]> granicus.if.org Git - postgresql/log
postgresql
26 years agoUpdate source code to Byron's v6.30.0250 sources plus minor cleanup
Thomas G. Lockhart [Tue, 6 Oct 1998 05:58:41 +0000 (05:58 +0000)]
Update source code to Byron's v6.30.0250 sources plus minor cleanup
 to get rid of unused variables.
Get clean compile on Linux (Thomas and Gerald).
Implement autoconf/configure for standalone builds and use the existing
 autoconf/configure system when in the Postgres source tree.
Code tests and functions with ApplixWare-4.4.1beta on a Linux box.
Changes should be backward compatible with WIN32 but still needs testing.

26 years agoInstall man pages from here now. "make install" will do that along with
Thomas G. Lockhart [Tue, 6 Oct 1998 05:44:05 +0000 (05:44 +0000)]
Install man pages from here now. "make install" will do that along with
the html/ps material...

26 years agoTry to configure interfaces/odbc using the --with-odbc switch.
Thomas G. Lockhart [Tue, 6 Oct 1998 05:42:58 +0000 (05:42 +0000)]
Try to configure interfaces/odbc using the --with-odbc switch.
Suppress installing man pages from here; do it from the doc/Makefile now.

26 years agoForgot to initialize double quote option to be enabled.
Thomas G. Lockhart [Tue, 6 Oct 1998 05:35:42 +0000 (05:35 +0000)]
Forgot to initialize double quote option to be enabled.

26 years agoAdd mention of the new "suppress double quotes" command line option "-n".
Thomas G. Lockhart [Tue, 6 Oct 1998 05:05:48 +0000 (05:05 +0000)]
Add mention of the new "suppress double quotes" command line option "-n".

26 years agoFix for double free from Tatsuo Ishii
Bruce Momjian [Tue, 6 Oct 1998 03:55:43 +0000 (03:55 +0000)]
Fix for double free from Tatsuo Ishii

26 years ago> > There is a bug in check_foreign_key of refint.c which is bundled with
Bruce Momjian [Tue, 6 Oct 1998 03:12:59 +0000 (03:12 +0000)]
> > There is a bug in check_foreign_key of refint.c which is bundled with
> > the standard distribution. It occurs when a trigger calling this
> > function recursively fires another trigger which calls the same
> > function. The calling check_foreign_key loses its plan informantion and
> > when it tries to use it the backend closes its channel. You can check it
> > with the sql script I am attaching below.
> > The solution to this is to do a find_plan again before executing it at
> > line 483 of refint.c.
> > Therefore two more lines should be added before line 483:

Anand Surelia

26 years agoHi all
Bruce Momjian [Tue, 6 Oct 1998 03:09:02 +0000 (03:09 +0000)]
Hi all

Is it too late to add a feature to pg_dump for 6.4??

I just spent most of the day learning pg_dump and modifing it so it
would
dump views also.

This is the first time I have ever contributed any code changes, so I'm
not sure of how to submit it.

The diff's and a readme as a tgz file are attached.

Thanks
Terry Mackintosh <terry@terrym.com>          http://www.terrym.com

26 years agoHere are the patches against the current source tree. I have run the
Bruce Momjian [Tue, 6 Oct 1998 03:02:29 +0000 (03:02 +0000)]
Here are the patches against the current source tree. I have run the
regression test on a FreeBSD box with both non-MULTIBYTE and
MULTIBYTE-enabled, and confirmed that the results are same.

However I do not tested on PCs(I don't have access to win). Please let
me know if the patches break anything on PCs.

Also please note that the patch for varchar.c is a fix for a nasty bug
of char(n) types that I introduced and I believe at least this should
be applied.

Tatsuo Ishii

26 years agoSubstantial rewrite of async.c to avoid problems with non-reentrant stdio
Tom Lane [Tue, 6 Oct 1998 02:40:09 +0000 (02:40 +0000)]
Substantial rewrite of async.c to avoid problems with non-reentrant stdio
and possibly other problems.  Minor changes in xact.c and postgres.c's
main loop to support new handling of async NOTIFY.

