OSDN Git Service

pg-rex/syncrep.git
26 years agoFix for views and outnodes.
Bruce Momjian [Wed, 7 Jan 1998 08:08:08 +0000 (08:08 +0000)]
Fix for views and outnodes.

26 years agoCleanup of outnode stuff.
Bruce Momjian [Wed, 7 Jan 1998 05:54:11 +0000 (05:54 +0000)]
Cleanup of outnode stuff.

26 years agoCleanup of outnode stuff.
Bruce Momjian [Wed, 7 Jan 1998 05:50:41 +0000 (05:50 +0000)]
Cleanup of outnode stuff.

26 years agoCleanup of outnode stuff.
Bruce Momjian [Wed, 7 Jan 1998 05:42:47 +0000 (05:42 +0000)]
Cleanup of outnode stuff.

26 years agoChange web page directions.
Bruce Momjian [Wed, 7 Jan 1998 03:48:29 +0000 (03:48 +0000)]
Change web page directions.

26 years agoShort little FAQ for FreeBSD
Marc G. Fournier [Wed, 7 Jan 1998 02:33:46 +0000 (02:33 +0000)]
Short little FAQ for FreeBSD

26 years agoFix sorting of multiple fields broken with UNION.
Bruce Momjian [Tue, 6 Jan 1998 23:58:05 +0000 (23:58 +0000)]
Fix sorting of multiple fields broken with UNION.

26 years agoCleanup of sort nodes and use of strtoul .
Bruce Momjian [Tue, 6 Jan 1998 23:19:49 +0000 (23:19 +0000)]
Cleanup of sort nodes and use of strtoul .

26 years agoABORT to ERROR in catalog/.
Bruce Momjian [Tue, 6 Jan 1998 19:42:33 +0000 (19:42 +0000)]
ABORT to ERROR in catalog/.

26 years agoRegression cleanup of WARN.
Bruce Momjian [Tue, 6 Jan 1998 19:25:19 +0000 (19:25 +0000)]
Regression cleanup of WARN.

26 years agoChange some labels in bootparse to make ctags happy. Clean up outfunc/readfunc code...
Bruce Momjian [Tue, 6 Jan 1998 18:53:02 +0000 (18:53 +0000)]
Change some labels in bootparse to make ctags happy.  Clean up outfunc/readfunc code and add missing fields for Query structure and new Union fields.  Fix optimizer bug shown in new \do command.  Change WARN to ERROR in contrib and regression stuff.

26 years agoMove variable.c to commands/ and aclchk.c to catalog/.
Bruce Momjian [Mon, 5 Jan 1998 18:53:12 +0000 (18:53 +0000)]
Move variable.c to commands/ and aclchk.c to catalog/.

26 years agoMove variable.c to commands/ and aclchk.c to catalog/.
Bruce Momjian [Mon, 5 Jan 1998 18:43:18 +0000 (18:43 +0000)]
Move variable.c to commands/ and aclchk.c to catalog/.

26 years agoChange some ABORTS to ERROR. Add line number when COPY Failure.
Bruce Momjian [Mon, 5 Jan 1998 16:40:20 +0000 (16:40 +0000)]
Change some ABORTS to ERROR.  Add line number when COPY Failure.

26 years agoRemove un-needed quotes from psql \d DEFAULT display.
Bruce Momjian [Mon, 5 Jan 1998 13:56:05 +0000 (13:56 +0000)]
Remove un-needed quotes from psql \d DEFAULT display.

26 years agoChange elog(WARN) to elog(ERROR) and elog(ABORT).
Bruce Momjian [Mon, 5 Jan 1998 04:10:07 +0000 (04:10 +0000)]
Change elog(WARN) to elog(ERROR) and elog(ABORT).

26 years agoChange elog(WARN) to elog(ERROR) and elog(ABORT).
Bruce Momjian [Mon, 5 Jan 1998 03:45:54 +0000 (03:45 +0000)]
Change elog(WARN) to elog(ERROR) and elog(ABORT).

26 years agoChange elog(WARN) to elog(ERROR) and elog(ABORT).
Bruce Momjian [Mon, 5 Jan 1998 03:35:55 +0000 (03:35 +0000)]
Change elog(WARN) to elog(ERROR) and elog(ABORT).

