OSDN Git Service

pg-rex/syncrep.git
23 years agoFirst non-stub implementation of shared free space map. It's not super
Tom Lane [Mon, 2 Jul 2001 20:50:46 +0000 (20:50 +0000)]
First non-stub implementation of shared free space map.  It's not super
useful as yet, since its primary source of information is (full) VACUUM,
which makes a concerted effort to get rid of free space before telling
the map about it ... next stop is concurrent VACUUM ...

23 years agoUpdate abstime expected results to match post-30-June-2001 reality.
Tom Lane [Mon, 2 Jul 2001 19:31:52 +0000 (19:31 +0000)]
Update abstime expected results to match post-30-June-2001 reality.
Probably the right fix is to remove 'current' special value entirely,
but I don't want to see regression test failures until that happens.

23 years agoMake the found-a-buffer-when-we-were-expecting-to-extend-the-rel path
Tom Lane [Mon, 2 Jul 2001 18:47:18 +0000 (18:47 +0000)]
Make the found-a-buffer-when-we-were-expecting-to-extend-the-rel path
actually work.  It had been throwing an Assert as of my recent changes
to bufmgr.c, but was not really right even before that AFAICT.

23 years agoUnder new theory of operation wherein postmaster forks children
Tom Lane [Sun, 1 Jul 2001 00:06:23 +0000 (00:06 +0000)]
Under new theory of operation wherein postmaster forks children
immediately, we will fork a child even if the database state does not
permit connections to be accepted (eg, we are in recovery mode).
The child process will correctly reject the connection and exit as
soon as it's finished collecting the connection request message.
However, this means that reaper() must be prepared to see child
process exit signals even while it's waiting for startup or shutdown
process to finish.  As was, a connection request arriving during a
database recovery or shutdown would cause postmaster abort.

23 years agoAllow default transaction isolation level (a.k.a. set session
Peter Eisentraut [Sat, 30 Jun 2001 22:03:26 +0000 (22:03 +0000)]
Allow default transaction isolation level (a.k.a. set session
characteristics) to be set through GUC.

23 years agofix markup
Peter Eisentraut [Sat, 30 Jun 2001 22:01:17 +0000 (22:01 +0000)]
fix markup

23 years agoAdd missing dependency and fix thinko.
Peter Eisentraut [Sat, 30 Jun 2001 21:58:06 +0000 (21:58 +0000)]
Add missing dependency and fix thinko.

23 years agoThe System V resources discussed here apply to NetBSD as well. In fact
D'Arcy J.M. Cain [Sat, 30 Jun 2001 21:15:57 +0000 (21:15 +0000)]
The System V resources discussed here apply to NetBSD as well.  In fact
I just finished changing the GENERIC files and the options(4) man page
on NetBSD to make this clearer.

23 years agoportability enhancements
Peter Eisentraut [Sat, 30 Jun 2001 19:01:27 +0000 (19:01 +0000)]
portability enhancements

23 years agoNLS for the psql \d family of commands. (E.g., the column headers will
Peter Eisentraut [Sat, 30 Jun 2001 17:26:12 +0000 (17:26 +0000)]
NLS for the psql \d family of commands.  (E.g., the column headers will
have localized strings.)  Also, modernize the system catalog queries where
appropriate, e.g., with outer joins.

23 years agogcc did not like new pg_stat macros, for good and sufficient reason.
Tom Lane [Fri, 29 Jun 2001 23:03:02 +0000 (23:03 +0000)]
gcc did not like new pg_stat macros, for good and sufficient reason.
Add 'do { ... } while (0)' decoration to eliminate compiler warnings.

23 years agoUpdate TODO list.
Bruce Momjian [Fri, 29 Jun 2001 21:28:16 +0000 (21:28 +0000)]
Update TODO list.

23 years agoFurther work on connecting the free space map (which is still just a
Tom Lane [Fri, 29 Jun 2001 21:08:25 +0000 (21:08 +0000)]
Further work on connecting the free space map (which is still just a
stub) into the rest of the system.  Adopt a cleaner approach to preventing
deadlock in concurrent heap_updates: allow RelationGetBufferForTuple to
select any page of the rel, and put the onus on it to lock both buffers
in a consistent order.  Remove no-longer-needed isExtend hack from
API of ReleaseAndReadBuffer.

