]> granicus.if.org Git - postgresql/log
postgresql
19 years agoAdd mention of compatibility problem with turning off backslash escapes:
Bruce Momjian [Thu, 17 Mar 2005 04:42:15 +0000 (04:42 +0000)]
Add mention of compatibility problem with turning off backslash escapes:

<   SQL-spec compliant, so allow such handling to be disabled.
>   SQL-spec compliant, so allow such handling to be disabled.  However,
>   disabling backslashes could break many third-party applications and tools.

19 years agoFactor duplicate snprintf code into functions.
Bruce Momjian [Thu, 17 Mar 2005 03:18:14 +0000 (03:18 +0000)]
Factor duplicate snprintf code into functions.

19 years agoAdd a reference to the documentation on alternate index operator classes in
Neil Conway [Thu, 17 Mar 2005 00:22:24 +0000 (00:22 +0000)]
Add a reference to the documentation on alternate index operator classes in
the locale docs. Patch from Chris KL, editorialization by Neil Conway.

19 years agoThis patch makes \d on tables and views sort fk constraints, triggers
Neil Conway [Wed, 16 Mar 2005 23:52:18 +0000 (23:52 +0000)]
This patch makes \d on tables and views sort fk constraints, triggers
and rules alphabetically in the output.  This makes it the same as
for indexes and stops the irritating random or reverse ordering it
currently has.

Chris KL

19 years agoForce initdb cause of encoding additions.
Bruce Momjian [Wed, 16 Mar 2005 22:59:42 +0000 (22:59 +0000)]
Force initdb cause of encoding additions.

19 years agoRevise TupleTableSlot code to avoid unnecessary construction and disassembly
Tom Lane [Wed, 16 Mar 2005 21:38:10 +0000 (21:38 +0000)]
Revise TupleTableSlot code to avoid unnecessary construction and disassembly
of tuples when passing data up through multiple plan nodes.  A slot can now
hold either a normal "physical" HeapTuple, or a "virtual" tuple consisting
of Datum/isnull arrays.  Upper plan levels can usually just copy the Datum
arrays, avoiding heap_formtuple() and possible subsequent nocachegetattr()
calls to extract the data again.  This work extends Atsushi Ogawa's earlier
patch, which provided the key idea of adding Datum arrays to TupleTableSlots.
(I believe however that something like this was foreseen way back in Berkeley
days --- see the old comment on ExecProject.)  A test case involving many
levels of join of fairly wide tables (about 80 columns altogether) showed
about 3x overall speedup, though simple queries will probably not be
helped very much.

I have also duplicated some code in heaptuple.c in order to provide versions
of heap_formtuple and friends that use "bool" arrays to indicate null
attributes, instead of the old convention of "char" arrays containing either
'n' or ' '.  This provides a better match to the convention used by
ExecEvalExpr.  While I have not made a concerted effort to get rid of uses
of the old routines, I think they should be deprecated and eventually removed.

19 years agoAdd sprintf support, that were were missing.
Bruce Momjian [Wed, 16 Mar 2005 21:27:23 +0000 (21:27 +0000)]
Add sprintf support, that were were missing.

Add support for snprintf '+', 'h', and %* length settings.

19 years agopgindent snprintf.c for consistency.
Bruce Momjian [Wed, 16 Mar 2005 15:12:18 +0000 (15:12 +0000)]
pgindent snprintf.c for consistency.

19 years agoFix snprintf for %*$.
Bruce Momjian [Wed, 16 Mar 2005 15:11:43 +0000 (15:11 +0000)]
Fix snprintf for %*$.

19 years agoAdd CVS \r\n regression tests.
Bruce Momjian [Wed, 16 Mar 2005 06:05:53 +0000 (06:05 +0000)]
Add CVS \r\n regression tests.

Andrew Dunstan

19 years agoFix snprintf to handle %$ properly by storing and reordering the
Bruce Momjian [Wed, 16 Mar 2005 06:00:58 +0000 (06:00 +0000)]
Fix snprintf to handle %$ properly by storing and reordering the
arguments.

Nicolai Tufar

19 years agoAdd missing include for new lc_ctype_is_c() function.
Bruce Momjian [Wed, 16 Mar 2005 01:49:10 +0000 (01:49 +0000)]
Add missing include for new lc_ctype_is_c() function.

