]> granicus.if.org Git - postgresql/log
postgresql
23 years agoI had a need to read such things as the backend locale and the catalog
Bruce Momjian [Fri, 23 Feb 2001 20:38:35 +0000 (20:38 +0000)]
I had a need to read such things as the backend locale and the catalog
version number from the current database, and couldn't find any existing
program to do that.

linda:~$ pg_controldata
Log file id:                          0
Log file segment:                     5
Last modified:                        Wed Feb  7 19:35:47 2001
Database block size:                  8192
Blocks per segment of large relation: 131072
Catalog version number:               200101061
LC_COLLATE:                           en_GB
LC_CTYPE:                             en_GB
Log archive directory:

Oliver Elphick                                Oliver.Elphick@lfix.co.uk

23 years agoAs long as we're fixing this space calculation, let's actually do it
Tom Lane [Fri, 23 Feb 2001 20:12:37 +0000 (20:12 +0000)]
As long as we're fixing this space calculation, let's actually do it
right.  We should MAXALIGN the individual items because we'll
allocate them individually, not as an array.

23 years agoUpdate TODO list.
Bruce Momjian [Fri, 23 Feb 2001 19:38:43 +0000 (19:38 +0000)]
Update TODO list.

23 years agoUpdate comments on locks.
Bruce Momjian [Fri, 23 Feb 2001 19:24:06 +0000 (19:24 +0000)]
Update comments on locks.

23 years agoUpdate TODO list.
Bruce Momjian [Fri, 23 Feb 2001 19:21:53 +0000 (19:21 +0000)]
Update TODO list.

23 years agoUpdate TODO list.
Bruce Momjian [Fri, 23 Feb 2001 19:14:52 +0000 (19:14 +0000)]
Update TODO list.

23 years agoBruce Momjian <pgman@candle.pha.pa.us> writes:
Bruce Momjian [Fri, 23 Feb 2001 18:28:46 +0000 (18:28 +0000)]
Bruce Momjian <pgman@candle.pha.pa.us> writes:
> Is there one LOCKMETHODCTL for every backend?  I thought there was only
> one of them.
>>
>> You're right, that line is erroneous; it should read
>>
>> size += MAX_LOCK_METHODS * MAXALIGN(sizeof(LOCKMETHODCTL));
>>
>> Not a significant error but it should be changed for clarity ...

23 years agoupdate to beta5 ... I think I ran cvs2cl right this time? :)
Marc G. Fournier [Fri, 23 Feb 2001 18:09:59 +0000 (18:09 +0000)]
update to beta5 ... I think I ran cvs2cl right this time? :)

23 years agoupdate to beta5
Marc G. Fournier [Fri, 23 Feb 2001 18:09:43 +0000 (18:09 +0000)]
update to beta5

23 years agoResolve a compile error.
Hiroshi Inoue [Fri, 23 Feb 2001 09:31:52 +0000 (09:31 +0000)]
Resolve a compile error.

23 years agoReindex of shared system indexes must be overwrite mode.
Hiroshi Inoue [Fri, 23 Feb 2001 09:26:14 +0000 (09:26 +0000)]
Reindex of shared system indexes must be overwrite mode.

23 years agoEnhanced UTF-8/SJIS mapping generator, contributed by
Tatsuo Ishii [Fri, 23 Feb 2001 08:44:33 +0000 (08:44 +0000)]
Enhanced UTF-8/SJIS mapping generator, contributed by
Eiji Tokuya" <e-tokuya@Mail.Sankyo-Unyu.co.jp>

23 years agoFreshen with a few recent platform reports:
Thomas G. Lockhart [Fri, 23 Feb 2001 02:17:27 +0000 (02:17 +0000)]
Freshen with a few recent platform reports:
 Linux/Alpha, Linux/Sparc, MacOS-X.

23 years agochinese-gb -> chinese_gb.
Bruce Momjian [Fri, 23 Feb 2001 01:21:58 +0000 (01:21 +0000)]
chinese-gb -> chinese_gb.