23 years agoFix VACUUM so that it can use pages as move targets even if they do not
Tom Lane [Fri, 29 Jun 2001 20:14:27 +0000 (20:14 +0000)]
Fix VACUUM so that it can use pages as move targets even if they do not
have any newly-dead tuples on them.  This is a longstanding deficiency
that prevents VACUUM from compacting a file as much as one would expect.
Change requires fixing repair_frag to not assume that fraged_pages is
a subset of vacuum_pages.
Also make some further cleanups of places that assumed page numbers fit
in int and tuple counts fit in uint32.

23 years agoAdd replication email.
Bruce Momjian [Fri, 29 Jun 2001 20:10:12 +0000 (20:10 +0000)]
Add replication email.

23 years agoAdd info on MATCH PARTIAL.
Bruce Momjian [Fri, 29 Jun 2001 20:07:23 +0000 (20:07 +0000)]
Add info on MATCH PARTIAL.

23 years agoSimpleDateFormat performance improvement, thread-safe.
Bruce Momjian [Fri, 29 Jun 2001 17:23:33 +0000 (17:23 +0000)]
SimpleDateFormat performance improvement, thread-safe.

Barry Lind

23 years agoFix longstanding error in VACUUM: sometimes would examine a buffer page
Tom Lane [Fri, 29 Jun 2001 16:34:30 +0000 (16:34 +0000)]
Fix longstanding error in VACUUM: sometimes would examine a buffer page
after writing/unpinning it.  An actual failure is unlikely, unless the
system is tremendously short of buffers ... but a bug is a bug.

23 years agoTurned high-frequently called pgstat functions into macros
Jan Wieck [Fri, 29 Jun 2001 16:29:37 +0000 (16:29 +0000)]
Turned high-frequently called pgstat functions into macros
for speed.

Jan

23 years agoFix erroneous GUC variable references from commandline-GUC patch.
Tom Lane [Fri, 29 Jun 2001 16:05:57 +0000 (16:05 +0000)]
Fix erroneous GUC variable references from commandline-GUC patch.

23 years agoRepair logic error for multi-key indexes. From Oleg Bartunov.
Tom Lane [Thu, 28 Jun 2001 16:00:07 +0000 (16:00 +0000)]
Repair logic error for multi-key indexes.  From Oleg Bartunov.

23 years agoAdd to thread item.
Bruce Momjian [Thu, 28 Jun 2001 15:19:11 +0000 (15:19 +0000)]
Add to thread item.

23 years agoInstall infrastructure for shared-memory free space map. Doesn't actually
Tom Lane [Wed, 27 Jun 2001 23:31:40 +0000 (23:31 +0000)]
Install infrastructure for shared-memory free space map.  Doesn't actually
do anything yet, but it has the necessary connections to initialization
and so forth.  Make some gestures towards allowing number of blocks in
a relation to be BlockNumber, ie, unsigned int, rather than signed int.
(I doubt I got all the places that are sloppy about it, yet.)  On the
way, replace the hardwired NLOCKS_PER_XACT fudge factor with a GUC
variable.

23 years agoNational language support for pg_dump and pg_restore. Combined with big
Peter Eisentraut [Wed, 27 Jun 2001 21:21:37 +0000 (21:21 +0000)]
National language support for pg_dump and pg_restore.  Combined with big
message clean up.

23 years agoRemove another unused include file with obsolete, useless,
Tom Lane [Wed, 27 Jun 2001 19:02:48 +0000 (19:02 +0000)]
Remove another unused include file with obsolete, useless,
confusing definitions in it.

23 years agoAdd jdbc item.
Bruce Momjian [Wed, 27 Jun 2001 18:07:11 +0000 (18:07 +0000)]
Add jdbc item.

23 years agoUpdate TODO list.
Bruce Momjian [Wed, 27 Jun 2001 18:06:49 +0000 (18:06 +0000)]
Update TODO list.

23 years agoUpdate TODO list.
Bruce Momjian [Wed, 27 Jun 2001 18:05:37 +0000 (18:05 +0000)]
Update TODO list.

