OSDN Git Service

pg-rex/syncrep.git
20 years agoattached is a patch that adds display of the groups a user belongs to to
Bruce Momjian [Mon, 1 Dec 2003 22:21:54 +0000 (22:21 +0000)]
attached is a patch that adds display of the groups a user belongs to to
\du and a \dg command to psql. It's against 7.4beta5.

Markus Bertheau <twanger@bluetwanger.de>

20 years agoAvoid assuming that type key_t is 32 bits, since it reportedly isn't
Tom Lane [Mon, 1 Dec 2003 22:15:38 +0000 (22:15 +0000)]
Avoid assuming that type key_t is 32 bits, since it reportedly isn't
on 64-bit Solaris.  Use a non-system-dependent datatype for UsedShmemSegID,
namely unsigned long (which we were already assuming could hold a shmem
key anyway, cf RecordSharedMemoryInLockFile).

20 years agoThis patch reduces some unsightly #ifdefs, and fixes two typos in
Bruce Momjian [Mon, 1 Dec 2003 22:14:40 +0000 (22:14 +0000)]
This patch reduces some unsightly #ifdefs, and fixes two typos in
comments in the psql code. This doesn't make any functional change, so
feel free to save it for 7.5

Neil Conway

20 years agoThis patch finishes off the work that I did with making view
Bruce Momjian [Mon, 1 Dec 2003 22:11:06 +0000 (22:11 +0000)]
This patch finishes off the work that I did with making view
definitions use pretty printing.

It does:

* Pretty index predicates
* Pretty rule definitions
* Uppercases PRIMARY KEY and UNIQUE to be consistent with CHECK and
FOREIGN KEY
* View rules are improved to match table rules:

Christopher Kings-Lynne

20 years agoThis patch adds a new GUC var, "default_with_oids", which follows the
Bruce Momjian [Mon, 1 Dec 2003 22:08:02 +0000 (22:08 +0000)]
This patch adds a new GUC var, "default_with_oids", which follows the
proposal for eventually deprecating OIDs on user tables that I posted
earlier to pgsql-hackers. pg_dump now always specifies WITH OIDS or
WITHOUT OIDS when dumping a table. The documentation has been updated.

Neil Conway

20 years agoTry to reduce confusion about what is a lock method identifier, a lock
Bruce Momjian [Mon, 1 Dec 2003 21:59:25 +0000 (21:59 +0000)]
Try to reduce confusion about what is a lock method identifier, a lock
method control structure, or a table of control structures.

. Use type LOCKMASK where an int is not a counter.

. Get rid of INVALID_TABLEID, use INVALID_LOCKMETHOD instead.

. Use INVALID_LOCKMETHOD instead of (LOCKMETHOD) NULL, because
  LOCKMETHOD is not a pointer.

. Define and use macro LockMethodIsValid.

. Rename LOCKMETHOD to LOCKMETHODID.

. Remove global variable LongTermTableId in lmgr.c, because it is
  never used.

. Make LockTableId static in lmgr.c, because it is used nowhere else.
  Why not remove it and use DEFAULT_LOCKMETHOD?

. Rename the lock method control structure from LOCKMETHODTABLE to
  LockMethodData.  Introduce a pointer type named LockMethod.

. Remove elog(FATAL) after InitLockTable() call in
  CreateSharedMemoryAndSemaphores(), because if something goes wrong,
  there is elog(FATAL) in LockMethodTableInit(), and if this doesn't
  help, an elog(ERROR) in InitLockTable() is promoted to FATAL.

. Make InitLockTable() void, because its only caller does not use its
  return value any more.

. Rename variables in lock.c to avoid statements like
        LockMethodTable[NumLockMethods] = lockMethodTable;
        lockMethodTable = LockMethodTable[lockmethod];

. Change LOCKMETHODID type to uint16 to fit into struct LOCKTAG.

. Remove static variables BITS_OFF and BITS_ON from lock.c, because
  I agree to this doubt:
 * XXX is a fetch from a static array really faster than a shift?

. Define and use macros LOCKBIT_ON/OFF.

Manfred Koizar

20 years agoRemove mention of factorial as casting example. New example needed.
Bruce Momjian [Mon, 1 Dec 2003 21:53:15 +0000 (21:53 +0000)]
Remove mention of factorial as casting example.  New example needed.