26 years agoAdd configure test to make sure fcntl(SETLK) is available.
Tom Lane [Tue, 6 Oct 1998 02:32:28 +0000 (02:32 +0000)]
Add configure test to make sure fcntl(SETLK) is available.

26 years agoAdd configure test to make sure fcntl(SETLK) is available,
Tom Lane [Tue, 6 Oct 1998 02:31:42 +0000 (02:31 +0000)]
Add configure test to make sure fcntl(SETLK) is available,
and make backend/libpq/pqcomm.c only try to lock the socket file when
the call exists.  Also, change open-RDONLY to open-WRONLY; at least
on my platform, you can't get a write lock on a file you didn't open
for writing.

26 years agoMany updates...
Tom Lane [Mon, 5 Oct 1998 04:11:47 +0000 (04:11 +0000)]
Many updates...

26 years agoInformation is now in the chapter on syntax in the User's Guide.
Thomas G. Lockhart [Mon, 5 Oct 1998 02:55:58 +0000 (02:55 +0000)]
Information is now in the chapter on syntax in the User's Guide.

26 years agoAdd new information for utility commands.
Thomas G. Lockhart [Mon, 5 Oct 1998 02:54:45 +0000 (02:54 +0000)]
Add new information for utility commands.
 Haven't yet done cleardbdir, ipcclean, pg_passwd, and pg_upgrade.
Add reference info for the SQL VACUUM command (oops, forgot it earlier).

26 years agoProperly interpret environment variables passed as the input location.
Thomas G. Lockhart [Mon, 5 Oct 1998 02:51:21 +0000 (02:51 +0000)]
Properly interpret environment variables passed as the input location.

26 years agoFix up warning and error messages to use single-quotes around strings.
Thomas G. Lockhart [Mon, 5 Oct 1998 02:49:36 +0000 (02:49 +0000)]
Fix up warning and error messages to use single-quotes around strings.

26 years agoFix up warning and error messages to use single-quotes aroun strings.
Thomas G. Lockhart [Mon, 5 Oct 1998 02:48:49 +0000 (02:48 +0000)]
Fix up warning and error messages to use single-quotes aroun strings.

26 years agoAdd note explaining that NoticeResponse messages can
Tom Lane [Sun, 4 Oct 1998 22:48:42 +0000 (22:48 +0000)]
Add note explaining that NoticeResponse messages can
arrive even while idle.

26 years agoDisable doubling of backslashes in PQprint, per discussion
Tom Lane [Sun, 4 Oct 1998 20:46:39 +0000 (20:46 +0000)]
Disable doubling of backslashes in PQprint, per discussion
on pgsql-interfaces a couple months ago.

26 years agohash.h's use of BSHIFT conflicts with <sys/param.h> on my
Tom Lane [Sun, 4 Oct 1998 20:19:08 +0000 (20:19 +0000)]
hash.h's use of BSHIFT conflicts with <sys/param.h> on my
system.  Finally got tired of looking at the compiler warning messages.
BSHIFT isn't all that useful, so I just took out the macro.

26 years agoOur 'install-sh' script is BSD-style, therefore requires -c
Tom Lane [Sun, 4 Oct 1998 18:28:10 +0000 (18:28 +0000)]
Our 'install-sh' script is BSD-style, therefore requires -c
to behave sanely.  configure was not treating it as BSD...

26 years agoAdd includes for prototypes for new IP type.
Bruce Momjian [Sun, 4 Oct 1998 16:24:32 +0000 (16:24 +0000)]
Add includes for prototypes for new IP type.

26 years agoUnixware patches from Billy G. Allie.
Bruce Momjian [Sun, 4 Oct 1998 15:38:58 +0000 (15:38 +0000)]
Unixware patches from Billy G. Allie.