Per Neil.

19 years agoPrevent locale-aware handling of upper, lower, and initcap when the
Bruce Momjian [Wed, 16 Mar 2005 00:02:49 +0000 (00:02 +0000)]
Prevent locale-aware handling of upper, lower, and initcap when the
locale is C.

Backpatch to 8.0.X because some operating systems were throwing errors
for such operations, rather than ignoring the locale when it was C.

19 years agoWrap the implementation of fork_process() inside #ifndef WIN32 -- this
Neil Conway [Wed, 16 Mar 2005 00:02:39 +0000 (00:02 +0000)]
Wrap the implementation of fork_process() inside #ifndef WIN32 -- this
should hopefully unbreak the Win32 build. Apologies for breaking it in
the first place.

19 years agoUpdate to 8.0.1.
Bruce Momjian [Tue, 15 Mar 2005 22:38:23 +0000 (22:38 +0000)]
Update to 8.0.1.

19 years agoAdd:
Bruce Momjian [Tue, 15 Mar 2005 05:31:27 +0000 (05:31 +0000)]
Add:

> * Allow sequential scans to take advantage of other concurrent
>   sequentiqal scans, also called "Synchronised Scanning"

19 years agoUpdate Win32 item:
Bruce Momjian [Tue, 15 Mar 2005 04:01:21 +0000 (04:01 +0000)]
Update Win32 item:

<         o Disallow encodings like UTF8 which PostgreSQL supports
<           but the operating system does not (already disallowed by
<    pginstaller)
>         o Add support for Unicode
<    To fix UTF8, the data needs to be converted to UTF16 and then
<    the Win32 wcscoll() can be used, and perhaps other functions
>    To fix this, the data needs to be converted to/from UTF16/UTF8
>           so the Win32 wcscoll() can be used, and perhaps other functions
<    locales but provides no ordering.
<
>    locales but provides no ordering or character set classes.

19 years agoClean up win1252 documentation. Mention how we determine the number of
Bruce Momjian [Tue, 15 Mar 2005 02:30:33 +0000 (02:30 +0000)]
Clean up win1252 documentation.  Mention how we determine the number of
bytes/character for each encoding.

19 years agoIssue free space notices to both the user and the server log file.
Bruce Momjian [Mon, 14 Mar 2005 20:15:09 +0000 (20:15 +0000)]
Issue free space notices to both the user and the server log file.

19 years agoMake pg_dump emit a useful error message, instead of just dumping core,
Tom Lane [Mon, 14 Mar 2005 18:57:33 +0000 (18:57 +0000)]
Make pg_dump emit a useful error message, instead of just dumping core,
if it finds a pg_rewrite entry for which there is no pg_class entry.
Per report from Andrew Slobodyanyk.

19 years agoAdd support for Win1252 encoding.
Bruce Momjian [Mon, 14 Mar 2005 18:31:25 +0000 (18:31 +0000)]
Add support for Win1252 encoding.

Roland Volkmann

19 years agoBump minor version numbers for 8.1 compared to 8.0.
Bruce Momjian [Mon, 14 Mar 2005 17:27:50 +0000 (17:27 +0000)]
Bump minor version numbers for 8.1 compared to 8.0.

19 years agoCompleted:
Bruce Momjian [Mon, 14 Mar 2005 16:37:02 +0000 (16:37 +0000)]
Completed:

> * -Cache last known per-tuple offsets to speed long tuple access

19 years agoFix mistakes in SGML markup. From David Fetter.
Neil Conway [Mon, 14 Mar 2005 06:49:48 +0000 (06:49 +0000)]
Fix mistakes in SGML markup. From David Fetter.

19 years agoNote that the -F and -R command line options only affect unaligned
Tom Lane [Mon, 14 Mar 2005 06:19:01 +0000 (06:19 +0000)]
Note that the -F and -R command line options only affect unaligned
output mode.  This was already stated in other places in the psql
reference page, but not here.

