crashing pg_dump if we can't connect to the database server, and
Neil Conway [Wed, 27 Jul 2005 05:14:12 +0000 (05:14 +0000)]
Avoid crashing pg_dump if we can't connect to the database server, and
no database has been explicitly specified. Per gripe from Omar Kilani.

19 years agoBasic documentation for ROLEs. The user-manag chapter still needs to
Tom Lane [Tue, 26 Jul 2005 23:24:02 +0000 (23:24 +0000)]
Basic documentation for ROLEs.  The user-manag chapter still needs to
be rewritten, but at least the reference pages are reasonably sane.

19 years agoMinor correction: cause ALTER ROLE role ROLE rolenames to behave
Tom Lane [Tue, 26 Jul 2005 22:37:50 +0000 (22:37 +0000)]
Minor correction: cause ALTER ROLE role ROLE rolenames to behave
sensibly, even though we don't document it.

19 years agoAdd a role property 'rolinherit' which, when false, denotes that the role
Tom Lane [Tue, 26 Jul 2005 16:38:29 +0000 (16:38 +0000)]
Add a role property 'rolinherit' which, when false, denotes that the role
doesn't automatically inherit the privileges of roles it is a member of;
for such a role, membership in another role can be exploited only by doing
explicit SET ROLE.  The default inherit setting is TRUE, so by default
the behavior doesn't change, but creating a user with NOINHERIT gives closer
adherence to our current reading of SQL99.  Documentation still lacking,
and I think the information schema needs another look.

19 years agoAdd pg_has_role() family of privilege inquiry functions modeled after the
Tom Lane [Tue, 26 Jul 2005 00:04:19 +0000 (00:04 +0000)]
Add pg_has_role() family of privilege inquiry functions modeled after the
existing ones for object privileges.  Update the information_schema for
roles --- pg_has_role() makes this a whole lot easier, removing the need
for most of the explicit joins with pg_user.  The views should be a tad
faster now, too.  Stephen Frost and Tom Lane.

19 years agoAdd SET ROLE. This is a partial commit of Stephen Frost's recent patch;
Tom Lane [Mon, 25 Jul 2005 22:12:34 +0000 (22:12 +0000)]
Add SET ROLE.  This is a partial commit of Stephen Frost's recent patch;
I'm still working on the has_role function and information_schema changes.

19 years agoAdd username for psql password prompt, if the username was specified.
Bruce Momjian [Mon, 25 Jul 2005 17:17:41 +0000 (17:17 +0000)]
Add username for psql password prompt, if the username was specified.

Adrian Maier

19 years agoAdd:
Bruce Momjian [Mon, 25 Jul 2005 15:27:06 +0000 (15:27 +0000)]
Add:

>  o Allow postgresql.conf file values to be changed via an SQL API
>  o Allow the server to be stopped/restarted via an SQL API

19 years agoAwhile back we replaced all uses of strcasecmp and strncasecmp with
Tom Lane [Mon, 25 Jul 2005 04:52:32 +0000 (04:52 +0000)]
Awhile back we replaced all uses of strcasecmp and strncasecmp with
pg_strcasecmp and pg_strncasecmp ... but I see some of the former have
crept back in.
Eternal vigilance is the price of locale independence, apparently.

19 years agoChange build of regress.so to use Makefile.shlib instead of depending
Tom Lane [Mon, 25 Jul 2005 00:58:27 +0000 (00:58 +0000)]
Change build of regress.so to use Makefile.shlib instead of depending
on the not-very-good .so pattern rules in the port-specific Makefiles.
(This leaves only pgxs' MODULES case needing those rules.)  Also,
compile pgsleep.c locally and add it to regress.so to avoid failure
on AIX.

19 years agoPartial fixes for contrib build on AIX: include -lm where needed.
Tom Lane [Sun, 24 Jul 2005 23:30:10 +0000 (23:30 +0000)]
Partial fixes for contrib build on AIX: include -lm where needed.
Per Rocco Altier.

19 years agoRegenerate badly-obsolete README file.
Tom Lane [Sun, 24 Jul 2005 17:18:52 +0000 (17:18 +0000)]
Regenerate badly-obsolete README file.

19 years agoWith the interval/day patch, the horology regression test no longer
Tom Lane [Sun, 24 Jul 2005 17:07:18 +0000 (17:07 +0000)]
With the interval/day patch, the horology regression test no longer
fails near DST transition days, so remove the advice about that testing
problem.  Also improve the description of variant-comparison-file
selection.