26 years agoAdd NOT NULL and DEFAULT to \d table.
Bruce Momjian [Mon, 5 Jan 1998 02:21:22 +0000 (02:21 +0000)]
Add NOT NULL and DEFAULT to \d table.

26 years agoVarious fixes NETBSD/sparc related
Marc G. Fournier [Sun, 4 Jan 1998 19:13:04 +0000 (19:13 +0000)]
Various fixes NETBSD/sparc related

From: Tom I Helbekkmo <tih@Hamartun.Priv.NO>

26 years agoFix for aggs on views and complex ones. I missed one file.
Bruce Momjian [Sun, 4 Jan 1998 04:53:50 +0000 (04:53 +0000)]
Fix for aggs on views and complex ones.  I missed one file.

26 years agoFix for count(*), aggs with views and multiple tables and sum(3).
Bruce Momjian [Sun, 4 Jan 1998 04:31:43 +0000 (04:31 +0000)]
Fix for count(*), aggs with views and multiple tables and sum(3).

26 years agoneed to include postgres.h before HAVE_DLD_H
Marc G. Fournier [Fri, 2 Jan 1998 03:40:04 +0000 (03:40 +0000)]
need to include postgres.h before HAVE_DLD_H
need to include string.h for strdup()

From: Bryan Henderson <bryanh@giraffe-data.com>

26 years agoTest for __ELF__ rather than LINUX_ELF in headers.
Thomas G. Lockhart [Thu, 1 Jan 1998 06:02:04 +0000 (06:02 +0000)]
Test for __ELF__ rather than LINUX_ELF in headers.
Comment-out dynamic link function declarations since they are all
 provided by the system.
Should we bother continuing to support non-elf Linux systems??

26 years agoRemove declaration for -DLINUX_ELF since testing for __ELF__ in code.
Thomas G. Lockhart [Thu, 1 Jan 1998 05:59:24 +0000 (05:59 +0000)]
Remove declaration for -DLINUX_ELF since testing for __ELF__ in code.

26 years agoDefine text_substr().
Thomas G. Lockhart [Thu, 1 Jan 1998 05:50:50 +0000 (05:50 +0000)]
Define text_substr().
Change declarations from "struct varlena *" to "text *".
Remove register variables since compiler should do better on its own.

26 years agoCoerce a function argument to avoid compiler warning.
Thomas G. Lockhart [Thu, 1 Jan 1998 05:48:14 +0000 (05:48 +0000)]
Coerce a function argument to avoid compiler warning.

26 years agoInclude text_substr() function and use instead of oracle_compat equivalent.
Thomas G. Lockhart [Thu, 1 Jan 1998 05:47:11 +0000 (05:47 +0000)]
Include text_substr() function and use instead of oracle_compat equivalent.

26 years agoChange precedence for boolean operators to match expected behavior.
Thomas G. Lockhart [Thu, 1 Jan 1998 05:44:54 +0000 (05:44 +0000)]
Change precedence for boolean operators to match expected behavior.
Change NOTICE/NOTICE/NOTICE/WARN elog messages to a single message
 with a few newline/tab breaks embedded in the string. Much cleaner I hope.

26 years agoAdd substring function for text based on oracle_compat package but
Thomas G. Lockhart [Thu, 1 Jan 1998 05:42:40 +0000 (05:42 +0000)]
Add substring function for text based on oracle_compat package but
 closer to standard behavior.
Substitute "text *" for "struct varlena *" in declarations.

26 years agoBracket #define sigsetjmp() to avoid redefinition. Linux starts out with
Thomas G. Lockhart [Thu, 1 Jan 1998 05:40:28 +0000 (05:40 +0000)]
Bracket #define sigsetjmp() to avoid redefinition.  Linux starts out with
 sigsetjmp as a macro, so gives compiler warning without bracketing.

26 years agoInclude SPI module when cleaning.
Thomas G. Lockhart [Thu, 1 Jan 1998 05:37:29 +0000 (05:37 +0000)]
Include SPI module when cleaning.

26 years agoMissed a subst() for the RANDOM substritution...
Marc G. Fournier [Wed, 31 Dec 1997 03:51:17 +0000 (03:51 +0000)]
Missed a subst() for the RANDOM substritution...