19 years agoAvoid O(N^2) overhead in repeated nocachegetattr calls when columns of
Tom Lane [Mon, 14 Mar 2005 04:41:13 +0000 (04:41 +0000)]
Avoid O(N^2) overhead in repeated nocachegetattr calls when columns of
a tuple are being accessed via ExecEvalVar and the attcacheoff shortcut
isn't usable (due to nulls and/or varlena columns).  To do this, cache
Datums extracted from a tuple in the associated TupleTableSlot.
Also some code cleanup in and around the TupleTable handling.
Atsushi Ogawa with some kibitzing by Tom Lane.

19 years agoDocument client-only encodings.
Bruce Momjian [Mon, 14 Mar 2005 03:59:22 +0000 (03:59 +0000)]
Document client-only encodings.

19 years agoFix typo on URL.
Bruce Momjian [Mon, 14 Mar 2005 03:07:25 +0000 (03:07 +0000)]
Fix typo on URL.

19 years agoAdd user:
Bruce Momjian [Mon, 14 Mar 2005 03:00:26 +0000 (03:00 +0000)]
Add user:

> * Support triggers on columns (Greg Sabino Mullane)

19 years agoFinalize character set documentation changes.
Bruce Momjian [Mon, 14 Mar 2005 02:14:42 +0000 (02:14 +0000)]
Finalize character set documentation changes.

19 years agoAllow ALTER FUNCTION to change a function's strictness, volatility, and
Neil Conway [Mon, 14 Mar 2005 00:19:37 +0000 (00:19 +0000)]
Allow ALTER FUNCTION to change a function's strictness, volatility, and
whether or not it is a security definer. Changing a function's strictness
is required by SQL2003, and the other capabilities make sense. Also, allow
an optional RESTRICT noise word to be specified, for SQL conformance.

Some trivial regression tests added and the documentation has been
updated.

19 years agoUpdate comments for new encoding names.
Bruce Momjian [Mon, 14 Mar 2005 00:19:13 +0000 (00:19 +0000)]
Update comments for new encoding names.

19 years agoAdd missing identification comment, remove entirely inappropriate include
Tom Lane [Sun, 13 Mar 2005 23:32:26 +0000 (23:32 +0000)]
Add missing identification comment, remove entirely inappropriate include
of postgres.h.

19 years agoAdd some missing #includes.
Tom Lane [Sun, 13 Mar 2005 23:27:38 +0000 (23:27 +0000)]
Add some missing #includes.

19 years agoIncrement all major version numbers in 8.0.X to force recompile of
Bruce Momjian [Sun, 13 Mar 2005 22:04:29 +0000 (22:04 +0000)]
Increment all major version numbers in 8.0.X to force recompile of
client aplications so 7.4.X releases can be installed on the same
machine as 8.0.X.

19 years agoForgot that I had intended to replace division by masking in hash calculation.
Tom Lane [Sun, 13 Mar 2005 19:59:40 +0000 (19:59 +0000)]
Forgot that I had intended to replace division by masking in hash calculation.

19 years agoRemove CENTER tag.
Bruce Momjian [Sun, 13 Mar 2005 19:27:53 +0000 (19:27 +0000)]
Remove CENTER tag.

19 years agoMake default_with_oids default to false -- user-created tables will now
Neil Conway [Sun, 13 Mar 2005 09:36:31 +0000 (09:36 +0000)]
Make default_with_oids default to false -- user-created tables will now
no longer include OIDs, unless WITH OIDS is specified or the
default_with_oids configuration parameter is enabled. Update the docs
accordingly.

19 years agoMore ordering adjustments.
Bruce Momjian [Sun, 13 Mar 2005 05:31:04 +0000 (05:31 +0000)]
More ordering adjustments.

19 years agoUpdate obsolete comment.
Neil Conway [Sun, 13 Mar 2005 05:19:26 +0000 (05:19 +0000)]
Update obsolete comment.

19 years agoFix markup.
Bruce Momjian [Sun, 13 Mar 2005 05:16:33 +0000 (05:16 +0000)]
Fix markup.

19 years agoMore markup changes.
Bruce Momjian [Sun, 13 Mar 2005 05:11:49 +0000 (05:11 +0000)]
More markup changes.

19 years agoMore cleanups.
Bruce Momjian [Sun, 13 Mar 2005 04:35:06 +0000 (04:35 +0000)]
More cleanups.