26 years agoAdd prototype include to inet types.
Bruce Momjian [Sun, 4 Oct 1998 15:35:12 +0000 (15:35 +0000)]
Add prototype include to inet types.

26 years agoRemove u_int32_t
Bruce Momjian [Sun, 4 Oct 1998 15:31:09 +0000 (15:31 +0000)]
Remove u_int32_t

26 years agocnf'ify cleanup
Bruce Momjian [Sun, 4 Oct 1998 03:30:56 +0000 (03:30 +0000)]
cnf'ify cleanup

26 years agoIntegrate new IP type from Tom Ivar Helbekkmo.
Bruce Momjian [Sat, 3 Oct 1998 05:41:01 +0000 (05:41 +0000)]
Integrate new IP type from Tom Ivar Helbekkmo.

26 years agoChanges from Michael Meskes:
Thomas G. Lockhart [Sat, 3 Oct 1998 02:33:51 +0000 (02:33 +0000)]
Changes from Michael Meskes:
Check strdup calls for out of memory.
Set library version to 2.6.2
Synced preproc.y and keywords.c with gram.y and keywords.c yet again.
Set version to 2.4.3

26 years agoUpdate for newest changes.
Thomas G. Lockhart [Sat, 3 Oct 1998 02:30:47 +0000 (02:30 +0000)]
Update for newest changes.

26 years ago Please apply the patch at the end. Disables use of system
Bruce Momjian [Fri, 2 Oct 1998 21:53:39 +0000 (21:53 +0000)]
Please apply the patch at the end. Disables use of system
    columns of views at all (not only oid, cmin etc. too).
    pgsql=> select cmin from pg_rules;
    ERROR:  system column cmin not available - pg_rules is a view
    pgsql=> select * from pg_rules where pg_rules.oid = pg_class.oid;
    ERROR:  system column oid not available - pg_rules is a view
    pgsql=>

Jan

26 years agoObsolete information completely superceded by the sgml sources.
Thomas G. Lockhart [Fri, 2 Oct 1998 16:46:34 +0000 (16:46 +0000)]
Obsolete information completely superceded by the sgml sources.
These weren't really "man page"-ish anyway, and I've verified that
 all information in them has moved to the newer sources.

26 years agoSurround all identifiers with double quotes.
Thomas G. Lockhart [Fri, 2 Oct 1998 16:43:41 +0000 (16:43 +0000)]
Surround all identifiers with double quotes.
 Formerly did so only for those which clearly required it, but that
 would still miss things like reserved key words which also require it.
Implement the "-n" switch to revert the double quote behavior
 to put DQs only where there is more than lower-case, digits,
 and underscores.

26 years ago Here's a combination of all the patches I'm currently waiting
Bruce Momjian [Fri, 2 Oct 1998 16:28:04 +0000 (16:28 +0000)]
Here's a combination of all the patches I'm currently waiting
    for against a just updated CVS tree. It contains

        Partial new rewrite system that handles subselects,  view
        aggregate  columns, insert into select from view, updates
        with set col = view-value and select rules restriction to
        view definition.

        Updates  for  rule/view  backparsing utility functions to
        handle subselects correct.

        New system views pg_tables and pg_indexes (where you  can
        see the complete index definition in the latter one).

        Enabling array references on query parameters.

        Bugfix for functional index.

        Little changes to system views pg_rules and pg_views.

    The rule system isn't a release-stopper any longer.

    But  another  stopper  is  that  I  don't  know if the latest
    changes to PL/pgSQL (not already in CVS) made it  compile  on
    AIX. Still wait for some response from Dave.

Jan

26 years ago the following little patch adds array references to query
Bruce Momjian [Fri, 2 Oct 1998 16:23:07 +0000 (16:23 +0000)]
the  following  little  patch  adds array references to query
    parameters. With it applied a function like

    CREATE FUNCTION getname(oid8, int4) RETURNS name AS
        'SELECT typname FROM pg_type WHERE oid = $1[$2]'
        LANGUAGE 'sql';

    is possible. Mainly I need this to enable array references in
    expressions for PL/pgSQL. Complete regression test ran O.K.