23 years agoUpdate TODO list.
Bruce Momjian [Wed, 27 Jun 2001 17:27:40 +0000 (17:27 +0000)]
Update TODO list.

23 years agoUpdate TODO list.
Bruce Momjian [Wed, 27 Jun 2001 16:24:17 +0000 (16:24 +0000)]
Update TODO list.

23 years agoHandle Procedure calls.
Hiroshi Inoue [Wed, 27 Jun 2001 07:38:07 +0000 (07:38 +0000)]
Handle Procedure calls.
Now the version is 7.01.0006.

23 years agoUpdate TODO list.
Bruce Momjian [Tue, 26 Jun 2001 19:43:11 +0000 (19:43 +0000)]
Update TODO list.

23 years agoUpdate TODO list.
Bruce Momjian [Tue, 26 Jun 2001 15:07:59 +0000 (15:07 +0000)]
Update TODO list.

23 years agoUpdate TODO list.
Bruce Momjian [Tue, 26 Jun 2001 04:04:49 +0000 (04:04 +0000)]
Update TODO list.

23 years agoUpdate TODO list.
Bruce Momjian [Tue, 26 Jun 2001 04:03:25 +0000 (04:03 +0000)]
Update TODO list.

23 years agoFix a couple remaining places where GUC variables were assigned to
Tom Lane [Mon, 25 Jun 2001 23:03:03 +0000 (23:03 +0000)]
Fix a couple remaining places where GUC variables were assigned to
directly, rather than through SetConfigOption().

23 years agoFix a couple remaining places where GUC variables were assigned to
Tom Lane [Mon, 25 Jun 2001 22:56:05 +0000 (22:56 +0000)]
Fix a couple remaining places where GUC variables were assigned to
directly, rather than through SetConfigOption().

23 years agoOptimizer can now estimate selectivity of IS NULL, IS NOT NULL,
Tom Lane [Mon, 25 Jun 2001 21:11:45 +0000 (21:11 +0000)]
Optimizer can now estimate selectivity of IS NULL, IS NOT NULL,
IS TRUE, etc, with some degree of verisimilitude.  Split out
selectivity support functions from builtins.h into a new header
file selfuncs.h, so as to reduce the number of header files builtins.h
must depend on.  Fix a few missing inclusions exposed thereby.
From Joe Conway, with some kibitzing from Tom Lane.

23 years agoSuppress gcc warning.
Tom Lane [Mon, 25 Jun 2001 20:20:02 +0000 (20:20 +0000)]
Suppress gcc warning.

23 years agoHigh memory usage
Bruce Momjian [Mon, 25 Jun 2001 01:53:59 +0000 (01:53 +0000)]
High memory usage

Here is a patch which inspired by Michael Stephens that should work

Dave Cramer

23 years agoBack out BYTEA binary compatibility changes.
Bruce Momjian [Sun, 24 Jun 2001 02:41:21 +0000 (02:41 +0000)]
Back out BYTEA binary compatibility changes.

23 years agoDon't use a temp file. It was created insecurely and was easy to do without.
Peter Eisentraut [Sat, 23 Jun 2001 23:29:48 +0000 (23:29 +0000)]
Don't use a temp file.  It was created insecurely and was easy to do without.

23 years agoSince a missing pg_pwd file is a valid situation, don't print an error
Peter Eisentraut [Sat, 23 Jun 2001 23:26:17 +0000 (23:26 +0000)]
Since a missing pg_pwd file is a valid situation, don't print an error
message in that case.

23 years ago> Marko Kreen <marko@l-t.ee> writes:
Bruce Momjian [Sat, 23 Jun 2001 22:23:49 +0000 (22:23 +0000)]
> Marko Kreen <marko@l-t.ee> writes:
> > secure_ctx changes too.  it will be PGC_BACKEND after '-p'.
>
> Oh, okay, I missed that part.  Could we see the total state of the
> patch --- ie, a diff against current CVS, not a bunch of deltas?
> I've gotten confused about what's in and what's out.

Ok, here it is.  Cleared the ctx comment too - after -p
it will be PGC_BACKEND in any case.