23 years agoRename to chinese-gb
Bruce Momjian [Fri, 23 Feb 2001 01:21:27 +0000 (01:21 +0000)]
Rename to chinese-gb

23 years agoMore comment cleanups.
Bruce Momjian [Thu, 22 Feb 2001 23:20:06 +0000 (23:20 +0000)]
More comment cleanups.

23 years agoMore comment improvements.
Bruce Momjian [Thu, 22 Feb 2001 23:02:33 +0000 (23:02 +0000)]
More comment improvements.

23 years agoClean up index/btree comments/macros, as approved.
Bruce Momjian [Thu, 22 Feb 2001 21:48:49 +0000 (21:48 +0000)]
Clean up index/btree comments/macros, as approved.

23 years agoChange /*---- commants to /* where appropriate. pgindent will tighten
Bruce Momjian [Thu, 22 Feb 2001 18:39:20 +0000 (18:39 +0000)]
Change /*---- commants to /* where appropriate.  pgindent will tighten
up the comments later.

23 years agoUpdate flowchart xfig file.
Bruce Momjian [Thu, 22 Feb 2001 18:09:20 +0000 (18:09 +0000)]
Update flowchart xfig file.

23 years ago The attachement is the Chinese (GB) patch for PgAccess, don't know
Bruce Momjian [Thu, 22 Feb 2001 15:33:14 +0000 (15:33 +0000)]
The attachement is the Chinese (GB) patch for PgAccess, don't know
if it's correct to post here.
It's simple to do the translation, And I've test in 7.0.2 & current CVS,
seems pretty good.
If anyone want this little thing, I'll very happy.
use it is very simple, just gunzip it and copy to
$PGDIR/share/pgaccess/lib/languages/ for current CVS version,
and $PGDIR/pgaccess/lib/languages/ for 7.0*
BTW: I havn't got the tools to translate it to BIG5 encoding, is there
anybody to to it?

He Weiping(Laser Henry)

23 years agoAvoid 'FATAL: out of free buffers: time to abort !" error
Hiroshi Inoue [Thu, 22 Feb 2001 08:59:40 +0000 (08:59 +0000)]
Avoid 'FATAL: out of free buffers: time to abort !" error
during WAL recovery.  Recovery failure is always serious.

23 years agoAdjust expected error message due to change of keyword case-folding.
Tom Lane [Thu, 22 Feb 2001 05:32:56 +0000 (05:32 +0000)]
Adjust expected error message due to change of keyword case-folding.

23 years agoAdd _REGEX_UTILS_H to avoid duplication.
Tatsuo Ishii [Thu, 22 Feb 2001 04:35:00 +0000 (04:35 +0000)]
Add _REGEX_UTILS_H to avoid duplication.

23 years agoUpdate mutibyte doc
Tatsuo Ishii [Thu, 22 Feb 2001 04:30:16 +0000 (04:30 +0000)]
Update mutibyte doc

23 years agoFix the bug report [ODBC] select from a table having more than 32 fields:
Hiroshi Inoue [Thu, 22 Feb 2001 03:39:30 +0000 (03:39 +0000)]
Fix the bug report [ODBC] select from a table having more than 32 fields:
reported by Matteo Cavalleri.
Great thanks to Tom for his accurate analysis.

23 years agoDocument timeofday(), mention start-of-transaction vs. current-time
Tom Lane [Wed, 21 Feb 2001 23:15:24 +0000 (23:15 +0000)]
Document timeofday(), mention start-of-transaction vs. current-time
semantics, a few other small improvements.

23 years agoCorrect bogus output formatting of timeofday() function. A number of
Tom Lane [Wed, 21 Feb 2001 22:03:00 +0000 (22:03 +0000)]
Correct bogus output formatting of timeofday() function.  A number of
microseconds < 100000 should be displayed as, eg, 13.000126, not 13.126.

23 years agoComment improvements.
Bruce Momjian [Wed, 21 Feb 2001 19:07:04 +0000 (19:07 +0000)]
Comment improvements.