Jan

26 years agoSummary
Bruce Momjian [Fri, 2 Oct 1998 16:18:20 +0000 (16:18 +0000)]
Summary

The ident() function in src/backend/libpq/hba.c doesn't cope when
postmaster is contacted on an IP alias. This patch fixes it.

 Malcolm Beattie

26 years agoFixes for Irix from Robert Bruccoleri
Bruce Momjian [Fri, 2 Oct 1998 15:38:04 +0000 (15:38 +0000)]
Fixes for Irix from Robert Bruccoleri

26 years agoHere's the new diff for getting the NS32K assembly code into the
Bruce Momjian [Fri, 2 Oct 1998 15:23:08 +0000 (15:23 +0000)]
Here's the new diff for getting the NS32K assembly code into the
spin-locks.  Notice that it's now inline assembler in s_lock.h,
rather than seperate code in s_lock.c.  It also shrank a little
bit...  Just rip out the S_LOCK() define and insert the tas() inline
function.  Please let me know if there are any problems with it.

Jon Buller

26 years agoregproc cleanups
Bruce Momjian [Fri, 2 Oct 1998 05:31:28 +0000 (05:31 +0000)]
regproc cleanups

26 years agoFix for regproc so proc name can be supplied if unique, if not, oid.
Bruce Momjian [Fri, 2 Oct 1998 05:10:11 +0000 (05:10 +0000)]
Fix for regproc so proc name can be supplied if unique, if not, oid.

26 years agoFix potential coredump in pg_conndefaults (assigning constant string
Tom Lane [Fri, 2 Oct 1998 01:37:17 +0000 (01:37 +0000)]
Fix potential coredump in pg_conndefaults (assigning constant string
to a field that will get free'd).  Also make it robust in cases where
values contain Tcl special characters.

26 years agoSearch contrib/tcl for tcl startup.
Bruce Momjian [Fri, 2 Oct 1998 01:22:43 +0000 (01:22 +0000)]
Search contrib/tcl for tcl startup.

26 years agoClean up some minor bugs concerning what was inside the main loop
Tom Lane [Fri, 2 Oct 1998 01:14:14 +0000 (01:14 +0000)]
Clean up some minor bugs concerning what was inside the main loop
and what wasn't.  Also try to improve the comments so that doesn't happen
again.  Changed SIGPIPE handling to SIG_IGN so that if frontend quits,
we will finish out the current command and return to main loop before
quitting.  This seems much safer than a forced abort mid-command.

26 years agoFix for constbyval.
Bruce Momjian [Thu, 1 Oct 1998 22:51:22 +0000 (22:51 +0000)]
Fix for constbyval.

26 years agoFix for constbyval .
Bruce Momjian [Thu, 1 Oct 1998 22:45:32 +0000 (22:45 +0000)]
Fix for constbyval .

26 years agofailed to add some of Jan's files :(
Marc G. Fournier [Thu, 1 Oct 1998 03:38:45 +0000 (03:38 +0000)]
failed to add some of Jan's files :(

26 years agoGet rid of some long-dead code that thinks NOTIFY is passed to the
Tom Lane [Thu, 1 Oct 1998 02:04:01 +0000 (02:04 +0000)]
Get rid of some long-dead code that thinks NOTIFY is passed to the
planner/optimizer/executor.  It isn't.
Besides, most of the removed code consists of comments about how
it's not right.

26 years agoMake HP-PA S_UNLOCK a little faster and less dependent on unspecified compiler codege...
Tom Lane [Thu, 1 Oct 1998 01:53:50 +0000 (01:53 +0000)]
Make HP-PA S_UNLOCK a little faster and less dependent on unspecified compiler codegen details.
Make default S_LOCK macro more robust against syntax mistakes.
(I cleared these changes with David Gould a few days ago.)