20 years agoAttached is a patch implementing factorial(), returning numeric. Points
Bruce Momjian [Mon, 1 Dec 2003 21:52:38 +0000 (21:52 +0000)]
Attached is a patch implementing factorial(), returning numeric. Points
to note:

1) arttype is numeric. I thought this was the best way of allowing
arbitarily large factorials, even though factorial(2^63) is a large
number. Happy to change to integers if this is overkill.
2) since we're accepting numeric arguments, the patch tests for floats.
If a numeric is passed with non-zero decimal portion, an error is raised
since (from memory) they are undefined.

Gavin Sherry

20 years agoMinor wordsmithing in datetime docs to try to address gripes raised by
Tom Lane [Mon, 1 Dec 2003 20:34:53 +0000 (20:34 +0000)]
Minor wordsmithing in datetime docs to try to address gripes raised by
cnliou.

20 years agonetmask() and hostmask() functions should return maximum-length masklen,
Tom Lane [Mon, 1 Dec 2003 18:50:19 +0000 (18:50 +0000)]
netmask() and hostmask() functions should return maximum-length masklen,
per gripe from Joe Sunday.

20 years agoFix typos noted by Halley Pacheco de Oliveira.
Tom Lane [Mon, 1 Dec 2003 17:58:27 +0000 (17:58 +0000)]
Fix typos noted by Halley Pacheco de Oliveira.

20 years agoForce zero_damaged_pages to be effectively ON during recovery from WAL,
Tom Lane [Mon, 1 Dec 2003 16:53:19 +0000 (16:53 +0000)]
Force zero_damaged_pages to be effectively ON during recovery from WAL,
since there is no need to worry about damaged pages when we are going to
overwrite them anyway from the WAL.  Per recent discussion.

20 years agoSeems there are three GUC variables that are defined as "Shows ..."
Bruce Momjian [Mon, 1 Dec 2003 03:55:21 +0000 (03:55 +0000)]
Seems there are three GUC variables that are defined as "Shows ..."
while you can actually set them with SET.

This applied patch changes the wording from "Show" to "Set".

20 years agoMake PQescapeBytea and byteaout consistent with each other, and
Joe Conway [Sun, 30 Nov 2003 20:55:09 +0000 (20:55 +0000)]
Make PQescapeBytea and byteaout consistent with each other, and
octal escape all octets outside the range 0x20 to 0x7e. This fixes
the problem pointed out by Sergey Yatskevich here:
http://archives.postgresql.org/pgsql-bugs/2003-11/msg00140.php

20 years agoBump all version numbers and version stamps mentioned in RELEASE_CHANGES.
Bruce Momjian [Sun, 30 Nov 2003 06:09:54 +0000 (06:09 +0000)]
Bump all version numbers and version stamps mentioned in RELEASE_CHANGES.

20 years agoReorder win32/bcc makefile mentions of thread.c for sanity.
Bruce Momjian [Sun, 30 Nov 2003 06:01:15 +0000 (06:01 +0000)]
Reorder win32/bcc makefile mentions of thread.c for sanity.

20 years agoAdd thread.c to Borland CC build.
Bruce Momjian [Sun, 30 Nov 2003 05:54:33 +0000 (05:54 +0000)]
Add thread.c to Borland CC build.

20 years agoRemove INTO recommendation for plpgsql language.
Bruce Momjian [Sun, 30 Nov 2003 05:45:22 +0000 (05:45 +0000)]
Remove INTO recommendation for plpgsql language.

20 years agoAdd:
Bruce Momjian [Sun, 30 Nov 2003 05:29:26 +0000 (05:29 +0000)]
Add:

> * Have pg_ctl look at PGHOST in case it is a socket directory

20 years agoFix URL.
Bruce Momjian [Sun, 30 Nov 2003 04:56:47 +0000 (04:56 +0000)]
Fix URL.

20 years ago*** empty log message ***
Bruce Momjian [Sun, 30 Nov 2003 04:54:07 +0000 (04:54 +0000)]
*** empty log message ***

20 years agoThere is an unused variable in an example function in the PL/PgSQL
Bruce Momjian [Sun, 30 Nov 2003 04:47:37 +0000 (04:47 +0000)]
There is an unused variable in an example function in the PL/PgSQL
documentation; this patch removes it.