23 years agoChange case-folding of keywords to conform to SQL99 and fix misbehavior
Tom Lane [Wed, 21 Feb 2001 18:53:47 +0000 (18:53 +0000)]
Change case-folding of keywords to conform to SQL99 and fix misbehavior
in Turkish locale.  Keywords are now checked under pure ASCII case-folding
rules ('A'-'Z'->'a'-'z' and nothing else).  However, once a word is
determined not to be a keyword, it will be case-folded under the current
locale, same as before.  See pghackers discussion 20-Feb-01.

23 years agoFix markup. (A <keyword> isn't what it is in SQL.) Add jungle of more
Peter Eisentraut [Wed, 21 Feb 2001 17:50:38 +0000 (17:50 +0000)]
Fix markup.  (A <keyword> isn't what it is in SQL.)  Add jungle of more
markup. ;-)

23 years agoSafe guard in case of imporper data is provided to elog()
Tatsuo Ishii [Wed, 21 Feb 2001 06:05:23 +0000 (06:05 +0000)]
Safe guard in case of imporper data is provided to elog()
in multi-byte build.

23 years agonew regression data for 7.1
Tatsuo Ishii [Wed, 21 Feb 2001 06:03:25 +0000 (06:03 +0000)]
new regression data for 7.1

23 years agochange destroydb -> dropdb
Tatsuo Ishii [Wed, 21 Feb 2001 06:02:04 +0000 (06:02 +0000)]
change destroydb -> dropdb

23 years agoAdd more details.
Peter Eisentraut [Tue, 20 Feb 2001 22:27:56 +0000 (22:27 +0000)]
Add more details.

23 years agoTurn on section numbering.
Peter Eisentraut [Tue, 20 Feb 2001 22:27:26 +0000 (22:27 +0000)]
Turn on section numbering.

23 years agoClean out any old versions of no-longer-installed header files that may
Tom Lane [Tue, 20 Feb 2001 20:37:13 +0000 (20:37 +0000)]
Clean out any old versions of no-longer-installed header files that may
be lurking in the install target directory.  But don't zap up-to-date
headers (so install-all-headers before regular install will work).
Per suggestion from Larry Rosenman.

23 years agoRemove unportable #warning directive.
Tom Lane [Tue, 20 Feb 2001 19:39:24 +0000 (19:39 +0000)]
Remove unportable #warning directive.

23 years agoMake sure -L and -I's for our source tree are always before system include
Peter Eisentraut [Tue, 20 Feb 2001 19:20:30 +0000 (19:20 +0000)]
Make sure -L and -I's for our source tree are always before system include
or library directories on the command line.

23 years agoRemove obsolete claim that it doesn't work to use gcc with HP's C++
Tom Lane [Tue, 20 Feb 2001 19:06:16 +0000 (19:06 +0000)]
Remove obsolete claim that it doesn't work to use gcc with HP's C++
compiler or g++ with HP's C compiler.

23 years agoSelect optimization by default for HP's C compiler.
Tom Lane [Tue, 20 Feb 2001 19:04:04 +0000 (19:04 +0000)]
Select optimization by default for HP's C compiler.

23 years agoUpdate TODO list.
Bruce Momjian [Tue, 20 Feb 2001 17:35:09 +0000 (17:35 +0000)]
Update TODO list.

23 years agoUpdate TODO list.
Bruce Momjian [Tue, 20 Feb 2001 17:25:42 +0000 (17:25 +0000)]
Update TODO list.

23 years agoChanges:
Bruce Momjian [Tue, 20 Feb 2001 15:34:14 +0000 (15:34 +0000)]
Changes:

* reverse the change #include <> -> "" in krb.c.
  It _must not_ include files in "."
* Makefile update.  Inconsistent var usage and SHLIB was
  not set.

Now it should work with all external libs.

arko Kreen

23 years agoTruncate incoming username and database name to NAMEDATALEN-1 characters
Tom Lane [Tue, 20 Feb 2001 01:34:40 +0000 (01:34 +0000)]
Truncate incoming username and database name to NAMEDATALEN-1 characters
so that we don't reject overlength names unnecessarily.

23 years agoUpdate HPUX FAQ.
Bruce Momjian [Tue, 20 Feb 2001 01:33:07 +0000 (01:33 +0000)]
Update HPUX FAQ.