26 years agopg_dump -z was careless about ownership of indexes.
Tom Lane [Thu, 1 Oct 1998 01:49:12 +0000 (01:49 +0000)]
pg_dump -z was careless about ownership of indexes.
Now it's a little less so.

26 years agoSimplify pg_result by using new libpq PQresultErrorMessage;
Tom Lane [Thu, 1 Oct 1998 01:45:38 +0000 (01:45 +0000)]
Simplify pg_result by using new libpq PQresultErrorMessage;
fix some memory leaks in pg_select.

26 years agoUpdate libpq to store an error message in PGresult, per pgsq-interfaces discussion...
Tom Lane [Thu, 1 Oct 1998 01:40:26 +0000 (01:40 +0000)]
Update libpq to store an error message in PGresult, per pgsq-interfaces discussion of 21-Sep.

26 years agoChange HPUX loader flags to trap null pointer derefs
Tom Lane [Thu, 1 Oct 1998 01:26:45 +0000 (01:26 +0000)]
Change HPUX loader flags to trap null pointer derefs

26 years agoUpdate to track newest gram.y.
Thomas G. Lockhart [Wed, 30 Sep 1998 05:48:28 +0000 (05:48 +0000)]
Update to track newest gram.y.

26 years agoAdd as many keywords as possible to column identifier or label lists.
Thomas G. Lockhart [Wed, 30 Sep 1998 05:47:57 +0000 (05:47 +0000)]
Add as many keywords as possible to column identifier or label lists.
Add "timestamp" to list of tokens in keywords.c.
 Before, TIMESTAMP WITH TIME ZONE did not actually parser.
Reorder token lists to be more alphabetical.
Remove ARCHIVE keyword which was deprecated in v6.3.

26 years agoMessage changed slightly since "rename" is now allowed as a column name.
Thomas G. Lockhart [Wed, 30 Sep 1998 05:43:23 +0000 (05:43 +0000)]
Message changed slightly since "rename" is now allowed as a column name.

26 years agoClean up pages. Add information for operator precedence.
Thomas G. Lockhart [Wed, 30 Sep 1998 05:41:54 +0000 (05:41 +0000)]
Clean up pages. Add information for operator precedence.
Split introduction sections into separate files to allow the legal notice
 and notation sections appear in all documents without having the history
 show up everplace too.
Add full list of reserved and non-reserved key words in syntax.sgml.
Add a separate chapter to the admin guide on security.

26 years agoAdd new SQL reference page and the first utility/app reference page
Thomas G. Lockhart [Wed, 30 Sep 1998 05:38:56 +0000 (05:38 +0000)]
Add new SQL reference page and the first utility/app reference page
 converted from the man page.

26 years agoFrom: Jan Wieck <jwieck@debis.com>
Marc G. Fournier [Tue, 29 Sep 1998 12:43:05 +0000 (12:43 +0000)]
From: Jan Wieck <jwieck@debis.com>

    here  is  the  patch  that  includes  PL/pgSQL into the build
    (currently with make errors ignored) and  adds  a  regression
    test for it. A clean build and regression ran fine here.

    Can you please apply it?

    The  tar  should  be  extracted  in  /usr/local/src/pgsql and
    creates the following files:

    src/pl/Makefile
            called by toplevel GNUmakefile and for now only calls
            src/pl/plpgsql/Makefile

    src/pl/plpgsql/Makefile
            calls  src/pl/plpgsql/src/Makefile  (here the call to
            make ignores build errors  -  this  must  be  changed
            later for the final release).

    src/test/regress/input/install_plpgsql.source
            SQL script installing PL/pgSQL language in regression
            database. Will be modified by  .../input/Makefile  to
            point  to  correct  PGLIB  directory where plpgsql.so
            gets installed.

    src/test/regress/output/install_plpgsql.source
            expected output for installation script.