Marko Kreen

23 years agoHere the latest version and first working release of Ora2Pg a tool to export
Bruce Momjian [Sat, 23 Jun 2001 02:56:23 +0000 (02:56 +0000)]
Here the latest version and first working release of Ora2Pg a tool to export
Oracle database to PostgreSQL.

    It currently dump the database schema (tables, views, sequences,
    indexes, grants), with primary, unique and foreign keys into PostgreSQL
    syntax without editing the SQL code generated. You can dump only a
    particular schema from the Oracle database.

    Functions, procedures and triggers with SQL or PLSQL code generated must
    be reviewed to match the PostgreSQL syntax. Some usefull recommandation
    on porting Oracle to PostgreSQL can be found at
http://techdocs.postgresql.org/
    under the "Converting from other Databases to PostgreSQL" Oracle part. I
just
    notice one thing more is that the trunc() function in Oracle is the same for
number
    or date so be carefull when porting to PostgreSQL to use trunc() for number
and
    date_trunc() for date.

I will add more precision in type NUMBER conversion based on length to match
as closest as possible all rich PostgreSQL numerics type. But it seems not to be

urgent as it seems that Oracle DBAs only create number with length 22 (default)
Space seems not to be their problem...

Gilles DAROLD

23 years agoAdd TEMPORARY sequences and have SERIAL on a temp table have a temporary
Bruce Momjian [Sat, 23 Jun 2001 00:07:34 +0000 (00:07 +0000)]
Add TEMPORARY sequences and have SERIAL on a temp table have a temporary
sequence.

23 years agofix markup
Peter Eisentraut [Sat, 23 Jun 2001 00:03:10 +0000 (00:03 +0000)]
fix markup

23 years agoupdates for new startup sequence, some reformatting
Peter Eisentraut [Fri, 22 Jun 2001 23:27:48 +0000 (23:27 +0000)]
updates for new startup sequence, some reformatting

23 years agoAttached is documentation describing plperlu differences from plperl.
Bruce Momjian [Fri, 22 Jun 2001 21:37:14 +0000 (21:37 +0000)]
Attached is documentation describing plperlu differences from plperl.

Alex Pilosov

23 years agoThe new files for statistical system views.
Jan Wieck [Fri, 22 Jun 2001 19:18:36 +0000 (19:18 +0000)]
The new files for statistical system views.

Jan

23 years agoStatistical system views (yet without the config stuff, but
Jan Wieck [Fri, 22 Jun 2001 19:16:24 +0000 (19:16 +0000)]
Statistical system views (yet without the config stuff, but
it's hard to keep such massive changes in sync with the tree
so I need to get it in and work from there now).

Jan

23 years agoMention that EXPLAIN results on small tables shouldn't be extrapolated
Tom Lane [Fri, 22 Jun 2001 18:53:36 +0000 (18:53 +0000)]
Mention that EXPLAIN results on small tables shouldn't be extrapolated
to large tables.  Recommend ANALYZE or VACUUM ANALYZE after populating
a table.

23 years agoInclude catalog/pg_type.h instead of manually extracting the interesting
Peter Eisentraut [Fri, 22 Jun 2001 17:48:39 +0000 (17:48 +0000)]
Include catalog/pg_type.h instead of manually extracting the interesting
oid values.

23 years agoChange SQLPrimaryKeys() so that it detects the primary key
Hiroshi Inoue [Fri, 22 Jun 2001 05:59:43 +0000 (05:59 +0000)]
Change SQLPrimaryKeys() so that it detects the primary key
other than tablename_pkey.

23 years agoAdd support to lock manager for conditionally locking a lock (ie,
Tom Lane [Fri, 22 Jun 2001 00:04:59 +0000 (00:04 +0000)]
Add support to lock manager for conditionally locking a lock (ie,
return without waiting if we can't get the lock immediately).
Not used yet, but will be needed for concurrent VACUUM.

23 years agoRemove unused include file for long-dead flavors of locking.
Tom Lane [Thu, 21 Jun 2001 21:01:36 +0000 (21:01 +0000)]
Remove unused include file for long-dead flavors of locking.

23 years agoTweak error message.
Tom Lane [Thu, 21 Jun 2001 19:45:45 +0000 (19:45 +0000)]
Tweak error message.