19 years agoAdd variant expected file to cope with different spelling of a Python
Tom Lane [Sun, 24 Jul 2005 14:53:42 +0000 (14:53 +0000)]
Add variant expected file to cope with different spelling of a Python
error message in Python 2.3 and before.  Per Michael Fuhr and buildfarm
results.

19 years agoFix rounding problem in interval_div by using rint(), and improve
Bruce Momjian [Sun, 24 Jul 2005 04:37:07 +0000 (04:37 +0000)]
Fix rounding problem in interval_div by using rint(), and improve
interval_mul function.

19 years agoFix logic error in tbm_intersect: the intersection of a normal page and
Tom Lane [Sun, 24 Jul 2005 02:25:26 +0000 (02:25 +0000)]
Fix logic error in tbm_intersect: the intersection of a normal page and
a lossy page has to be lossy, because we don't know exactly which tuples
on the page should remain part of the bitmap.  Per Jie Zhang.

19 years agoFix some failures to initialize table entries induced by recent autovacuum
Tom Lane [Sun, 24 Jul 2005 00:33:28 +0000 (00:33 +0000)]
Fix some failures to initialize table entries induced by recent autovacuum
integration.  Not clear this explains recent stats problems, but it's
definitely wrong.

19 years agoSimple constraint exclusion. For now, only child tables of inheritance
Tom Lane [Sat, 23 Jul 2005 21:05:48 +0000 (21:05 +0000)]
Simple constraint exclusion.  For now, only child tables of inheritance
scans are candidates for exclusion; this should be fixed eventually.
Simon Riggs, with some help from Tom Lane.

19 years agoRemove unintended code addition.
Bruce Momjian [Sat, 23 Jul 2005 15:31:16 +0000 (15:31 +0000)]
Remove unintended code addition.

19 years agoMacro alignment cleanup.
Bruce Momjian [Sat, 23 Jul 2005 15:29:47 +0000 (15:29 +0000)]
Macro alignment cleanup.

19 years agoImprove computations of interval_div to prevent rounding problem on AIX.
Bruce Momjian [Sat, 23 Jul 2005 14:53:21 +0000 (14:53 +0000)]
Improve computations of interval_div to prevent rounding problem on AIX.

19 years agoAndrew pointed out that the current fix didn't handle dates that were
Bruce Momjian [Sat, 23 Jul 2005 14:25:34 +0000 (14:25 +0000)]
Andrew pointed out that the current fix didn't handle dates that were
near daylight savings time boudaries.  This handles it properly, e.g.

        test=> select '2005-04-03 04:00:00'::timestamp at time zone
        'America/Los_Angeles';
                timezone
        ------------------------
         2005-04-03 07:00:00-04
        (1 row)

19 years agoIn the stats test, delay for the stats collector to catch up using a
Tom Lane [Sat, 23 Jul 2005 14:18:57 +0000 (14:18 +0000)]
In the stats test, delay for the stats collector to catch up using a
function that actually sleeps, instead of busy-waiting.  Perhaps this
will resolve some of the intermittent stats failures we keep seeing.

19 years agoFix AT TIME ZONE for timestamps without time zones:
Bruce Momjian [Sat, 23 Jul 2005 02:02:27 +0000 (02:02 +0000)]
Fix AT TIME ZONE for timestamps without time zones:

test=> select (CURRENT_DATE + '05:00'::time)::timestamp at time zone
'Canada/Pacific';
        timezone
------------------------
 2005-07-22 08:00:00-04
(1 row)

19 years agoFix AT TIME ZONE for timestamps without time zones:
Bruce Momjian [Fri, 22 Jul 2005 21:16:15 +0000 (21:16 +0000)]
Fix AT TIME ZONE for timestamps without time zones:

test=> select ('2005-07-20 00:00:00'::timestamp without time zone) at
time zone 'Europe/Paris';
        timezone
------------------------
 2005-07-19 22:00:00-04

Udpate documentation.

19 years agoFix insufficient check for overflow in tm2abstime(), per report from jw.
Tom Lane [Fri, 22 Jul 2005 19:55:50 +0000 (19:55 +0000)]
Fix insufficient check for overflow in tm2abstime(), per report from jw.