26 years agoChange @RANDOM@ to @MISSING_RANDOM@ because of bash
Marc G. Fournier [Tue, 30 Dec 1997 04:08:55 +0000 (04:08 +0000)]
Change @RANDOM@ to @MISSING_RANDOM@ because of bash

26 years agoOops, missed adding a file
Marc G. Fournier [Tue, 30 Dec 1997 04:03:01 +0000 (04:03 +0000)]
Oops, missed adding a file

From: Ryan Kirkpatrick <rkirkpat@nag.cs.colorado.edu>

26 years agobring in alpha/linux s_lock changes
Marc G. Fournier [Tue, 30 Dec 1997 04:01:28 +0000 (04:01 +0000)]
bring in alpha/linux s_lock changes

From: Ryan Kirkpatrick <rkirkpat@nag.cs.colorado.edu>

26 years agoSlightly delayed patches from Todd...damn holidays :)
Marc G. Fournier [Tue, 30 Dec 1997 02:26:56 +0000 (02:26 +0000)]
Slightly delayed patches from Todd...damn holidays :)

  o  A new patch that contains the following changes:
        -- The pg_pwd file is now cached in the postmaster's memory.
        -- pg_pwd is reloaded when the postmaster detects a flag file creat()'ed
           by a backend.
        -- qsort() is used to sort loaded password entries, and bsearch() is
           is used to find entries in the pg_pwd cache.
        -- backends now copy the pg_user relation to pg_pwd.pid, and then
           rename the temp file to be pg_pwd.
        -- The delimiter for pg_pwd has been changed to a tab character.

26 years agoChange some mallocs to palloc.
Bruce Momjian [Mon, 29 Dec 1997 05:13:57 +0000 (05:13 +0000)]
Change some mallocs to palloc.

26 years agoFix for ORDER BY in UNION.
Bruce Momjian [Mon, 29 Dec 1997 04:31:50 +0000 (04:31 +0000)]
Fix for ORDER BY in UNION.

26 years agoUNION cleanup again.
Bruce Momjian [Mon, 29 Dec 1997 02:09:54 +0000 (02:09 +0000)]
UNION cleanup again.

26 years agoCleanup of UNION ALL fix. Manual page updates.
Bruce Momjian [Mon, 29 Dec 1997 01:13:37 +0000 (01:13 +0000)]
Cleanup of UNION ALL fix.  Manual page updates.

26 years agoUNION work for UNION ALL and other union stuff.
Bruce Momjian [Sat, 27 Dec 1997 06:41:41 +0000 (06:41 +0000)]
UNION work for UNION ALL and other union stuff.

26 years agoKnew I missed somewhere...remove PORTNAME references from GNUmakefile.in, since
Marc G. Fournier [Fri, 26 Dec 1997 22:35:53 +0000 (22:35 +0000)]
Knew I missed somewhere...remove PORTNAME references from GNUmakefile.in, since
PORTNAME should not be used anywhere anymore...I hope :)

26 years agoPrototypes for new funcs.
Vadim B. Mikheev [Fri, 26 Dec 1997 08:45:27 +0000 (08:45 +0000)]
Prototypes for new funcs.

26 years agoCleaning GCC warnings.
Vadim B. Mikheev [Fri, 26 Dec 1997 06:02:26 +0000 (06:02 +0000)]
Cleaning GCC warnings.

26 years agoextern char* crypt_getpwdfilename(void);
Vadim B. Mikheev [Fri, 26 Dec 1997 05:51:51 +0000 (05:51 +0000)]
extern char* crypt_getpwdfilename(void);
                                  ^^^^ gcc likes this

26 years agoImplementation of UNIONs.
Bruce Momjian [Wed, 24 Dec 1997 06:06:58 +0000 (06:06 +0000)]
Implementation of UNIONs.

26 years agoReverse out Thomas's group clause change.
Bruce Momjian [Tue, 23 Dec 1997 21:49:03 +0000 (21:49 +0000)]
Reverse out Thomas's group clause change.

26 years agoMake no-tty not use quiet in psql, fix group by copy failure, fix ccsym to delete...
Bruce Momjian [Tue, 23 Dec 1997 21:38:53 +0000 (21:38 +0000)]
Make no-tty not use quiet in psql, fix group by copy failure, fix ccsym to delete tmp files.