23 years agoFix strangely formatted comment.
Bruce Momjian [Thu, 21 Jun 2001 18:25:54 +0000 (18:25 +0000)]
Fix strangely formatted comment.

23 years agoA bit of code beautification/cleanup of obsolete comments. Rethink
Tom Lane [Thu, 21 Jun 2001 16:43:24 +0000 (16:43 +0000)]
A bit of code beautification/cleanup of obsolete comments.  Rethink
ordering of startup operations in one or two places.

23 years agoShared libraries should not be linked explicitly against -lc on FreeBSD
Peter Eisentraut [Wed, 20 Jun 2001 20:25:11 +0000 (20:25 +0000)]
Shared libraries should not be linked explicitly against -lc on FreeBSD
(see http://mail.gnu.org/pipermail/libtool/2000-July/001824.html).  But
apparently it should still be present with -Bsymbolic.

23 years ago-Bsymbolic requires -lc on NetBSD. Also add rpath for non-ELF.
Peter Eisentraut [Wed, 20 Jun 2001 20:15:45 +0000 (20:15 +0000)]
-Bsymbolic requires -lc on NetBSD.  Also add rpath for non-ELF.

from Johnny C. Lam <lamj@stat.cmu.edu> via D'Arcy

23 years agoRe-add explicit declaration of filename_completion_function(), which seems
Peter Eisentraut [Wed, 20 Jun 2001 18:39:14 +0000 (18:39 +0000)]
Re-add explicit declaration of filename_completion_function(), which seems
to be missing in some header files (on OpenBSD 2.8?).

23 years agoThere is no RTLD_GLOBAL on OpenBSD, says Alex Pilosov <alex@pilosoft.com>.
Peter Eisentraut [Wed, 20 Jun 2001 18:33:31 +0000 (18:33 +0000)]
There is no RTLD_GLOBAL on OpenBSD, says Alex Pilosov <alex@pilosoft.com>.

23 years agoSwedish translation for psql (my interpretation of Swedish, surely)
Peter Eisentraut [Wed, 20 Jun 2001 18:25:26 +0000 (18:25 +0000)]
Swedish translation for psql (my interpretation of Swedish, surely)

23 years agoHandle reading of startup packet and authentication exchange after forking
Peter Eisentraut [Wed, 20 Jun 2001 18:07:56 +0000 (18:07 +0000)]
Handle reading of startup packet and authentication exchange after forking
a new postmaster child process.  This should eliminate problems with
authentication blocking (e.g., ident, SSL init) and also reduce problems
with the accept queue filling up under heavy load.

The option to send elog output to a different file per backend (postgres -o)
has been disabled for now because the initialization would have to happen
in a different order and it's not clear we want to keep this anyway.

23 years agoMake sure that everything says version 3.2.
D'Arcy J.M. Cain [Wed, 20 Jun 2001 11:20:34 +0000 (11:20 +0000)]
Make sure that everything says version 3.2.

23 years agoAdd NUMERICOID to this script. This script can be run occasionally to
D'Arcy J.M. Cain [Wed, 20 Jun 2001 11:19:56 +0000 (11:19 +0000)]
Add NUMERICOID to this script.  This script can be run occasionally to
make sure that we are using the right #defines in pgmodule.c but the
OIDs are never actually expected to change.

23 years agoChange the driver so that large error messages are returned
Hiroshi Inoue [Wed, 20 Jun 2001 07:06:40 +0000 (07:06 +0000)]
Change the driver so that large error messages are returned
by multiple SQLError calls.

23 years agoApparently, on some systems, ExtUtils::Embed and MakeMaker are slightly
Bruce Momjian [Wed, 20 Jun 2001 00:26:06 +0000 (00:26 +0000)]
Apparently, on some systems, ExtUtils::Embed and MakeMaker are slightly
broken, and its impossible to make a shared library when compiling with
both CCDLFLAGS and LDDLFAGS, you have to pick one or the other.

Alex Pilosov

23 years agoAttached is a patch to deal with the ones that I missed the first time
Bruce Momjian [Wed, 20 Jun 2001 00:04:44 +0000 (00:04 +0000)]
Attached is a patch to deal with the ones that I missed the first time
around.  I tested this patch under Cygwin and Linux.