    src/test/regress/sql/plpgsql.sql
            the main regression  test.  It  tests  functions  and
            triggers written in PL/pgSQL including views that use
            supportfunctions in this language.

    src/test/regress/expected/plpgsql.out
            the expected output for the above regression test.

    make_plpgsql.diff
            patch that adds some lines to

            src/GNUmakefile.in
            src/test/regress/expected/Makefile
            src/test/regress/input/Makefile
            src/test/regress/output/Makefile
            src/test/regress/sql/Makefile
            src/test/regress/sql/tests

26 years agopgsql_perl5-1.8.0
Edmund Mergl [Sun, 27 Sep 1998 19:12:35 +0000 (19:12 +0000)]
pgsql_perl5-1.8.0

26 years agomulti-byte fix from Tatsuo Ishii
Bruce Momjian [Fri, 25 Sep 1998 15:51:02 +0000 (15:51 +0000)]
multi-byte fix from Tatsuo Ishii

26 years agoAdd new Makefile from Jan.
Bruce Momjian [Fri, 25 Sep 1998 15:50:02 +0000 (15:50 +0000)]
Add new Makefile from Jan.

26 years agoBracket filename in LOAD error message with single quotes
Thomas G. Lockhart [Fri, 25 Sep 1998 13:47:27 +0000 (13:47 +0000)]
Bracket filename in LOAD error message with single quotes
 to clarify actual path used.

26 years agoUpdate to track gram.y.
Thomas G. Lockhart [Fri, 25 Sep 1998 13:45:12 +0000 (13:45 +0000)]
Update to track gram.y.
 Had removed PARSEDEBUG statements...

26 years agoCleanup markup and minor editing to prepare for first release.
Thomas G. Lockhart [Fri, 25 Sep 1998 13:42:46 +0000 (13:42 +0000)]
Cleanup markup and minor editing to prepare for first release.

26 years agoInclude the graphics directory in build lookup path.
Thomas G. Lockhart [Fri, 25 Sep 1998 13:41:25 +0000 (13:41 +0000)]
Include the graphics directory in build lookup path.

26 years agoClean up existing debugging print statements.
Thomas G. Lockhart [Fri, 25 Sep 1998 13:38:32 +0000 (13:38 +0000)]
Clean up existing debugging print statements.
Modify comment blocks to insulate from pgindent.

26 years agoClean up code in analyze.c for SERIAL data type.
Thomas G. Lockhart [Fri, 25 Sep 1998 13:36:08 +0000 (13:36 +0000)]
Clean up code in analyze.c for SERIAL data type.
Remove _all_ PARSEDEBUG print statements.

26 years agoFix comment for regproc.
Bruce Momjian [Fri, 25 Sep 1998 03:36:33 +0000 (03:36 +0000)]
Fix comment for regproc.

26 years agoMB patches from Tatsuo Ishii
Bruce Momjian [Fri, 25 Sep 1998 01:46:25 +0000 (01:46 +0000)]
MB patches from Tatsuo Ishii

26 years agoFix for datetime from Tatsuo Ishii
Bruce Momjian [Wed, 23 Sep 1998 17:51:46 +0000 (17:51 +0000)]
Fix for datetime from Tatsuo Ishii

26 years agoFix for oidArray from Christopher Oliver.
Bruce Momjian [Wed, 23 Sep 1998 17:50:12 +0000 (17:50 +0000)]
Fix for oidArray from Christopher Oliver.

26 years agoAllow 8-key indexes.
Bruce Momjian [Wed, 23 Sep 1998 04:22:14 +0000 (04:22 +0000)]
Allow 8-key indexes.

26 years agoAdd in, I think, support for %lld in snprintf(), specifically with
Marc G. Fournier [Wed, 23 Sep 1998 03:11:24 +0000 (03:11 +0000)]
Add in, I think, support for %lld in snprintf(), specifically with
Irix in mind...