19 years agoFix compare_fuzzy_path_costs() to behave a bit more sanely. The original
Tom Lane [Fri, 22 Jul 2005 19:12:02 +0000 (19:12 +0000)]
Fix compare_fuzzy_path_costs() to behave a bit more sanely.  The original
coding would ignore startup cost differences of less than 1% of the
estimated total cost; which was OK for normal planning but highly not OK
if a very small LIMIT was applied afterwards, so that startup cost becomes
the name of the game.  Instead, compare startup and total costs fuzzily
but independently.  This changes the plan selected for two queries in the
regression tests; adjust expected-output files for resulting changes in
row order.  Per reports from Dawid Kuroczko and Sam Mason.

19 years agoMore removal of unneeded parentheses.
Bruce Momjian [Fri, 22 Jul 2005 19:00:55 +0000 (19:00 +0000)]
More removal of unneeded parentheses.

19 years agoMore spacing adjustments
Bruce Momjian [Fri, 22 Jul 2005 15:15:38 +0000 (15:15 +0000)]
More spacing adjustments

19 years agoUpdate date/time comments.
Bruce Momjian [Fri, 22 Jul 2005 05:08:26 +0000 (05:08 +0000)]
Update date/time comments.

19 years agoMore minor spacing improvements.
Bruce Momjian [Fri, 22 Jul 2005 05:03:09 +0000 (05:03 +0000)]
More minor spacing improvements.

19 years agoCode spacing improvement, particularly *tm spacing.
Bruce Momjian [Fri, 22 Jul 2005 03:46:34 +0000 (03:46 +0000)]
Code spacing improvement, particularly *tm spacing.

19 years agoMore comment update of time macros.
Bruce Momjian [Thu, 21 Jul 2005 20:37:21 +0000 (20:37 +0000)]
More comment update of time macros.

19 years agoUpdate DAYS_PER_MONTH comment.
Bruce Momjian [Thu, 21 Jul 2005 18:06:13 +0000 (18:06 +0000)]
Update DAYS_PER_MONTH comment.

Add SECS_PER_YEAR and MINS_PER_HOUR macros.

19 years agoAdd comment about void* use in MemSet.
Bruce Momjian [Thu, 21 Jul 2005 15:16:30 +0000 (15:16 +0000)]
Add comment about void* use in MemSet.

19 years agoAdd:
Bruce Momjian [Thu, 21 Jul 2005 13:43:16 +0000 (13:43 +0000)]
Add:

<
> * Research storing disk pages with no alignment/padding

19 years agoFix integer timestamp build for macro changes.
Bruce Momjian [Thu, 21 Jul 2005 05:18:26 +0000 (05:18 +0000)]
Fix integer timestamp build for macro changes.

19 years agoAdd comment marking non-exact time conversion macros.
Bruce Momjian [Thu, 21 Jul 2005 04:48:42 +0000 (04:48 +0000)]
Add comment marking non-exact time conversion macros.

19 years agoRemove unnecessary parentheses in assignments.
Bruce Momjian [Thu, 21 Jul 2005 04:41:43 +0000 (04:41 +0000)]
Remove unnecessary parentheses in assignments.
Add spaces where needed.
Reference time interval variables as tinterval.

19 years agoFix storage size for btree_gist interval indexes. Fix penalty
Tom Lane [Thu, 21 Jul 2005 04:15:04 +0000 (04:15 +0000)]
Fix storage size for btree_gist interval indexes.  Fix penalty
calculations for interval and time/timetz to behave sanely for both
integer and float timestamps; up to now I think it's been doing
something pretty strange...

19 years agoAdd time/date macros for code clarity:
Bruce Momjian [Thu, 21 Jul 2005 03:56:25 +0000 (03:56 +0000)]
Add time/date macros for code clarity:

#define DAYS_PER_YEAR   365.25
#define MONTHS_PER_YEAR 12
#define DAYS_PER_MONTH  30
#define HOURS_PER_DAY   24

19 years agoFix breakage of INSTALL document build.
Tom Lane [Wed, 20 Jul 2005 23:57:46 +0000 (23:57 +0000)]
Fix breakage of INSTALL document build.

19 years ago/contrib/btree_gist adjustments for new 'day' interval value.
Bruce Momjian [Wed, 20 Jul 2005 18:17:39 +0000 (18:17 +0000)]
/contrib/btree_gist adjustments for new 'day' interval value.

19 years agoUpdate catalog version for INTERVAL day addition.
Bruce Momjian [Wed, 20 Jul 2005 17:24:39 +0000 (17:24 +0000)]
Update catalog version for INTERVAL day addition.