Note that I only changed dblink's Makefile in the most minimal way
to fix the link problem under Cygwin (i.e., use the link rule from
Makefile.shlib instead).  dblink's Makefile should probably be further
patched to be consistent with the other PostgreSQL Makefiles.

Jason Tishler

23 years agoBack out SET ALL patch because it is breaking things.
Bruce Momjian [Tue, 19 Jun 2001 23:40:10 +0000 (23:40 +0000)]
Back out SET ALL patch because it is breaking things.

23 years agoAdd IS UNKNOWN, IS NOT UNKNOWN boolean tests, fix the existing boolean
Tom Lane [Tue, 19 Jun 2001 22:39:12 +0000 (22:39 +0000)]
Add IS UNKNOWN, IS NOT UNKNOWN boolean tests, fix the existing boolean
tests to return the correct results per SQL9x when given NULL inputs.
Reimplement these tests as well as IS [NOT] NULL to have their own
expression node types, instead of depending on special functions.
From Joe Conway, with a little help from Tom Lane.

23 years agoFix badly broken RelationGetRelationName().
Tom Lane [Tue, 19 Jun 2001 21:28:41 +0000 (21:28 +0000)]
Fix badly broken RelationGetRelationName().

23 years agoClean up some longstanding problems in shared-cache invalidation.
Tom Lane [Tue, 19 Jun 2001 19:42:16 +0000 (19:42 +0000)]
Clean up some longstanding problems in shared-cache invalidation.
SI messages now include the relevant database OID, so that operations
in one database do not cause useless cache flushes in backends attached
to other databases.  Declare SI messages properly using a union, to
eliminate the former assumption that Oid is the same size as int or Index.
Rewrite the nearly-unreadable code in inval.c, and document it better.
Arrange for catcache flushes at end of command/transaction to happen before
relcache flushes do --- this avoids loading a new tuple into the catcache
while setting up new relcache entry, only to have it be flushed again
immediately.

23 years agoMove temprel name define from temprel.h to rel.h.
Bruce Momjian [Tue, 19 Jun 2001 12:03:41 +0000 (12:03 +0000)]
Move temprel name define from temprel.h to rel.h.

23 years agoBring RelationGetRelationName into sync with new temp rel names.
Tom Lane [Tue, 19 Jun 2001 05:11:50 +0000 (05:11 +0000)]
Bring RelationGetRelationName into sync with new temp rel names.
Quick hack -- need to decide which header should include the other.

23 years agoCorrected the check for *message truncated* for the future use.
Hiroshi Inoue [Tue, 19 Jun 2001 03:17:12 +0000 (03:17 +0000)]
Corrected the check for *message truncated* for the future use.

23 years agoChanged the error handling as follows.
Hiroshi Inoue [Tue, 19 Jun 2001 02:17:06 +0000 (02:17 +0000)]
Changed the error handling as follows.
1) ERRORs cause an SQL_ERROR and the SQLSTATE='S1000'.
2) NOTICEs cause an SQL_SUCCESS_WITH_INFO and the succeeding
   SQLError() returns the NOTICE message.

23 years agoRESET ALL secondary patch:
Bruce Momjian [Mon, 18 Jun 2001 23:42:32 +0000 (23:42 +0000)]
RESET ALL secondary patch:

Here is Tomified version of my 2 pending patches.
Dropped the set_.._real change as it is not needed.
Desc would be:

* use GUC for settings from cmdline

Marko Kreen

23 years agoWell, after persuading cvsup and cvs that it _is_ possible to have local
Bruce Momjian [Mon, 18 Jun 2001 21:40:06 +0000 (21:40 +0000)]
Well, after persuading cvsup and cvs that it _is_ possible to have local
modifiable repositories, I have a clean untrusted plperl patch to offer
you :)

Highlights:
* There's one perl interpreter used for both trusted and untrusted
procedures. I do think its unnecessary to keep two perl
interpreters around. If someone can break out from trusted "Safe" perl
mode, well, they can do what they want already. If someone disagrees, I
can change this.