26 years agoAdd newline/tab to middle of long failure message to pretty it up a bit.
Thomas G. Lockhart [Tue, 23 Dec 1997 20:00:06 +0000 (20:00 +0000)]
Add newline/tab to middle of long failure message to pretty it up a bit.

26 years agoChange field name in RetrieveStmt from selectClause to unionClause.
Thomas G. Lockhart [Tue, 23 Dec 1997 19:58:12 +0000 (19:58 +0000)]
Change field name in RetrieveStmt from selectClause to unionClause.
Add unionall boolean field to SubSelect structure.

26 years agoLengthen some time conversion routine names now that longer names
Thomas G. Lockhart [Tue, 23 Dec 1997 19:55:51 +0000 (19:55 +0000)]
Lengthen some time conversion routine names now that longer names
 (> 16 characters) are allowed in pg_proc (thanks Bruce!).

26 years agoFix numeric comparision (... != NULL) to be (... > 0). gcc is happier now.
Thomas G. Lockhart [Tue, 23 Dec 1997 19:53:30 +0000 (19:53 +0000)]
Fix numeric comparision (... != NULL) to be (... > 0). gcc is happier now.
Add non-working code which is commented-out to copy GroupClause structures.

26 years agoAdd routines to print AExpr, Ident, and AConst parsing structures.
Thomas G. Lockhart [Tue, 23 Dec 1997 19:50:54 +0000 (19:50 +0000)]
Add routines to print AExpr, Ident, and AConst parsing structures.

26 years agoImplement CREATE TABLE ... AS SELECT borrowing code from SubSelect
Thomas G. Lockhart [Tue, 23 Dec 1997 19:47:32 +0000 (19:47 +0000)]
Implement CREATE TABLE ... AS SELECT borrowing code from SubSelect
 and from SELECT ... INTO ... support code.
Allow NOT, IS NULL, IS NOT NULL in constraints.
Define unionall boolean flag in SubSelect structure.
Implement row descriptors: (a, b, c) = (x, y, z).
Change IS TRUE, IS FALSE, etc. to expressions using "=" rather than
 function calls to istrue() or isfalse() to allow optimization.
Force type for TRUE and FALSE to bool.

26 years agoChange variable name reference from selectClause to more appropriate unionClause.
Thomas G. Lockhart [Tue, 23 Dec 1997 19:39:42 +0000 (19:39 +0000)]
Change variable name reference from selectClause to more appropriate unionClause.
Add a few lines of test code playing with the unique flag for unions.
 Does not do much helpful yet.

26 years agoFix up some elog error messages.
Thomas G. Lockhart [Tue, 23 Dec 1997 19:36:20 +0000 (19:36 +0000)]
Fix up some elog error messages.

26 years agoExpand a few date/time routine names to > 16 characters now that this
Thomas G. Lockhart [Tue, 23 Dec 1997 19:29:15 +0000 (19:29 +0000)]
Expand a few date/time routine names to > 16 characters now that this
 is legal. Fix a little whitespace in other declarations.

26 years agoFix format to produce actual German style. Affects new "GERMAN" feature only.
Thomas G. Lockhart [Tue, 23 Dec 1997 19:26:31 +0000 (19:26 +0000)]
Fix format to produce actual German style. Affects new "GERMAN" feature only.

26 years agoChange logic slightly to avoid one unnecessary calculation. No big deal.
Thomas G. Lockhart [Tue, 23 Dec 1997 19:24:43 +0000 (19:24 +0000)]
Change logic slightly to avoid one unnecessary calculation. No big deal.

26 years agoChange a few routine names back to full length now that pg_proc allows
Thomas G. Lockhart [Tue, 23 Dec 1997 19:23:09 +0000 (19:23 +0000)]
Change a few routine names back to full length now that pg_proc allows
 names over 16 characters. datet_datetime() -> datetime_datetime(), etc.

26 years agoRemove regression in function.
Bruce Momjian [Tue, 23 Dec 1997 03:27:23 +0000 (03:27 +0000)]
Remove regression in function.

26 years agoFix notty output to show status result. -q option still turns it off.
Bruce Momjian [Mon, 22 Dec 1997 20:03:53 +0000 (20:03 +0000)]
Fix notty output to show status result.  -q option still turns it off.