26 years agoFix for SELECT NOT boolfield FROM table
Bruce Momjian [Tue, 22 Sep 1998 21:48:27 +0000 (21:48 +0000)]
Fix for SELECT NOT boolfield FROM table

26 years agoCleanup for oid8[] from Tatsuo Ishii.
Bruce Momjian [Tue, 22 Sep 1998 20:28:15 +0000 (20:28 +0000)]
Cleanup for oid8[] from Tatsuo Ishii.

26 years agoHere are some patches to fix up the regression tests so that the union
Bruce Momjian [Tue, 22 Sep 1998 16:52:59 +0000 (16:52 +0000)]
Here are some patches to fix up the regression tests so that the union
test passes.  Interestingly, the fix involves no changes or special
cases in the union test and actually removes a special case for the
numerology test.  Thus, following the strategy outlined below is a
definite improvement over the previous situation.

Cheers,
Brook

26 years agoAdd .h to string include.
Bruce Momjian [Tue, 22 Sep 1998 16:41:19 +0000 (16:41 +0000)]
Add .h to string include.

26 years agoEditing and markup cleanup.
Thomas G. Lockhart [Tue, 22 Sep 1998 15:48:03 +0000 (15:48 +0000)]
Editing and markup cleanup.

26 years agoFix for AND/OR handling.
Bruce Momjian [Mon, 21 Sep 1998 15:41:28 +0000 (15:41 +0000)]
Fix for AND/OR handling.

26 years agoFrom: Michael Meskes <meskes@usa.net>
Marc G. Fournier [Mon, 21 Sep 1998 05:55:23 +0000 (05:55 +0000)]
From: Michael Meskes <meskes@usa.net>

docs improvement for ecpg..

26 years agoFrom: Michael Meskes <meskes@usa.net>
Marc G. Fournier [Mon, 21 Sep 1998 05:52:55 +0000 (05:52 +0000)]
From: Michael Meskes <meskes@usa.net>

+ Mon Aug 31 09:40:04 CEST 1998
+
+       - Minor patch to Makefile
+       - Put pgc.l in sync with scan.l
+
+ Tue Sep  1 11:31:05 CEST 1998
+
+       - Fixed another bug in preproc.y
+
+ Thu Sep  3 12:21:16 CEST 1998
+
+       - Sync preproc.y with gram.y
+
+ Mon Sep 14 09:21:02 CEST 1998
+
+       - Sync preproc.y with gram.y yet again
+
+ Thu Sep 17 08:55:33 CEST 1998
+
+       - Synced preproc.y and gram.y one more time
+
+ Thu Sep 17 19:23:24 CEST 1998
+
+       - Added missing ´;´ in preproc.y
+       - Set version to 2.4.2

26 years agoFix for \z formatting from Tom Lane.
Bruce Momjian [Mon, 21 Sep 1998 02:25:29 +0000 (02:25 +0000)]
Fix for \z formatting from Tom Lane.

26 years agoThis patch covers several to-do items that I had for libpgtcl:
Bruce Momjian [Mon, 21 Sep 1998 01:02:04 +0000 (01:02 +0000)]
This patch covers several to-do items that I had for libpgtcl:

* It works under both Tcl 7.6 and Tcl 8.0 now.  (The code claims to
  work under Tcl 7.5 as well, but I have no way to test that ---
  if anyone still cares, please check it with 7.5.)

* pg_listen suppresses extra LISTEN commands and correctly sends an
  UNLISTEN when the last listen request for a relation is cancelled.
  (Note this means it will not work with pre-6.4 backends, but that
  was true already because it depends on the current libpq, which
  only speaks protocol 2.0.)

* Added -error option to pg_result so that there's some way to find
  out what you did wrong ;-)

* Miscellaneous cleanups of code comments and overenthusiastic #includes.