19 years agoMore improvements.
Bruce Momjian [Sun, 13 Mar 2005 04:10:23 +0000 (04:10 +0000)]
More improvements.

19 years agoMore additions to the table.
Bruce Momjian [Sun, 13 Mar 2005 03:44:51 +0000 (03:44 +0000)]
More additions to the table.

19 years agoKeep changing the markup until I like it. :-)
Bruce Momjian [Sun, 13 Mar 2005 03:02:08 +0000 (03:02 +0000)]
Keep changing the markup until I like it.  :-)

19 years agoMore table markup improvements.
Bruce Momjian [Sun, 13 Mar 2005 02:54:34 +0000 (02:54 +0000)]
More table markup improvements.

19 years agoMore table markup fixes.
Bruce Momjian [Sun, 13 Mar 2005 02:33:03 +0000 (02:33 +0000)]
More table markup fixes.

19 years agoRework "aliases" column for encodings.
Bruce Momjian [Sun, 13 Mar 2005 02:20:50 +0000 (02:20 +0000)]
Rework "aliases" column for encodings.

19 years agoFix markup typo.
Bruce Momjian [Sun, 13 Mar 2005 02:07:04 +0000 (02:07 +0000)]
Fix markup typo.

19 years agoAdd missing conversion documentation for certain encodings.
Bruce Momjian [Sun, 13 Mar 2005 02:02:44 +0000 (02:02 +0000)]
Add missing conversion documentation for certain encodings.

19 years agoReorder documented encodings to be alphabetical.
Bruce Momjian [Sun, 13 Mar 2005 01:30:59 +0000 (01:30 +0000)]
Reorder documented encodings to be alphabetical.

Remove warning about pre-7.2 LATIN5 usage.

19 years agoDocument aliases for our supported encodings.
Bruce Momjian [Sun, 13 Mar 2005 01:26:30 +0000 (01:26 +0000)]
Document aliases for our supported encodings.

Add a few encodings that were not documented.

19 years agoWhen cloning template0 (or other fully-frozen databases), set the new
Tom Lane [Sat, 12 Mar 2005 21:33:55 +0000 (21:33 +0000)]
When cloning template0 (or other fully-frozen databases), set the new
database's datallowconn and datfrozenxid to the current transaction ID
instead of copying the source database's values.  This is OK because we
assume the source DB contains no normal transaction IDs whatsoever.
This keeps VACUUM from immediately starting to complain about unvacuumed
databases in the situation where we are more than 2 billion transactions
out from the XID stamp of template0.  Per discussion with Milen Radev
(although his complaint turned out to be due to something else, but the
problem is real anyway).

19 years agoFix ALTER DATABASE RENAME to allow the operation if user is a superuser
Tom Lane [Sat, 12 Mar 2005 21:11:50 +0000 (21:11 +0000)]
Fix ALTER DATABASE RENAME to allow the operation if user is a superuser
who for some reason isn't marked usecreatedb.  Per report from Alexander
Pravking.  Also fix sloppy coding in have_createdb_privilege().

19 years agoAdjust the API for aggregate function calls so that a C-coded function
Tom Lane [Sat, 12 Mar 2005 20:25:06 +0000 (20:25 +0000)]
Adjust the API for aggregate function calls so that a C-coded function
can tell whether it is being used as an aggregate or not.  This allows
such a function to avoid re-pallocing a pass-by-reference transition
value; normally it would be unsafe for a function to scribble on an input,
but in the aggregate case it's safe to reuse the old transition value.
Make int8inc() do this.  This gets a useful improvement in the speed of
COUNT(*), at least on narrow tables (it seems to be swamped by I/O when
the table rows are wide).  Per a discussion in early December with
Neil Conway.  I also fixed int_aggregate.c to check this, thereby
turning it into something approaching a supportable technique instead
of being a crude hack.

19 years agoAdd pg_buffercache contrib module, from Mark Kirkwood.
Neil Conway [Sat, 12 Mar 2005 15:36:24 +0000 (15:36 +0000)]
Add pg_buffercache contrib module, from Mark Kirkwood.