26 years agoFix for test 0 error.
Bruce Momjian [Mon, 22 Dec 1997 05:56:43 +0000 (05:56 +0000)]
Fix for test 0 error.

26 years agoFix for select 1=1 or 2=2, select 1=1 and 2=2, and select sum(2+2).
Bruce Momjian [Mon, 22 Dec 1997 05:42:25 +0000 (05:42 +0000)]
Fix for select 1=1 or 2=2, select 1=1 and 2=2, and select sum(2+2).

26 years agoRemove some recursion in optimizer and clean up some code there.
Bruce Momjian [Sun, 21 Dec 1997 05:18:48 +0000 (05:18 +0000)]
Remove some recursion in optimizer and clean up some code there.

26 years agoPotential fix for Bruce's "test" problem
Marc G. Fournier [Sat, 20 Dec 1997 18:36:26 +0000 (18:36 +0000)]
Potential fix for Bruce's "test" problem

26 years agoFix aggregates on inherited tables.
Bruce Momjian [Sat, 20 Dec 1997 07:59:44 +0000 (07:59 +0000)]
Fix aggregates on inherited tables.

26 years agoMove more to dynloader subdir
Marc G. Fournier [Sat, 20 Dec 1997 04:48:11 +0000 (04:48 +0000)]
Move more  to dynloader subdir

26 years agoMake sure distclean works as well as clean
Marc G. Fournier [Sat, 20 Dec 1997 04:43:20 +0000 (04:43 +0000)]
Make sure distclean works as well as clean

26 years agoMissed a HAVE_
Marc G. Fournier [Sat, 20 Dec 1997 04:26:27 +0000 (04:26 +0000)]
Missed a HAVE_

26 years agofix so that dynloader.h is included instead of port-protos.h
Marc G. Fournier [Sat, 20 Dec 1997 04:13:31 +0000 (04:13 +0000)]
fix so that dynloader.h is included instead of port-protos.h

26 years agoFix use of variable in Makefile.in
Marc G. Fournier [Sat, 20 Dec 1997 04:10:35 +0000 (04:10 +0000)]
Fix use of variable in Makefile.in

26 years agoMake sure configure reflects changes
Marc G. Fournier [Sat, 20 Dec 1997 03:41:06 +0000 (03:41 +0000)]
Make sure configure reflects changes

26 years agoOops, forgot to add what I took away
Marc G. Fournier [Sat, 20 Dec 1997 03:38:20 +0000 (03:38 +0000)]
Oops, forgot to add what I took away

26 years agoMore cleanups...
Marc G. Fournier [Sat, 20 Dec 1997 03:36:35 +0000 (03:36 +0000)]
More cleanups...

26 years agoMore ports switched over
Marc G. Fournier [Sat, 20 Dec 1997 03:23:46 +0000 (03:23 +0000)]
More ports switched over

26 years agoMove over the BSDi port files
Marc G. Fournier [Sat, 20 Dec 1997 03:13:59 +0000 (03:13 +0000)]
Move over the BSDi port files

26 years agoThis should get the linux port to work...
Marc G. Fournier [Sat, 20 Dec 1997 02:36:26 +0000 (02:36 +0000)]
This should get the linux port to work...

26 years agoGet it so that everything else *should* work with my changes...
Marc G. Fournier [Sat, 20 Dec 1997 02:33:56 +0000 (02:33 +0000)]
Get it so that everything else *should* work with my changes...

26 years agoMajor cleanout of PORTNAME variables from Makefiles...bound to screw up
Marc G. Fournier [Sat, 20 Dec 1997 00:29:35 +0000 (00:29 +0000)]
Major cleanout of PORTNAME variables from Makefiles...bound to screw up
some of the ports...

26 years agoFirst clean compile without a "PORTNAME" variable being set...
Marc G. Fournier [Sat, 20 Dec 1997 00:11:13 +0000 (00:11 +0000)]
First clean compile without a "PORTNAME" variable being set...