Neil Conway

20 years agoAdd:
Bruce Momjian [Sun, 30 Nov 2003 04:44:38 +0000 (04:44 +0000)]
Add:

> * Make LENGTH() of CHAR() not count trailing spaces
>

20 years agoRemove win32-only mention for pgadmin.
Bruce Momjian [Sun, 30 Nov 2003 04:35:53 +0000 (04:35 +0000)]
Remove win32-only mention for pgadmin.

20 years ago*** empty log message ***
Bruce Momjian [Sun, 30 Nov 2003 04:34:14 +0000 (04:34 +0000)]
*** empty log message ***

20 years agoUpdate wording:
Bruce Momjian [Sun, 30 Nov 2003 04:07:04 +0000 (04:07 +0000)]
Update wording:

< * Allow clients to query WITH HOLD cursors and prepared statements
> * Allow clients to query a list of WITH HOLD cursors and prepared statements

20 years agoUpdates:
Bruce Momjian [Sun, 30 Nov 2003 03:41:10 +0000 (03:41 +0000)]
Updates:

<   manuals (Rory)
>   manuals
496c496
< * Jan is Jan Wieck <JanWieck@Yahoo.com> of PeerDirect Corp.
> * Jan is Jan Wieck <JanWieck@Yahoo.com> of Afilias, Inc.

20 years agomake sure the $Id tags are converted to $PostgreSQL as well ...
PostgreSQL Daemon [Sat, 29 Nov 2003 22:41:33 +0000 (22:41 +0000)]
make sure the $Id tags are converted to $PostgreSQL as well ...

20 years agoPut out a more useful version indication in the welcome banner for a
Tom Lane [Sat, 29 Nov 2003 21:40:43 +0000 (21:40 +0000)]
Put out a more useful version indication in the welcome banner for a
standalone backend --- the CVS revision number of postgres.c is not real
useful to anyone.

20 years ago$Header: -> $PostgreSQL Changes ...
PostgreSQL Daemon [Sat, 29 Nov 2003 19:52:15 +0000 (19:52 +0000)]
$Header: -> $PostgreSQL Changes ...

20 years agotesting $PostgreSQL$ tag
PostgreSQL Daemon [Fri, 28 Nov 2003 20:32:09 +0000 (20:32 +0000)]
testing $PostgreSQL$ tag

20 years agoAttached is a patch to correct white space issues in FAQ_MSWIN.
Bruce Momjian [Fri, 28 Nov 2003 20:22:18 +0000 (20:22 +0000)]
Attached is a patch to correct white space issues in FAQ_MSWIN.

Daniel Convissor

20 years agoUpdate Russian FAQ, both branches.
Bruce Momjian [Fri, 28 Nov 2003 20:20:33 +0000 (20:20 +0000)]
Update Russian FAQ, both branches.

Viktor Vislobokov

20 years agoUpdate most recent version to 7.4.
Bruce Momjian [Fri, 28 Nov 2003 20:09:30 +0000 (20:09 +0000)]
Update most recent version to 7.4.

20 years agoFix free instead of pfree
Teodor Sigaev [Fri, 28 Nov 2003 12:09:02 +0000 (12:09 +0000)]
Fix free instead of pfree

20 years agoFix regression in dblink_disconnect() reported by Eduardo Stern:
Joe Conway [Fri, 28 Nov 2003 05:03:02 +0000 (05:03 +0000)]
Fix regression in dblink_disconnect() reported by Eduardo Stern:
persistent_conn was left dangling after a disconnect in the
unnamed connection case, causing a subsequent disconnect to
crash the backend.

20 years agoRemove possibility to specify an installation root after --with-krbx and
Peter Eisentraut [Thu, 27 Nov 2003 19:44:56 +0000 (19:44 +0000)]
Remove possibility to specify an installation root after --with-krbx and
--with-openssl options.  This creates too much risk to pick up the wrong
directory accidentally (for example when there are lib64 directories), and
does not really help much with contemporary installation layouts.

20 years agoMake sure vpath_build is set correctly even if config.status --no-create
Peter Eisentraut [Thu, 27 Nov 2003 18:14:02 +0000 (18:14 +0000)]
Make sure vpath_build is set correctly even if config.status --no-create
was specified.