23 years agoFix pg_passwd's failure to cope with usernames > 8 chars.
Tom Lane [Tue, 20 Feb 2001 01:16:49 +0000 (01:16 +0000)]
Fix pg_passwd's failure to cope with usernames > 8 chars.

23 years agoRemove inclusion of <varargs.h> on SunOS; this does not work since we
Tom Lane [Tue, 20 Feb 2001 00:28:07 +0000 (00:28 +0000)]
Remove inclusion of <varargs.h> on SunOS; this does not work since we
use the ANSI varargs style (<stdarg.h>) not the old style.  Tatsuo had
reported this change was necessary back in the 7.0 beta cycle (4/13/00)
but for some reason, making the edit never got done.

23 years agoDocument --enable-cassert.
Tom Lane [Mon, 19 Feb 2001 22:25:43 +0000 (22:25 +0000)]
Document --enable-cassert.

23 years agoUpdate TODO list.
Bruce Momjian [Mon, 19 Feb 2001 21:52:57 +0000 (21:52 +0000)]
Update TODO list.

23 years agoUpdate TODO list.
Bruce Momjian [Mon, 19 Feb 2001 20:46:14 +0000 (20:46 +0000)]
Update TODO list.

23 years agoChange plpgsql's GET DIAGNOSTICS statement to use SQL99-compatible
Tom Lane [Mon, 19 Feb 2001 19:49:53 +0000 (19:49 +0000)]
Change plpgsql's GET DIAGNOSTICS statement to use SQL99-compatible
syntax.  Fix the RESULT_OID case, which never worked.  Add documentation.

23 years agoUpdate TODO list.
Bruce Momjian [Mon, 19 Feb 2001 19:06:47 +0000 (19:06 +0000)]
Update TODO list.

23 years agoUpdate TODO list.
Bruce Momjian [Mon, 19 Feb 2001 16:55:08 +0000 (16:55 +0000)]
Update TODO list.

23 years agoUpdate TODO list.
Bruce Momjian [Mon, 19 Feb 2001 15:23:08 +0000 (15:23 +0000)]
Update TODO list.

23 years agoSynced gram.y and preproc.y.
Michael Meskes [Mon, 19 Feb 2001 07:30:20 +0000 (07:30 +0000)]
Synced gram.y and preproc.y.

23 years agoFix nasty bug in configure.
Tatsuo Ishii [Mon, 19 Feb 2001 05:16:40 +0000 (05:16 +0000)]
Fix nasty bug in configure.
      STRTOL='strtoul.o' --> STRTOUL='strtoul.o'

23 years agoUpdate example of partially constraining join order to use a subselect
Tom Lane [Mon, 19 Feb 2001 00:24:30 +0000 (00:24 +0000)]
Update example of partially constraining join order to use a subselect
in FROM instead of an auxiliary view.  We didn't have subselect-in-FROM
when I wrote this originally...

23 years agoAND/OR truth table needs another row IMHO.
Tom Lane [Mon, 19 Feb 2001 00:01:18 +0000 (00:01 +0000)]
AND/OR truth table needs another row IMHO.

23 years agoAdd copyright mentions, per Tom Lane.
Bruce Momjian [Sun, 18 Feb 2001 18:34:02 +0000 (18:34 +0000)]
Add copyright mentions, per Tom Lane.

23 years agoAllow extract() to accept the same field selectors as date_part(), not just
Peter Eisentraut [Sun, 18 Feb 2001 18:06:10 +0000 (18:06 +0000)]
Allow extract() to accept the same field selectors as date_part(), not just
the ones specified by SQL.

23 years agoAdd --template option to createdb script to allow access to WITH TEMPLATE
Tom Lane [Sun, 18 Feb 2001 17:53:55 +0000 (17:53 +0000)]
Add --template option to createdb script to allow access to WITH TEMPLATE
option of CREATE DATABASE.  In pg_regress, create regression database
from template0 to ensure that any installation-local cruft in template1
will not mess up the tests.