BTW, I bumped the package version number from 1.2 to 1.3.  Is this
premature?  Does someone run around and do that routinely before
each pgsql release?

regards, tom lane

26 years agoautoconf
Bruce Momjian [Sun, 20 Sep 1998 05:01:04 +0000 (05:01 +0000)]
autoconf

26 years agoFix for memory leak, configure path fix, from Tom Lane.
Bruce Momjian [Sun, 20 Sep 1998 04:57:40 +0000 (04:57 +0000)]
Fix for memory leak, configure path fix, from Tom Lane.

26 years agoFix memory leak from Tom Lane.
Bruce Momjian [Sun, 20 Sep 1998 04:51:12 +0000 (04:51 +0000)]
Fix memory leak from Tom Lane.

26 years agoI also noticed that pg_dump contains a copy of the same
Bruce Momjian [Sun, 20 Sep 1998 03:18:43 +0000 (03:18 +0000)]
I also noticed that pg_dump contains a copy of the same
prompt_for_password code that psql does.  We fixed psql a month or
two back to permit usernames and passwords longer than 8 characters.
I propagated the same fix into pg_dump.

Tom Lane

26 years agoRegression test cleanup.
Bruce Momjian [Fri, 18 Sep 1998 21:47:27 +0000 (21:47 +0000)]
Regression test cleanup.

26 years agoNS32K lock patch reversed.
Bruce Momjian [Fri, 18 Sep 1998 17:18:41 +0000 (17:18 +0000)]
NS32K lock patch reversed.

26 years agoFixes for VC and libpq.
Bruce Momjian [Fri, 18 Sep 1998 16:46:07 +0000 (16:46 +0000)]
Fixes for VC and libpq.

Magnus Hagander

26 years agoFix for NS32K machine.
Bruce Momjian [Fri, 18 Sep 1998 05:36:24 +0000 (05:36 +0000)]
Fix for NS32K machine.

26 years agofix for aix snprintf
Bruce Momjian [Fri, 18 Sep 1998 05:20:22 +0000 (05:20 +0000)]
fix for aix snprintf

26 years agoRemove /usr/local/* from templates.
Bruce Momjian [Fri, 18 Sep 1998 05:00:52 +0000 (05:00 +0000)]
Remove /usr/local/* from templates.

26 years agoFix for \df.
Bruce Momjian [Fri, 18 Sep 1998 04:43:20 +0000 (04:43 +0000)]
Fix for \df.

26 years agoNew gram.c
Bruce Momjian [Fri, 18 Sep 1998 03:02:47 +0000 (03:02 +0000)]
New gram.c

26 years agoUpdate entries for MkLinux and Linux/ix86 for Postgres v6.4.
Thomas G. Lockhart [Wed, 16 Sep 1998 14:48:41 +0000 (14:48 +0000)]
Update entries for MkLinux and Linux/ix86 for Postgres v6.4.

26 years agoAdd Unix-specific installation and configuration info for psqlodbc.
Thomas G. Lockhart [Wed, 16 Sep 1998 14:44:00 +0000 (14:44 +0000)]
Add Unix-specific installation and configuration info for psqlodbc.

26 years agoMarkup and editing adjustments...
Thomas G. Lockhart [Wed, 16 Sep 1998 14:43:12 +0000 (14:43 +0000)]
Markup and editing adjustments...

26 years agoConsolidate Jan's rules test into one file.
Thomas G. Lockhart [Wed, 16 Sep 1998 14:35:37 +0000 (14:35 +0000)]
Consolidate Jan's rules test into one file.
Remove dependency on a specific Postgres user name in the results
 (Check result against CURRENT_USER with a boolean instead).

26 years agoSupport specifying PRIMARY KEY for the SERIAL type.
Thomas G. Lockhart [Wed, 16 Sep 1998 14:29:35 +0000 (14:29 +0000)]
Support specifying PRIMARY KEY for the SERIAL type.
Change DEFAULT NULL to send back a NULL pointer
 rather than a string "NULL". This seems to work, where sending
 the string led to type conversion problems (and probably the wrong
 thing anyway).