20 years agoConditionalize variable that is only used conditionally, to avoid warning.
Peter Eisentraut [Thu, 27 Nov 2003 18:12:50 +0000 (18:12 +0000)]
Conditionalize variable that is only used conditionally, to avoid warning.

20 years agoFixes about word with several infiniteve
Teodor Sigaev [Thu, 27 Nov 2003 16:04:40 +0000 (16:04 +0000)]
Fixes about word with several infiniteve

20 years agoSecond try at fixing no-room-to-move-down PANIC in compact_fsm_storage.
Tom Lane [Wed, 26 Nov 2003 20:50:11 +0000 (20:50 +0000)]
Second try at fixing no-room-to-move-down PANIC in compact_fsm_storage.
Ward's report that it can still happen in RC2 forces me to realize that
this is not a can't-happen condition after all, and that the compaction
code had better cope rather than panicking.

20 years agoAdded missing SPI_finish() calls to get_tuple_of_interest(). Fixes bug
Joe Conway [Wed, 26 Nov 2003 20:43:25 +0000 (20:43 +0000)]
Added missing SPI_finish() calls to get_tuple_of_interest(). Fixes bug
reported by Andrea Grassi.

20 years agoAdd NetBSD Sparc as supported.
Peter Eisentraut [Wed, 26 Nov 2003 15:56:48 +0000 (15:56 +0000)]
Add NetBSD Sparc as supported.

20 years agoCast field-length variables used in printf to int, because sometimes
Peter Eisentraut [Wed, 26 Nov 2003 15:55:01 +0000 (15:55 +0000)]
Cast field-length variables used in printf to int, because sometimes
they might be of a wider type.

20 years agoUtility for convert myspell dictionaries to ispell, full README will be later
Teodor Sigaev [Wed, 26 Nov 2003 14:06:16 +0000 (14:06 +0000)]
Utility for convert myspell dictionaries to ispell, full README will be later

20 years agoPatches from Yutaka Tanida. Create primary key indexes after data
Tatsuo Ishii [Wed, 26 Nov 2003 06:53:18 +0000 (06:53 +0000)]
Patches from Yutaka Tanida. Create primary key indexes after data
insertion to reduce initialization time.

20 years agoRepair subselect.c's occasional assignment of the wrong vartypmod to
Tom Lane [Tue, 25 Nov 2003 23:59:12 +0000 (23:59 +0000)]
Repair subselect.c's occasional assignment of the wrong vartypmod to
Vars created to fill subplan args lists.  This is an ancient error, going
back at least to 7.0, but is more easily triggered in 7.4 than before
because we no longer compare varlevelsup when deciding whether a Param
slot can be re-used.  Fixes bug reported by Klint Gore.

20 years agoGet rid of hashkeys field of Hash plan node, since it's redundant with
Tom Lane [Tue, 25 Nov 2003 21:00:54 +0000 (21:00 +0000)]
Get rid of hashkeys field of Hash plan node, since it's redundant with
the hashclauses field of the parent HashJoin.  This avoids problems with
duplicated links to SubPlans in hash clauses, as per report from
Andrew Holm-Hansen.

20 years agoFix language.
Peter Eisentraut [Tue, 25 Nov 2003 19:27:18 +0000 (19:27 +0000)]
Fix language.

20 years agoNew translation
Peter Eisentraut [Tue, 25 Nov 2003 19:19:21 +0000 (19:19 +0000)]
New translation

20 years agoMore adjustment of error messages
Peter Eisentraut [Tue, 25 Nov 2003 19:18:26 +0000 (19:18 +0000)]
More adjustment of error messages

20 years agoInstall all the headers files that the ones that are already installed
Peter Eisentraut [Tue, 25 Nov 2003 19:09:02 +0000 (19:09 +0000)]
Install all the headers files that the ones that are already installed
depend on.

20 years agoIgnore too long lexeme
Teodor Sigaev [Tue, 25 Nov 2003 13:33:15 +0000 (13:33 +0000)]
Ignore too long lexeme

20 years agoUpgrade to DocBook V4.2 SGML.
Peter Eisentraut [Mon, 24 Nov 2003 19:08:02 +0000 (19:08 +0000)]
Upgrade to DocBook V4.2 SGML.

20 years agoAvoid using string literal with embedded newline.
Tom Lane [Mon, 24 Nov 2003 17:25:14 +0000 (17:25 +0000)]
Avoid using string literal with embedded newline.