23 years agoEnhance documentation of date/time functions. Add extract, current_date,
Peter Eisentraut [Sun, 18 Feb 2001 17:17:17 +0000 (17:17 +0000)]
Enhance documentation of date/time functions.  Add extract, current_date,
current_time, current_timestamp.  Add more examples.  Remove age(), because
it doesn't work like it is documented, and no one could explain it to me.

23 years agoMake <lineannotation> italic. Fix syntax error in print style.
Peter Eisentraut [Sun, 18 Feb 2001 17:15:03 +0000 (17:15 +0000)]
Make <lineannotation> italic.  Fix syntax error in print style.

23 years agoAdd note that partial indices are not currently supported.
Tom Lane [Sun, 18 Feb 2001 06:07:37 +0000 (06:07 +0000)]
Add note that partial indices are not currently supported.

23 years agoUpdate description of fsync option for 7.1.
Tom Lane [Sun, 18 Feb 2001 05:30:12 +0000 (05:30 +0000)]
Update description of fsync option for 7.1.

23 years agoChange default commit_delay to zero, update documentation.
Tom Lane [Sun, 18 Feb 2001 04:50:43 +0000 (04:50 +0000)]
Change default commit_delay to zero, update documentation.

23 years agoChange s_lock to not use any zero-delay select() calls; these are just a
Tom Lane [Sun, 18 Feb 2001 04:39:42 +0000 (04:39 +0000)]
Change s_lock to not use any zero-delay select() calls; these are just a
waste of cycles on single-CPU machines, and of dubious utility on multi-CPU
machines too.
Tweak s_lock_stuck so that caller can specify timeout interval, and
increase interval before declaring stuck spinlock for buffer locks and XLOG
locks.
On systems that have fdatasync(), use that rather than fsync() to sync WAL
log writes.  Ensure that WAL file is entirely allocated during XLogFileInit.

23 years agoRemove bogus set_ps_display call --- changing displayed status here is
Tom Lane [Sun, 18 Feb 2001 04:28:31 +0000 (04:28 +0000)]
Remove bogus set_ps_display call --- changing displayed status here is
either wrong or unnecessary in most cases, and on systems where setting
status takes a kernel call, the overhead of setting status three times
per command rather than two is annoying.

23 years agoCatch some stray references to .htm files.
Tom Lane [Sat, 17 Feb 2001 16:52:53 +0000 (16:52 +0000)]
Catch some stray references to .htm files.

23 years agoNo more .htm files, so don't try to add them to the doc tarballs.
Tom Lane [Sat, 17 Feb 2001 16:47:57 +0000 (16:47 +0000)]
No more .htm files, so don't try to add them to the doc tarballs.

23 years agoUpdate TODO list.
Bruce Momjian [Sat, 17 Feb 2001 14:35:48 +0000 (14:35 +0000)]
Update TODO list.

23 years agoFix a bug in psql. unescape() does not work for multi-byte encodings.
Tatsuo Ishii [Sat, 17 Feb 2001 10:03:33 +0000 (10:03 +0000)]
Fix a bug in psql. unescape() does not work for multi-byte encodings.

23 years agoFix README in Japanese to sync with README.pgbench updated by Tom.
Tatsuo Ishii [Sat, 17 Feb 2001 06:50:03 +0000 (06:50 +0000)]
Fix README in Japanese to sync with README.pgbench updated by Tom.
(thanks!)

23 years agoSeems a bad idea to assume that select(2) doesn't touch the input masks
Tom Lane [Sat, 17 Feb 2001 03:37:22 +0000 (03:37 +0000)]
Seems a bad idea to assume that select(2) doesn't touch the input masks
if it returns EINTR.

23 years agoAdd current seek position to FDDEBUG output for FileRead,
Tom Lane [Sat, 17 Feb 2001 01:00:04 +0000 (01:00 +0000)]
Add current seek position to FDDEBUG output for FileRead,
FileWrite, FileSeek.