* Opcode is not statically loaded anymore. Instead, we load Dynaloader,
which then can grab Opcode (and anything else you can 'use') on its own.

* Checked to work on FreeBSD 4.3 + perl 5.5.3 , OpenBSD 2.8 + perl5.6.1,
RedHat 6.2 + perl 5.5.3

* Uses ExtUtils::Embed to find what options are necessary to link with
perl shared libraries

* createlang is also updated, it can create untrusted perl using 'plperlu'

* Example script (assuming you have Mail::Sendmail installed):
create function foo() returns text as '
         use Mail::Sendmail;

         %mail = ( To      => q(you@yourname.com),
                   From    => q(me@here.com),
                   Message => "This is a very short message"
                  );
         sendmail(%mail) or die $Mail::Sendmail::error;
return          "OK. Log says:\n", $Mail::Sendmail::log;
' language 'plperlu';

Alex Pilosov

23 years agoThe attached patch enables the contrib subtree to build cleanly under
Bruce Momjian [Mon, 18 Jun 2001 21:38:02 +0000 (21:38 +0000)]
The attached patch enables the contrib subtree to build cleanly under
Cygwin with the possible exception of mSQL-interface.  Since I don't
have mSQL installed, I skipped this tool.

Except for dealing with a missing getopt.h (oid2name) and HUGE (seg),
the bulk of the patch uses the standard PostgreSQL approach to deal with
Windows DLL issues.

I tested the build aspect of this patch under Cygwin and Linux without
any ill affects.  Note that I did not actually attempt to test the code
for functionality.

The procedure to apply the patch is as follows:

    $ # save the attachment as /tmp/contrib.patch
    $ # change directory to the top of the PostgreSQL source tree
    $ patch -p0 </tmp/contrib.patch

Jason

23 years agoUntar copyright.html file and make good version.
Bruce Momjian [Mon, 18 Jun 2001 21:34:34 +0000 (21:34 +0000)]
Untar copyright.html file and make good version.

23 years agoHere's a small patch for dblink:
Bruce Momjian [Mon, 18 Jun 2001 19:09:50 +0000 (19:09 +0000)]
Here's a small patch for dblink:

   - fixed dblink invalid pointer causing corrupt elog message
   - fixed dblink_tok improper handling of null results
   - fixed examples in README.dblink

Joe Conway

23 years agoFix some references to USE_AUSTRALIAN_RULES that Bruce missed;
Tom Lane [Mon, 18 Jun 2001 19:05:11 +0000 (19:05 +0000)]
Fix some references to USE_AUSTRALIAN_RULES that Bruce missed;
plus a few trivial improvements in wording.

23 years agoUpdate FAQ.
Bruce Momjian [Mon, 18 Jun 2001 18:56:03 +0000 (18:56 +0000)]
Update FAQ.

23 years agoAdd new dblink to /contrib Makefile.'
Bruce Momjian [Mon, 18 Jun 2001 17:20:56 +0000 (17:20 +0000)]
Add new dblink to /contrib Makefile.'

23 years agoAdd GUC setting for Australian timezones. Uses new GUC boolean callback
Bruce Momjian [Mon, 18 Jun 2001 16:14:44 +0000 (16:14 +0000)]
Add GUC setting for Australian timezones.  Uses new GUC boolean callback
functions to clear date cache.  Allow regression tests to pass when
timezone set.

23 years agoAllow removal of system-named pg_* temp tables. Rename temp file/dir as
Bruce Momjian [Mon, 18 Jun 2001 16:13:21 +0000 (16:13 +0000)]
Allow removal of system-named pg_* temp tables.  Rename temp file/dir as
pgsql_tmp.

23 years agoImprove wording of authentication files.
Bruce Momjian [Mon, 18 Jun 2001 16:11:30 +0000 (16:11 +0000)]
Improve wording of authentication files.