20 years agoFix markup.
Peter Eisentraut [Mon, 24 Nov 2003 16:56:00 +0000 (16:56 +0000)]
Fix markup.

20 years agoOverdue code review for ALTER SEQUENCE patch. Don't generate illegal Node
Tom Lane [Mon, 24 Nov 2003 16:54:07 +0000 (16:54 +0000)]
Overdue code review for ALTER SEQUENCE patch.  Don't generate illegal Node
tree for CYCLE option; don't assume zeros are invalid values for sequence
fields other than increment_by; don't reset cache_value when not told to;
simplify code for testing whether to apply defaults.

20 years agoUse --with-docdir to choose installation location of documentation; put
Peter Eisentraut [Mon, 24 Nov 2003 14:52:58 +0000 (14:52 +0000)]
Use --with-docdir to choose installation location of documentation; put
back --infodir, which several automatic build environments expect to exist.
Add --without-docdir to prevent installation of documentation, which is
helpful for things like RPM that have their own method of installing
documentation.

20 years agoRepair missed renamings of show_statement_stats and show_executor_stats.
Tom Lane [Mon, 24 Nov 2003 14:49:51 +0000 (14:49 +0000)]
Repair missed renamings of show_statement_stats and show_executor_stats.

20 years agoRename USE_THREADS to ENABLE_THREAD_SAFETY to avoid name clash with Perl.
Peter Eisentraut [Mon, 24 Nov 2003 13:16:22 +0000 (13:16 +0000)]
Rename USE_THREADS to ENABLE_THREAD_SAFETY to avoid name clash with Perl.
Fixes compilation failure with --enable-thread-safety --with-perl and Perl
5.6.1.

20 years agoMake the messages and the options parsing a bit more standard.
Peter Eisentraut [Sun, 23 Nov 2003 22:17:59 +0000 (22:17 +0000)]
Make the messages and the options parsing a bit more standard.

20 years agoAdd maintainer-clean target.
Peter Eisentraut [Sun, 23 Nov 2003 21:42:13 +0000 (21:42 +0000)]
Add maintainer-clean target.

20 years agoAdd NLS support.
Peter Eisentraut [Sun, 23 Nov 2003 21:41:30 +0000 (21:41 +0000)]
Add NLS support.

20 years agoCOMMENT ON casts, conversions, languages, operator classes, and
Tom Lane [Fri, 21 Nov 2003 22:32:49 +0000 (22:32 +0000)]
COMMENT ON casts, conversions, languages, operator classes, and
large objects.  Dump all these in pg_dump; also add code to pg_dump
user-defined conversions.  Make psql's large object code rely on
the backend for inserting/deleting LOB comments, instead of trying to
hack pg_description directly.  Documentation and regression tests added.

Christopher Kings-Lynne, code reviewed by Tom

20 years agoRemove unused variable.
Tom Lane [Fri, 21 Nov 2003 17:41:31 +0000 (17:41 +0000)]
Remove unused variable.

20 years agoMark item as done, update description for 7.5:
Bruce Momjian [Wed, 19 Nov 2003 17:35:40 +0000 (17:35 +0000)]
Mark item as done, update description for 7.5:

< A dash (-) marks changes that will appear in the upcoming 7.4 release.
> A dash (-) marks changes that will appear in the upcoming 7.5 release.
437c437
< * Use background process to write dirty shared buffers to disk
> * -Use background process to write dirty shared buffers to disk

20 years agoRemove items completed in 7.4:
Bruce Momjian [Wed, 19 Nov 2003 17:34:01 +0000 (17:34 +0000)]
Remove items completed in 7.4:

< * -Allow elog() to return error codes, module name, file name, line
<   number, not just messages (Tom)
< * -Add error codes (Tom)
< * -Make error messages more consistent
40d35
< * -Add GUC log_statement_and_duration to print statement and >= min duration
51d45
< * -Allow easy display of usernames in a group
53d46
< * -Add start time to pg_stat_activity
55d47
< * -Have standalone backend read postgresql.conf (Tom)
59d50
< * -Allow CIDR format to be used in pg_hba.conf
74d64
< * -Add IPv6 capability to INET/CIDR types
77d66
< * -Change NUMERIC data type to use base 10,000 internally
82d70
< * -Add GUC variables to control floating number output digits (Pedro Ferreira)
90,92d77
< * -Allow current datestyle to restrict dates;  prevent month/day swapping
<   from making invalid dates valid
< * -Prevent month/day swapping of ISO dates to make invalid dates valid
102d86
<  o -Allow arrays to be ORDER'ed
104d87
<  o -Support construction of array result values in expressions (Joe)
128d110
< * -Remove Cyrillic recode support
146,147d127
< * -Allow CREATE INDEX zman_index ON test (date_trunc( 'day', zman ) datetime_ops)
<   fails index can't store constant parameters
155d134
< * -Add btree index support for reltime, tinterval, regproc (Tom)
157d135
< * -Certain indexes will not shrink, e.g. indexes on ever-increasing
161d138
< * -Allow LIKE indexing optimization for non-ASCII locales using special index
173d149
< * -Improve concurrency of hash indexes (Tom)
181d156
< * -Allow LIMIT/OFFSET to use expressions (Tom)
187d161
< * -Return proper effected tuple count from complex commands [return]
191d164
< * -Make a transaction-safe TRUNCATE (Rod)
196,197d168
< * -Allow UPDATE to use SET col = DEFAULT
< * -Add config variable to prevent auto-adding missing FROM-clause tables
199d169
< * -Have SELECT '13 minutes'::interval display zero seconds in ISO datestyle
224,225d193
<  o -Add ALTER TABLE tab SET WITHOUT OIDS (Rod)
<  o -Add ALTER SEQUENCE to modify min/max/increment/cache/cycle values
237d204
<  o -Allow CLUSTER to cluster all tables (Alvaro Herrera)
248d214
<  o -MOVE 0 should not move to end of cursor (Bruce)
252d217
<  o -Allow cursors outside transactions
264,265d228
<  o -Allow EXPLAIN EXECUTE to see prepared plans
<   o -Allow SHOW of some non-modifiable variables, like pg_controldata
280d242
<  o -Allow array declarations and other data types in PL/PgSQL DECLARE
282,283d243
<  o -Make PL/PgSQL %TYPE schema-aware
<  o -Allow PL/PgSQL to support array element assignment (Joe)
294,298d253
< * -Allow psql to show transaction status if backend protocol changes made
< * -Add schema, cast, and conversion backslash commands to psql (Christopher)
< * -Allow pg_dump to dump a specific schema (Neil Conway)
< * -Allow psql to do table completion for SELECT * FROM schema_part and
<   table completion for SELECT * FROM schema_name.
300,302d254
< * -Allow SSL-enabled clients to turn off SSL transfers
< * -Modify pg_get_triggerdef() to take a boolean to pretty-print,
<   and use that as part of pg_dump along with psql
327d278
<  o -Add SQLSTATE
330d280
<  o -Make casts work in variable initializations
334,336d283
<  o -Allow multi-threaded use of SQLCA
<  o -Understand structure definitions outside a declare section
<  o -Allow :var[:index] or :var[<integer>] as cvariable for an array var
356d302
< * -Support statement-level triggers (Neil)
460,461d405
< * -Add checkpoint_min_warning postgresql.conf option to warn about checkpoints
<   that are too frequent (Bruce)
472,476d415
< * -Improve Subplan list handling
< * -Allow Subplans to use efficient joins(hash, merge) with upper variable
< * -Add hash for evaluating GROUP BY aggregates (Tom)
< * -Allow merge and hash joins on expressions not just simple variables (Tom)
< * -Make IN/NOT IN have similar performance to EXISTS/NOT EXISTS (Tom)
480d418
< * -Inline simple SQL functions to avoid overhead (Tom)
495d432
< * -Get faster regex() code from Henry Spencer <henry@zoo.utoronto.ca>
511,512d447
< * -Modify regression tests to prevent failures do to minor numeric rounding
< * -Add OpenBSD's getpeereid() call for local socket authentication
537,542d471
<  o -Show transaction status in psql
<  o -Allow binding of query parameters, support for prepared queries
<  o -Remove hard-coded limits on user/db/password names
<  o -Remove unused elements of startup packet (unused, tty, passlength)
<  o -Fix COPY/fastpath protocol
<  o -Error codes
544d472
<  o -Special passing of binary values in platform-neutral format (bytea?)
547d474
<  o -Report server version number, database encoding, client encoding

20 years agoSERIAL no longer creates an index by default, as of 7.3.
Bruce Momjian [Wed, 19 Nov 2003 16:50:48 +0000 (16:50 +0000)]
SERIAL no longer creates an index by default, as of 7.3.