23 years agoJust noticed that use of 'volatile' in HPPA S_UNLOCK() was causing gcc
Tom Lane [Fri, 16 Feb 2001 23:50:40 +0000 (23:50 +0000)]
Just noticed that use of 'volatile' in HPPA S_UNLOCK() was causing gcc
to generate unnecessarily stupid code.  Tweak macro to describe a series
of store-constant ops, not store/load/store/load/store/load/store.

23 years agoUpdate README's install instructions to agree with current reality.
Tom Lane [Fri, 16 Feb 2001 21:41:53 +0000 (21:41 +0000)]
Update README's install instructions to agree with current reality.

23 years agoUpdate bsdi shared memory stuff.
Bruce Momjian [Fri, 16 Feb 2001 19:43:52 +0000 (19:43 +0000)]
Update bsdi shared memory stuff.

23 years agoUpdate BSDI FAQ.
Bruce Momjian [Fri, 16 Feb 2001 19:27:19 +0000 (19:27 +0000)]
Update BSDI FAQ.

23 years agoDefend against starting a non-MULTIBYTE-enabled backend in a database
Tom Lane [Fri, 16 Feb 2001 18:50:40 +0000 (18:50 +0000)]
Defend against starting a non-MULTIBYTE-enabled backend in a database
with encoding other than SQL_ASCII.  Per recent discussion in pghackers.

23 years agoSome more updates...
Peter Mount [Fri, 16 Feb 2001 16:45:01 +0000 (16:45 +0000)]
Some more updates...

Fri Feb 17 15:11:00 GMT 2001 peter@retep.org.uk
        - Reduced the object overhead in PreparedStatement by reusing the same
          StringBuffer object throughout. Similarly SimpleDateStamp's are alse
          reused in a thread save manner.
        - Implemented in PreparedStatement: setNull(), setDate/Time/Timestamp
          using Calendar, setBlob(), setCharacterStream()
        - Clob's are now implemented in ResultSet & PreparedStatement!
        - Implemented a lot of DatabaseMetaData & ResultSetMetaData methods.
          We have about 18 unimplemented methods left in JDBC2 at the current
          time.

23 years agoUpdate BSDI SGML.
Bruce Momjian [Fri, 16 Feb 2001 16:10:06 +0000 (16:10 +0000)]
Update BSDI SGML.

23 years agoichar() has been renamed to chr(), so fix translation table.
Tom Lane [Fri, 16 Feb 2001 03:49:40 +0000 (03:49 +0000)]
ichar() has been renamed to chr(), so fix translation table.

23 years agoFix bugs in pltcl's new return_null command: it was liable to go belly up
Tom Lane [Fri, 16 Feb 2001 03:26:40 +0000 (03:26 +0000)]
Fix bugs in pltcl's new return_null command: it was liable to go belly up
if the return datatype's input converter was at all strict, because the
converter would get called on junk data when returning NULL.  Also
ensure that it gives an error rather than coredumping if someone tries
to use it in a trigger function.

23 years agoFix erroneous sort request in pltcl selftest.
Tom Lane [Fri, 16 Feb 2001 03:22:41 +0000 (03:22 +0000)]
Fix erroneous sort request in pltcl selftest.

23 years agoClean up two rather nasty bugs in operator selection code.
Tom Lane [Fri, 16 Feb 2001 03:16:58 +0000 (03:16 +0000)]
Clean up two rather nasty bugs in operator selection code.

1. If there is exactly one pg_operator entry of the right name and oprkind,
oper() and related routines would return that entry whether its input type
had anything to do with the request or not.  This is just premature
optimization: we shouldn't return the single candidate until after we verify
that it really is a valid candidate, ie, is at least coercion-compatible
with the given types.

2. oper() and related routines only promise a coercion-compatible result.
Unfortunately, there were quite a few callers that assumed the returned
operator is binary-compatible with the given datatype; they would proceed
to call it without making any datatype coercions.  These callers include
sorting, grouping, aggregation, and VACUUM ANALYZE.  In general I think
it is appropriate for these callers to require an exact or binary-compatible
match, so I've added a new routine compatible_oper() that only succeeds if
it can find an operator that doesn't require any run-time conversions.
Callers now call oper() or compatible_oper() depending on whether they are
prepared to deal with type conversion or not.