19 years agoSome builds (depends on crypto engine support?) of OpenSSL
Neil Conway [Sat, 12 Mar 2005 06:53:54 +0000 (06:53 +0000)]
Some builds (depends on crypto engine support?) of OpenSSL
0.9.7x have EVP_DigestFinal function which which clears all of
EVP_MD_CTX.  This makes pgcrypto crash in functions which
re-use one digest context several times: hmac() and crypt()
with md5 algorithm.

Following patch fixes it by carring the digest info around
EVP_DigestFinal and re-initializing cipher.

Marko Kreen.

19 years agoDocumention all our supported encodings.
Bruce Momjian [Sat, 12 Mar 2005 06:28:17 +0000 (06:28 +0000)]
Documention all our supported encodings.

19 years agoDone:
Bruce Momjian [Sat, 12 Mar 2005 05:41:38 +0000 (05:41 +0000)]
Done:

<  o Allow COPY FROM ... CSV to interpret newlines and carriage
>  o -Allow COPY FROM ... CSV to interpret newlines and carriage

19 years agoHandle carriage returns and line feeds in COPY CSV mode.
Bruce Momjian [Sat, 12 Mar 2005 05:41:34 +0000 (05:41 +0000)]
Handle carriage returns and line feeds in COPY CSV mode.

Andrew Dunstan

19 years agoDone:
Bruce Momjian [Sat, 12 Mar 2005 05:25:49 +0000 (05:25 +0000)]
Done:

* -Add a warning when the free space map is too small

19 years agoAdd warning about the need to increase "max_fsm_relations" and
Bruce Momjian [Sat, 12 Mar 2005 05:21:52 +0000 (05:21 +0000)]
Add warning about the need to increase "max_fsm_relations" and
"max_fsm_relations" for vacuums.  Also improve VACUUM VERBOSE final
message text.

Ron Mayer

19 years agoFix snprintf() to properly handle precision specification for %f.
Bruce Momjian [Sat, 12 Mar 2005 04:00:56 +0000 (04:00 +0000)]
Fix snprintf() to properly handle precision specification for %f.

19 years agoFix problem with infinite recursion between write_syslogger_file and
Tom Lane [Sat, 12 Mar 2005 01:54:44 +0000 (01:54 +0000)]
Fix problem with infinite recursion between write_syslogger_file and
elog if the former has trouble writing its file.  Code review for
Magnus' patch to redirect stderr to syslog on Windows (Bruce's version
seems right, but did some minor prettification).

Backpatch both changes to 8.0 branch.

19 years agoHere's the patch to fix a lot of markup errors in the HTML FAQs. Doesn't
Bruce Momjian [Fri, 11 Mar 2005 21:46:54 +0000 (21:46 +0000)]
Here's the patch to fix a lot of markup errors in the HTML FAQs. Doesn't
change content (at least not supposed to).

Magnus Hagander

19 years agoAdd fprintf() custom version to libpgport.
Bruce Momjian [Fri, 11 Mar 2005 19:13:43 +0000 (19:13 +0000)]
Add fprintf() custom version to libpgport.

Document use of macros for pg_printf functions.

Bump major versions of all interfaces to handle movement of get_progname
from libpq to libpgport in 8.0, and probably other libpgport changes in 8.1.

19 years agoDefine snprintf() to call pg_snprintf() so our own snprintf-like
Bruce Momjian [Fri, 11 Mar 2005 17:20:35 +0000 (17:20 +0000)]
Define snprintf() to call pg_snprintf() so our own snprintf-like
implementation doesn't export out via libpq and get used by a user
application.

19 years agoAdd 'static' to initdb.c file-global variables.
Bruce Momjian [Fri, 11 Mar 2005 15:36:27 +0000 (15:36 +0000)]
Add 'static' to initdb.c file-global variables.

19 years agoMark file-global function and variables as static.
Bruce Momjian [Fri, 11 Mar 2005 15:25:31 +0000 (15:25 +0000)]
Mark file-global function and variables as static.

19 years agoFix typos.
Bruce Momjian [Fri, 11 Mar 2005 13:09:39 +0000 (13:09 +0000)]
Fix typos.

Hashem Masoud

19 years agoFix typos.
Bruce Momjian [Fri, 11 Mar 2005 11:59:16 +0000 (11:59 +0000)]
Fix typos.