19 years agoAdd 'day' field to INTERVAL so 1 day interval can be distinguished from
Bruce Momjian [Wed, 20 Jul 2005 16:42:32 +0000 (16:42 +0000)]
Add 'day' field to INTERVAL so 1 day interval can be distinguished from
24 hours. This is very helpful for daylight savings time:

select '2005-05-03 00:00:00 EST'::timestamp with time zone + '24 hours';
      ?column?
----------------------
2005-05-04 01:00:00-04

select '2005-05-03 00:00:00 EST'::timestamp with time zone + '1 day';
      ?column?
----------------------
2005-05-04 01:00:00-04

Michael Glaesemann

19 years agoFix interval division and multiplication, before:
Bruce Momjian [Wed, 20 Jul 2005 03:50:24 +0000 (03:50 +0000)]
Fix interval division and multiplication, before:

test=> select '4 months'::interval / 5;
   ?column?
---------------
 1 mon -6 days
(1 row)

after:

test=> select '4 months'::interval / 5;
 ?column?
----------
 24 days
(1 row)

The problem was the use of rint() to round, and then find the remainder,
causing the negative values.

19 years agoDocumentation tweak: make <command>CREATE OPERATOR CLASS</command> into
Neil Conway [Tue, 19 Jul 2005 01:27:59 +0000 (01:27 +0000)]
Documentation tweak: make <command>CREATE OPERATOR CLASS</command> into
an <xref/>.

19 years agoFix some bogosities in geometric-function documentation: add an entry
Tom Lane [Mon, 18 Jul 2005 22:34:14 +0000 (22:34 +0000)]
Fix some bogosities in geometric-function documentation: add an entry
for circle(polygon), which was missing; remove bogus entry for
point(lseg, lseg), which does not exist, and the documentation seemed to
describe lseg_interpt, which we already document as an operator not a
function.  Also remove entry for box_intersect, which likewise is
preferentially used via the operator #.

19 years ago\pset numericsep -> numericlocale.
Bruce Momjian [Mon, 18 Jul 2005 20:57:53 +0000 (20:57 +0000)]
\pset numericsep -> numericlocale.

19 years agoFix len so decimal length is only added when a period appears in the output.
Bruce Momjian [Mon, 18 Jul 2005 19:27:37 +0000 (19:27 +0000)]
Fix len so decimal length is only added when a period appears in the output.

19 years agoAvoid use of E'', and thereby creating an unnecessary version-dependency,
Tom Lane [Mon, 18 Jul 2005 19:12:09 +0000 (19:12 +0000)]
Avoid use of E'', and thereby creating an unnecessary version-dependency,
by using LIKE...ESCAPE instead.  Per suggestion by andrew@supernews.

19 years agoAvoid need for E'' construct by using regexp instead of LIKE. More
Tom Lane [Mon, 18 Jul 2005 19:09:09 +0000 (19:09 +0000)]
Avoid need for E'' construct by using regexp instead of LIKE.  More
readable, and more like the other places in this file.

19 years agoFix numiericsep length computations.
Bruce Momjian [Mon, 18 Jul 2005 18:58:45 +0000 (18:58 +0000)]
Fix numiericsep length computations.

Eugen Nedelcu

19 years agoAdjust psql describe queries so that any pg_foo_is_visible() condition
Tom Lane [Mon, 18 Jul 2005 17:40:14 +0000 (17:40 +0000)]
Adjust psql describe queries so that any pg_foo_is_visible() condition
is applied last, after other constraints such as name patterns.  This
is useful first because the pg_foo_is_visible() functions are relatively
expensive, and second because it minimizes the prospects for race
conditions.  The change is fragile though since it makes unwarranted
assumptions about planner behavior, ie, that WHERE clauses will be
executed in the original order if there's not reason to change it.
This should fix ... or at least hide ... an intermittent failure in the
prepared_xacts regression test, while we think about what else to do.

19 years agopgcrypto documentation update. Marko Kreen
Tom Lane [Mon, 18 Jul 2005 17:17:12 +0000 (17:17 +0000)]
pgcrypto documentation update.  Marko Kreen

19 years agoFortuna fixes. Marko Kreen
Tom Lane [Mon, 18 Jul 2005 17:12:54 +0000 (17:12 +0000)]
Fortuna fixes.  Marko Kreen

19 years agoSmall cleanups for pgcrypto. Marko Kreen
Tom Lane [Mon, 18 Jul 2005 17:09:01 +0000 (17:09 +0000)]
Small cleanups for pgcrypto.  Marko Kreen

19 years agoFix C++-style comments, per Rocco Altier.
Tom Lane [Mon, 18 Jul 2005 16:35:06 +0000 (16:35 +0000)]
Fix C++-style comments, per Rocco Altier.