23 years agoDo some restructuring to improve performance of the catcaches. Teach
Tom Lane [Mon, 18 Jun 2001 03:35:07 +0000 (03:35 +0000)]
Do some restructuring to improve performance of the catcaches.  Teach
CatalogCacheFlushRelation (formerly called SystemCacheRelationFlushed)
how to distinguish tuples it should flush from those it needn't; this
means a relcache flush event now only removes the catcache entries
it ought to, rather than zapping the caches completely as it used to.
Testing with the regression tests indicates that this considerably
improves the lifespan of catcache entries.  Also, rearrange catcache
data structures so that the limit on number of cached tuples applies
globally across all the catcaches, rather than being per-catcache.
It was a little silly to have the same size limit on both, say,
pg_attribute caches and pg_am caches (there being only four possible
rows in the latter...).  Doing LRU removal across all the caches
instead of locally in each one should reduce cache reload traffic
in the more heavily used caches and improve the efficiency of
cache memory use.

23 years agoFix *escape* handling in copy_statement_with_parameters(was my fault).
Hiroshi Inoue [Mon, 18 Jun 2001 02:16:09 +0000 (02:16 +0000)]
Fix *escape* handling in copy_statement_with_parameters(was my fault).

23 years agoKeep the list of to-be-NOTIFYed names in a plain List palloc'd in
Tom Lane [Sun, 17 Jun 2001 22:27:15 +0000 (22:27 +0000)]
Keep the list of to-be-NOTIFYed names in a plain List palloc'd in
TopTransactionContext, rather than using Dllist.  This simplifies and
speeds up the code, and eliminates a former risk of coredump when
out of memory (since the old code didn't bother to check for malloc
failure).  It also moves us one step closer to retiring Dllist...

23 years agoMake inet/cidr << and <<= operators indexable. From Alex Pilosov <alex@pilosoft...
Tom Lane [Sun, 17 Jun 2001 02:05:20 +0000 (02:05 +0000)]
Make inet/cidr << and <<= operators indexable.  From Alex Pilosov <alex@pilosoft.com>.

23 years agoTweak startup sequence so that running out of PROC array slots is
Tom Lane [Sat, 16 Jun 2001 22:58:17 +0000 (22:58 +0000)]
Tweak startup sequence so that running out of PROC array slots is
detected sooner in backend startup, and is treated as an expected error
(it gives 'Sorry, too many clients already' now).  This allows us not
to have to enforce the MaxBackends limit exactly in the postmaster.
Also, remove ProcRemove() and fold its functionality into ProcKill().
There's no good reason for a backend not to be responsible for removing
its PROC entry, and there are lots of good reasons for the postmaster
not to be touching shared-memory data structures.

23 years agoIt turns out that the relcache thinks it can distinguish different
Tom Lane [Sat, 16 Jun 2001 18:59:31 +0000 (18:59 +0000)]
It turns out that the relcache thinks it can distinguish different
rules and triggers by OID.  So, even though we have no cross-references
in the system catalogs to pg_rewrite.oid or pg_trigger.oid, we'd better
have unique indexes on them.  Put back pg_rewrite_oid_index, which I
mistakenly removed a few days ago, and add pg_trigger_oid_index.

23 years agoOops, forgot to commit doco updates for has_table_privilege.
Tom Lane [Fri, 15 Jun 2001 21:03:07 +0000 (21:03 +0000)]
Oops, forgot to commit doco updates for has_table_privilege.

23 years agofixed bug in connect.c
Michael Meskes [Fri, 15 Jun 2001 09:55:51 +0000 (09:55 +0000)]
fixed bug in connect.c

23 years agoAdd bpchar to list of string types.
D'Arcy J.M. Cain [Fri, 15 Jun 2001 02:23:18 +0000 (02:23 +0000)]
Add bpchar to list of string types.
Thanks to Steve McClure <smcclure@racemi.com> for the patch.

23 years agoAdd missing PG_SETMASK(&BlockSig) to SIGHUP_handler().
Tom Lane [Thu, 14 Jun 2001 19:59:24 +0000 (19:59 +0000)]
Add missing PG_SETMASK(&BlockSig) to SIGHUP_handler().

23 years agoRevoke public read access from pg_statistic, create new system view
Tom Lane [Thu, 14 Jun 2001 19:47:25 +0000 (19:47 +0000)]
Revoke public read access from pg_statistic, create new system view
pg_stats to provide controlled (and, hopefully, more readable) access
to statistics.  Comments on definition of pg_stats welcome.
I didn't force initdb, but the rules regress test will fail until you
do one.