Robert Treat

19 years agoFix typos.
Bruce Momjian [Fri, 11 Mar 2005 11:43:45 +0000 (11:43 +0000)]
Fix typos.

Robert Treat

19 years agoFix typo.
Bruce Momjian [Fri, 11 Mar 2005 11:42:03 +0000 (11:42 +0000)]
Fix typo.

19 years agoSlight refactoring and optimization of some code in WaitOnLock().
Neil Conway [Fri, 11 Mar 2005 03:52:06 +0000 (03:52 +0000)]
Slight refactoring and optimization of some code in WaitOnLock().

19 years agoAdd:
Bruce Momjian [Fri, 11 Mar 2005 03:45:25 +0000 (03:45 +0000)]
Add:

> * Fix cross-compiling of time zone database via 'zic'

19 years agoMake the behavior of HAVING without GROUP BY conform to the SQL spec.
Tom Lane [Thu, 10 Mar 2005 23:21:26 +0000 (23:21 +0000)]
Make the behavior of HAVING without GROUP BY conform to the SQL spec.
Formerly, if such a clause contained no aggregate functions we mistakenly
treated it as equivalent to WHERE.  Per spec it must cause the query to
be treated as a grouped query of a single group, the same as appearance
of aggregate functions would do.  Also, the HAVING filter must execute
after aggregate function computation even if it itself contains no
aggregate functions.

19 years agoAdd spinlock support for Itanium processor with Intel compiler.
Bruce Momjian [Thu, 10 Mar 2005 21:41:01 +0000 (21:41 +0000)]
Add spinlock support for Itanium processor with Intel compiler.

Vikram Kalsi

19 years agoRefactor fork()-related code. We need to do various housekeeping tasks
Neil Conway [Thu, 10 Mar 2005 07:14:03 +0000 (07:14 +0000)]
Refactor fork()-related code. We need to do various housekeeping tasks
before we can invoke fork() -- flush stdio buffers, save and restore the
profiling timer on Linux with LINUX_PROFILE, and handle BeOS stuff. This
patch moves that code into a single function, fork_process(), instead of
duplicating it at the various callsites of fork().

This patch doesn't address the EXEC_BACKEND case; there is room for
further cleanup there.

19 years agoAdd comma.
Bruce Momjian [Tue, 8 Mar 2005 13:27:31 +0000 (13:27 +0000)]
Add comma.

19 years agoFix markup of URL.
Bruce Momjian [Tue, 8 Mar 2005 03:52:57 +0000 (03:52 +0000)]
Fix markup of URL.

19 years agoAdd URL for patches queues.
Bruce Momjian [Tue, 8 Mar 2005 01:15:33 +0000 (01:15 +0000)]
Add URL for patches queues.

19 years agoUnbreak out-of-tree builds, by fixing a typo.
Neil Conway [Mon, 7 Mar 2005 23:18:06 +0000 (23:18 +0000)]
Unbreak out-of-tree builds, by fixing a typo.

19 years agoReplace an instance of $Id$ and an instance of $Header$ with $PostgreSQL$
Neil Conway [Mon, 7 Mar 2005 22:59:43 +0000 (22:59 +0000)]
Replace an instance of $Id$ and an instance of $Header$ with $PostgreSQL$

19 years agoProperly implement "Response files" for bcc. Add URL's to describe the
Bruce Momjian [Mon, 7 Mar 2005 21:10:58 +0000 (21:10 +0000)]
Properly implement "Response files" for bcc.  Add URL's to describe the
feature for Win32 and bcc.

19 years agoAdjust creation/destruction of TupleDesc data structure to reduce the
Tom Lane [Mon, 7 Mar 2005 04:42:17 +0000 (04:42 +0000)]
Adjust creation/destruction of TupleDesc data structure to reduce the
number of palloc calls.  This has a salutory impact on plpgsql operations
with record variables (which create and destroy tupdescs constantly)
and probably helps a bit in some other cases too.

19 years agoRename canonical encodings, per Peter:
Bruce Momjian [Mon, 7 Mar 2005 04:30:55 +0000 (04:30 +0000)]
Rename canonical encodings, per Peter:

UNICODE => UTF8
ALT => WIN866
WIN => WIN1251
TCVN => WIN1258

The old codes continue to work.