19 years agoMemSet() must not cast its pointer argument to int32* until after it has
Tom Lane [Mon, 18 Jul 2005 15:53:28 +0000 (15:53 +0000)]
MemSet() must not cast its pointer argument to int32* until after it has
checked that the pointer is actually word-aligned.  Casting a non-aligned
pointer to int32* is technically illegal per the C spec, and some recent
versions of gcc actually generate bad code for the memset() when given
such a pointer.  Per report from Andrew Morrow.

19 years agoMake pg_regress accept a command-line option for the temporary installation's
Tom Lane [Sun, 17 Jul 2005 18:28:45 +0000 (18:28 +0000)]
Make pg_regress accept a command-line option for the temporary installation's
port number, and use a default value for it that is dependent on the
configuration-time DEF_PGPORT.  Should make the world safe for running
parallel 'make check' in different branches.  Back-patch as far as 7.4
so that this actually is useful.

19 years agoFix create_unique_plan() so it doesn't generate useless entries in the
Tom Lane [Fri, 15 Jul 2005 22:02:51 +0000 (22:02 +0000)]
Fix create_unique_plan() so it doesn't generate useless entries in the
output targetlist of the Unique or HashAgg plan.  This code was OK when
written, but subsequent changes to use "physical tlists" where possible
had broken it: given an input subplan that has extra variables added to
avoid a projection step, it would copy those extra variables into the
upper tlist, which is pointless since a projection has to happen anyway.

19 years agoFix SGML spacing.
Bruce Momjian [Fri, 15 Jul 2005 19:02:12 +0000 (19:02 +0000)]
Fix SGML spacing.

19 years agoCheck for out-of-range varoattno in deparse_context_for_subplan.
Tom Lane [Fri, 15 Jul 2005 18:39:59 +0000 (18:39 +0000)]
Check for out-of-range varoattno in deparse_context_for_subplan.
I have seen this case in CVS tip due to new "physical tlist" optimization
for subqueries.  I believe it probably can't happen in existing releases,
but the check is not going to hurt anything, so backpatch to 8.0 just
in case.

19 years agoFix overenthusiastic optimization of 'x IN (SELECT DISTINCT ...)' and related
Tom Lane [Fri, 15 Jul 2005 17:09:26 +0000 (17:09 +0000)]
Fix overenthusiastic optimization of 'x IN (SELECT DISTINCT ...)' and related
cases: we can't just consider whether the subquery's output is unique on its
own terms, we have to check whether the set of output columns we are going to
use will be unique.  Per complaint from Luca Pireddu and test case from
Michael Fuhr.

19 years agoAdjust permissions checking for ALTER OWNER commands: instead of
Tom Lane [Thu, 14 Jul 2005 21:46:30 +0000 (21:46 +0000)]
Adjust permissions checking for ALTER OWNER commands: instead of
requiring superuserness always, allow an owner to reassign ownership
to any role he is a member of, if that role would have the right to
create a similar object.  These three requirements essentially state
that the would-be alterer has enough privilege to DROP the existing
object and then re-CREATE it as the new role; so we might as well
let him do it in one step.  The ALTER TABLESPACE case is a bit
squirrely, but the whole concept of non-superuser tablespace owners
is pretty dubious anyway.  Stephen Frost, code review by Tom Lane.

19 years agoEnable multi-byte thousands_sep and decimal_point for numericsep.
Bruce Momjian [Thu, 14 Jul 2005 21:12:41 +0000 (21:12 +0000)]
Enable multi-byte thousands_sep and decimal_point for numericsep.

19 years agoMake better thousands_sep default based on decimal marker value.
Bruce Momjian [Thu, 14 Jul 2005 15:54:21 +0000 (15:54 +0000)]
Make better thousands_sep default based on decimal marker value.

19 years agoChange numericsep to a boolean, and make it locale-aware.
Bruce Momjian [Thu, 14 Jul 2005 08:42:37 +0000 (08:42 +0000)]
Change numericsep to a boolean, and make it locale-aware.

19 years agoAdd numeric separator support for latex and troff output methods.
Bruce Momjian [Thu, 14 Jul 2005 07:32:01 +0000 (07:32 +0000)]
Add numeric separator support for latex and troff output methods.