26 years agoremoved the port-protos.h from bsd a little too quick...remedied now
Marc G. Fournier [Fri, 19 Dec 1997 23:17:42 +0000 (23:17 +0000)]
removed the port-protos.h from bsd a little too quick...remedied now

26 years agoIf this doesn't break something, nothing will *muhahaha*
Marc G. Fournier [Fri, 19 Dec 1997 23:10:16 +0000 (23:10 +0000)]
If this doesn't break something, nothing will *muhahaha*

Just got rid of PORTNAME altogether for FreeBSD and AIX...more to follow

26 years agoThese files will get link'd to 'dynloader.[ch]', from configure, which
Marc G. Fournier [Fri, 19 Dec 1997 22:58:27 +0000 (22:58 +0000)]
These files will get link'd to 'dynloader.[ch]', from configure, which
will then get link'd into SUBSYS.o

26 years agoMore removal of port dependent stuff
Marc G. Fournier [Fri, 19 Dec 1997 22:56:03 +0000 (22:56 +0000)]
More removal of port dependent stuff

26 years agoMore cleanups. Move alot of the prototype definitions straight into
Marc G. Fournier [Fri, 19 Dec 1997 22:47:00 +0000 (22:47 +0000)]
More cleanups.  Move alot of the prototype definitions straight into
config.h, right beside their appropriate HAVE_ #ifdefs...one central
location that is included everywhere...

26 years agoGet rid of PORTNAME from Makefile.global.in...
Marc G. Fournier [Fri, 19 Dec 1997 21:29:28 +0000 (21:29 +0000)]
Get rid of PORTNAME from Makefile.global.in...

26 years agoAdd regression output from FreeBSD
Marc G. Fournier [Fri, 19 Dec 1997 21:23:19 +0000 (21:23 +0000)]
Add regression output from FreeBSD

26 years agoReplace foo/bar to l1/l2.
Bruce Momjian [Fri, 19 Dec 1997 16:54:15 +0000 (16:54 +0000)]
Replace foo/bar to l1/l2.

26 years agoHere is the difference between expected->results of the regression
Marc G. Fournier [Fri, 19 Dec 1997 15:50:18 +0000 (15:50 +0000)]
Here is the difference between expected->results of the regression
tests under sparc_solaris

26 years agoMore cleanups...only need one rusagestub.h
Marc G. Fournier [Fri, 19 Dec 1997 13:35:17 +0000 (13:35 +0000)]
More cleanups...only need one rusagestub.h

26 years agoremove port from OBJS line...
Marc G. Fournier [Fri, 19 Dec 1997 13:28:06 +0000 (13:28 +0000)]
remove port from OBJS line...

26 years agoNice clean compile...still need to run regression tests, but its a major
Marc G. Fournier [Fri, 19 Dec 1997 03:42:14 +0000 (03:42 +0000)]
Nice clean compile...still need to run regression tests, but its a major
start

26 years agoUpdate configure so that it auto-detects and configures for missing
Marc G. Fournier [Fri, 19 Dec 1997 03:08:03 +0000 (03:08 +0000)]
Update configure so that it auto-detects and configures for missing
functions...

26 years agoOkay, that's it for tonight...broke a few ports, most likely, but that
Marc G. Fournier [Fri, 19 Dec 1997 03:06:22 +0000 (03:06 +0000)]
Okay, that's it for tonight...broke a few ports, most likely, but that
is expected...a few ports don't have "ports" anymore, with most things
auto-configured by configure like it should be :)

26 years agoMore cleanups
Marc G. Fournier [Fri, 19 Dec 1997 02:56:35 +0000 (02:56 +0000)]
More cleanups

26 years agoMore cleanups...several ports are major redundancies of other ports
Marc G. Fournier [Fri, 19 Dec 1997 02:46:25 +0000 (02:46 +0000)]
More cleanups...several ports are major redundancies of other ports

26 years agoport.c no longer required...individual functions that were in there are
Marc G. Fournier [Fri, 19 Dec 1997 02:27:20 +0000 (02:27 +0000)]
port.c no longer required...individual functions that were in there are
auto-generated via configure

individual functions in individual files

26 years agoMore cleanups. I can now compile without PORTNAME being defined n
Marc G. Fournier [Fri, 19 Dec 1997 02:09:10 +0000 (02:09 +0000)]
More cleanups.  I can now compile without PORTNAME being defined n
Makefile.global.

End result, if all goes well, should allow for much easier porting, since
there will no longer be a concept of a "port".  Most, if not everything,
*should* be determined by configure, or by the compiler itself.  Still
work to be done though :)