19 years agoHere's a tiny fix for a harmless typo in catalog.c:
Neil Conway [Mon, 7 Mar 2005 04:15:34 +0000 (04:15 +0000)]
Here's a tiny fix for a harmless typo in catalog.c:

Too much space is allocated for tablespace file path, I guess the
directory name used to be "pg_tablespaces" instead of "pg_tblspc" at
some point.

Heikki Linnakangas

19 years agoAdd a note that regression test row ordering differences can be caused
Tom Lane [Mon, 7 Mar 2005 02:00:28 +0000 (02:00 +0000)]
Add a note that regression test row ordering differences can be caused
by non-default planner parameter settings.

19 years agoRevise hash join code so that we can increase the number of batches
Tom Lane [Sun, 6 Mar 2005 22:15:05 +0000 (22:15 +0000)]
Revise hash join code so that we can increase the number of batches
on-the-fly, and thereby avoid blowing out memory when the planner has
underestimated the hash table size.  Hash join will now obey the
work_mem limit with some faithfulness.  Per my recent proposal
(hash aggregate part isn't done yet though).

19 years agoPrevent rank change in case of duplicate search terms
Teodor Sigaev [Sat, 5 Mar 2005 15:48:32 +0000 (15:48 +0000)]
Prevent rank change in case of duplicate search terms

19 years agoReplace the BufMgrLock with separate locks on the lookup hashtable and
Tom Lane [Fri, 4 Mar 2005 20:21:07 +0000 (20:21 +0000)]
Replace the BufMgrLock with separate locks on the lookup hashtable and
the freelist, plus per-buffer spinlocks that protect access to individual
shared buffer headers.  This requires abandoning a global freelist (since
the freelist is a global contention point), which shoots down ARC and 2Q
as well as plain LRU management.  Adopt a clock sweep algorithm instead.
Preliminary results show substantial improvement in multi-backend situations.

19 years agoMove snprintf int64 compatibility letters into a NOT_USED block.
Bruce Momjian [Wed, 2 Mar 2005 23:56:53 +0000 (23:56 +0000)]
Move snprintf int64 compatibility letters into a NOT_USED block.

19 years agoFix misspelling of sysctl, per Stefan Hans.
Tom Lane [Wed, 2 Mar 2005 19:58:54 +0000 (19:58 +0000)]
Fix misspelling of sysctl, per Stefan Hans.

19 years agoUse test && test rather than test -a, per Peter.
Bruce Momjian [Wed, 2 Mar 2005 15:42:35 +0000 (15:42 +0000)]
Use test && test rather than test -a, per Peter.

19 years agoFix for %I64d snprintf.
Bruce Momjian [Wed, 2 Mar 2005 15:32:39 +0000 (15:32 +0000)]
Fix for %I64d snprintf.

19 years agosnprintf() %I64d code fix.
Bruce Momjian [Wed, 2 Mar 2005 15:07:09 +0000 (15:07 +0000)]
snprintf() %I64d code fix.

19 years agoUse our own snprintf() only if NLS is enabled, and support %qd and %I64d.
Bruce Momjian [Wed, 2 Mar 2005 14:48:22 +0000 (14:48 +0000)]
Use our own snprintf() only if NLS is enabled, and support %qd and %I64d.

19 years agoPrevent large allocation in snprintf to hold positional parameters.
Bruce Momjian [Wed, 2 Mar 2005 05:22:22 +0000 (05:22 +0000)]
Prevent large allocation in snprintf to hold positional parameters.
Allocated size based on format string.

19 years agoAnother go at making pred_test() handle all reasonable combinations
Tom Lane [Wed, 2 Mar 2005 04:10:53 +0000 (04:10 +0000)]
Another go at making pred_test() handle all reasonable combinations
of AND and OR clauses.  The key point here is that an OR on the
predicate side has to be treated gingerly: we may be able to prove
that the OR is implied even when no one of its components is implied.
For example (x OR y) implies (x OR y OR z) even though no one of x,
y, or z can be individually proven.  This code handles both the
example shown recently by Sergey Koshcheyev and the one shown last
October by Dawid Kuroczko.