20 years agoBackground writer process
Jan Wieck [Wed, 19 Nov 2003 15:55:08 +0000 (15:55 +0000)]
Background writer process

This first part of the background writer does no syncing at all.
It's only purpose is to keep the LRU heads clean so that regular
backends seldom to never have to call write().

Jan

20 years agoFixed typo in create schema parsing.
Michael Meskes [Wed, 19 Nov 2003 13:18:13 +0000 (13:18 +0000)]
Fixed typo in create schema parsing.

20 years agoAdd FreeBSD mention for initdb.c.
Bruce Momjian [Mon, 17 Nov 2003 20:35:28 +0000 (20:35 +0000)]
Add FreeBSD mention for initdb.c.

20 years agoOptimize. Improve ispell support for compound words. This work was sponsored by ABC...
Teodor Sigaev [Mon, 17 Nov 2003 17:34:35 +0000 (17:34 +0000)]
Optimize. Improve ispell support for compound words. This work was sponsored by ABC Startsiden AS.

20 years agoChange "IN/NOT IN" to "IN / NOT IN" for clarity, both branches.
Bruce Momjian [Mon, 17 Nov 2003 04:50:56 +0000 (04:50 +0000)]
Change "IN/NOT IN" to "IN / NOT IN" for clarity, both branches.

20 years agoAdd:
Bruce Momjian [Mon, 17 Nov 2003 02:05:26 +0000 (02:05 +0000)]
Add:

> * Add GUC setting to make created tables default to WITHOUT OIDS

20 years agoFix double dash:
Bruce Momjian [Mon, 17 Nov 2003 02:01:56 +0000 (02:01 +0000)]
Fix double dash:

<  o --Add ALTER TABLE tab SET WITHOUT OIDS (Rod)
>  o -Add ALTER TABLE tab SET WITHOUT OIDS (Rod)

20 years agoFix datetime input parsing to accept YYYY-MONTHNAME-DD and related syntaxes,
Tom Lane [Sun, 16 Nov 2003 20:29:16 +0000 (20:29 +0000)]
Fix datetime input parsing to accept YYYY-MONTHNAME-DD and related syntaxes,
which had been unintentionally broken by recent changes to tighten up the
DateStyle rules for all-numeric date input.  Add documentation and
regression tests for this, too.

20 years agoSync HEAD with update committed only in 7.4 branch.
Tom Lane [Sun, 16 Nov 2003 17:16:47 +0000 (17:16 +0000)]
Sync HEAD with update committed only in 7.4 branch.

20 years agoAnother try at fixing the Linux overcommit docs, per Andrew Dunstan.
Tom Lane [Sun, 16 Nov 2003 17:01:50 +0000 (17:01 +0000)]
Another try at fixing the Linux overcommit docs, per Andrew Dunstan.

20 years agoChanged parameter name for shared cache status report interval to
Jan Wieck [Sun, 16 Nov 2003 16:41:01 +0000 (16:41 +0000)]
Changed parameter name for shared cache status report interval to

    debug_shared_buffers = <seconds>

as per previous discussion.

Jan

20 years agoUpdate HISTORY
Bruce Momjian [Sun, 16 Nov 2003 04:49:35 +0000 (04:49 +0000)]
Update HISTORY

20 years agopgadmin III released.
Bruce Momjian [Sun, 16 Nov 2003 04:41:12 +0000 (04:41 +0000)]
pgadmin III released.

20 years agoUpdate info about SysV configuration procedures on OS X 10.3.
Tom Lane [Sun, 16 Nov 2003 03:58:10 +0000 (03:58 +0000)]
Update info about SysV configuration procedures on OS X 10.3.

20 years agoSome editorializing on the Linux memory overcommit docs, from Neil Conway.
Tom Lane [Sun, 16 Nov 2003 03:38:34 +0000 (03:38 +0000)]
Some editorializing on the Linux memory overcommit docs, from Neil Conway.