The upshot of these bugs is revealed by the following silliness in PL/Tcl's
selftest: it creates an operator @< on int4, and then tries to use it to
sort a char(N) column.  The system would let it do that :-( (and evidently
has done so since 6.3 :-( :-().  The result in this case was just a silly
sort order, but the reverse combination would've provoked coredump from
trying to dereference integers.  With this fix you get more reasonable
behavior:
pltcl_test=# select * from T_pkey1 order by key1, key2 using @<;
ERROR:  Unable to identify an operator '@<' for types 'bpchar' and 'bpchar'
        You will have to retype this query using an explicit cast

23 years agoAdd casting for numeric/float4/float8 type value
Hiroshi Inoue [Fri, 16 Feb 2001 03:10:09 +0000 (03:10 +0000)]
Add casting for numeric/float4/float8 type value
automatically to compensate the lack of automatic
conversion functionality of PostgreSQL server.
For example if there's a numeric type binding
   1.2567 --> 1.2567::numeric.
I hope this change would enable the use of numeric
type in MS-Access etc.

Thanks Hiroki Kataoka for his checking my code.

23 years agoUpdate bsdi faq.
Bruce Momjian [Fri, 16 Feb 2001 00:46:21 +0000 (00:46 +0000)]
Update bsdi faq.

23 years agoUpdate bsdi faq.
Bruce Momjian [Fri, 16 Feb 2001 00:17:57 +0000 (00:17 +0000)]
Update bsdi faq.

23 years agoTake OUTER JOIN semantics into account when estimating the size of join
Tom Lane [Fri, 16 Feb 2001 00:03:08 +0000 (00:03 +0000)]
Take OUTER JOIN semantics into account when estimating the size of join
relations.  It's not very bright, but at least it now knows that
A LEFT JOIN B must produce at least as many rows as are in A ...

23 years agoUpdate FAQ.
Bruce Momjian [Thu, 15 Feb 2001 22:21:23 +0000 (22:21 +0000)]
Update FAQ.

23 years agoUpdate FAQ.
Bruce Momjian [Thu, 15 Feb 2001 22:15:10 +0000 (22:15 +0000)]
Update FAQ.

23 years agoAdd some notes about memory management of RI plans.
Tom Lane [Thu, 15 Feb 2001 21:57:43 +0000 (21:57 +0000)]
Add some notes about memory management of RI plans.

23 years agoUpdate comments about memory management.
Tom Lane [Thu, 15 Feb 2001 21:47:08 +0000 (21:47 +0000)]
Update comments about memory management.

23 years agoUpdate notes about memory context scheme.
Tom Lane [Thu, 15 Feb 2001 21:38:26 +0000 (21:38 +0000)]
Update notes about memory context scheme.

23 years agoUpdate obsolete wording of error message.
Tom Lane [Thu, 15 Feb 2001 21:11:17 +0000 (21:11 +0000)]
Update obsolete wording of error message.

23 years agoAlthough we can't support out-of-line TOAST storage in indexes (yet),
Tom Lane [Thu, 15 Feb 2001 20:57:01 +0000 (20:57 +0000)]
Although we can't support out-of-line TOAST storage in indexes (yet),
compressed storage works perfectly well.  Might as well have a coherent
strategy for applying it, rather than the haphazard store-what-you-get
approach that was in the code before.  The strategy I've set up here is
to attempt compression of any compressible index value exceeding
BLCKSZ/16, or about 500 bytes by default.

23 years agoFurther polishing of documentation about new fmgr call convention.
Tom Lane [Thu, 15 Feb 2001 19:03:35 +0000 (19:03 +0000)]
Further polishing of documentation about new fmgr call convention.

23 years agoReduce default selectivity estimates for geometric operators; it seems
Tom Lane [Thu, 15 Feb 2001 17:55:17 +0000 (17:55 +0000)]
Reduce default selectivity estimates for geometric operators; it seems
the old ones were not small enough to ensure r-tree and gist indexes would
get picked when available.  These numbers are totally bogus anyway, but
in the absence of any real estimation technique, we'd like to select
indexes when available ...