20 years agoMore release-notes copy editing, from Neil Conway.
Tom Lane [Sun, 16 Nov 2003 03:23:48 +0000 (03:23 +0000)]
More release-notes copy editing, from Neil Conway.

20 years agoA little bit more copy-editing...
Tom Lane [Sat, 15 Nov 2003 20:01:11 +0000 (20:01 +0000)]
A little bit more copy-editing...

20 years agoMinor wording improvement suggested by a comment in the interactive docs.
Tom Lane [Sat, 15 Nov 2003 19:46:31 +0000 (19:46 +0000)]
Minor wording improvement suggested by a comment in the interactive docs.

20 years agoMake creation of statistics collection socket more robust, by allowing it
Tom Lane [Sat, 15 Nov 2003 17:24:07 +0000 (17:24 +0000)]
Make creation of statistics collection socket more robust, by allowing it
to try additional addresses returned from getaddrinfo() if the first one
fails at the bind() or connect() steps.  Per yesterday's discussion.

20 years agoCorrect names of people with funny characters.
Peter Eisentraut [Sat, 15 Nov 2003 01:16:37 +0000 (01:16 +0000)]
Correct names of people with funny characters.

20 years agoTranslation updates
Peter Eisentraut [Fri, 14 Nov 2003 23:59:12 +0000 (23:59 +0000)]
Translation updates

20 years agoRepair error in syntax documentation for CREATE TABLE: one does not put
Tom Lane [Fri, 14 Nov 2003 22:56:11 +0000 (22:56 +0000)]
Repair error in syntax documentation for CREATE TABLE: one does not put
a comma between multiple column_constraint's for a single column.
Per report from Tomislaw Kitynski.

20 years agoTry to improve error handling for failures of backend subprocess.
Tom Lane [Fri, 14 Nov 2003 18:32:34 +0000 (18:32 +0000)]
Try to improve error handling for failures of backend subprocess.

20 years agoAdd CHECK_FOR_INTERRUPTS() to bootstrap command loop, so that control-C
Tom Lane [Fri, 14 Nov 2003 18:19:45 +0000 (18:19 +0000)]
Add CHECK_FOR_INTERRUPTS() to bootstrap command loop, so that control-C
can terminate the bootstrap run.

20 years agoAdd release date for 7.4.
Bruce Momjian [Fri, 14 Nov 2003 17:47:04 +0000 (17:47 +0000)]
Add release date for 7.4.

20 years agoAdd fflush() before popen() calls; avoids any possible problem with
Tom Lane [Fri, 14 Nov 2003 17:30:41 +0000 (17:30 +0000)]
Add fflush() before popen() calls; avoids any possible problem with
double or out-of-sequence output with child process.

20 years agoAdd missing logic to handle fixing permissions on an already-existing
Tom Lane [Fri, 14 Nov 2003 17:19:35 +0000 (17:19 +0000)]
Add missing logic to handle fixing permissions on an already-existing
data directory.  Also fix handling of error conditions associated with
data directory checking step (can't use a boolean to distinguish four
possible result states...)

20 years agoFix misstatement about the required value of SEMVMX. Given our current
Tom Lane [Fri, 14 Nov 2003 15:43:22 +0000 (15:43 +0000)]
Fix misstatement about the required value of SEMVMX.  Given our current
usage of PGSemaMagic, the minimum required value is actually 537, but
say 1000 just to use a round number.  255 is *not* good enough.

20 years agoAdded documentation for the new interface between the buffer manager
Jan Wieck [Fri, 14 Nov 2003 04:32:11 +0000 (04:32 +0000)]
Added documentation for the new interface between the buffer manager
and the cache replacement strategy as well as a description of the
ARC algorithm and the special tailoring of that done for PostgreSQL.

Jan

20 years agoUpdate HISTORY to match release.sgml.
Bruce Momjian [Fri, 14 Nov 2003 03:41:11 +0000 (03:41 +0000)]
Update HISTORY to match release.sgml.

20 years agoUpdate bug template to 7.4 --- that was the only one missing.
Bruce Momjian [Fri, 14 Nov 2003 03:30:54 +0000 (03:30 +0000)]
Update bug template to 7.4 --- that was the only one missing.

20 years agoTranslation updates
Peter Eisentraut [Fri, 14 Nov 2003 02:08:17 +0000 (02:08 +0000)]
Translation updates