OSDN Git Service

pg-rex/syncrep.git
24 years agoAnother round of those unportable config/build changes :-/
Peter Eisentraut [Sun, 9 Jul 2000 13:14:19 +0000 (13:14 +0000)]
Another round of those unportable config/build changes :-/

* Add option to build with OpenSSL out of the box. Fix thusly exposed
  bit rot. Although it compiles now, getting this to do something
  useful is left as an exercise.

* Fix Kerberos options to defer checking for required libraries until
  all the other libraries are checked for.

* Change default odbcinst.ini and krb5.srvtab path to PREFIX/etc.

* Install work around for Autoconf's install-sh relative path anomaly.
  Get rid of old INSTL_*_OPTS variables, now that we don't need them
  anymore.

* Use `gunzip -c' instead of g?zcat. Reportedly broke on AIX.

* Look for only one of readline.h or readline/readline.h, not both.

* Make check for PS_STRINGS cacheable. Don't test for the header files
  separately.

* Disable fcntl(F_SETLK) test on Linux.

* Substitute the standard GCC warnings set into CFLAGS in configure,
  don't add it on in Makefile.global.

* Sweep through contrib tree to teach makefiles standard semantics.

... and in completely unrelated news:

* Make postmaster.opts arbitrary options-aware. I still think we need to
  save the environment as well.

24 years agoMake view/rule permission checking behave properly with
Tom Lane [Sun, 9 Jul 2000 04:56:32 +0000 (04:56 +0000)]
Make view/rule permission checking behave properly with
subqueries in the rule.

24 years agoFix bogus DatumGetInt32 coercion.
Tom Lane [Sun, 9 Jul 2000 04:17:53 +0000 (04:17 +0000)]
Fix bogus DatumGetInt32 coercion.

24 years agoRemove reference to tag which is not available to the User's Guide
Thomas G. Lockhart [Sat, 8 Jul 2000 16:00:15 +0000 (16:00 +0000)]
Remove reference to tag which is not available to the User's Guide
 when built on its own. UG now builds cleanly.

24 years agoUpdate TODO list.
Bruce Momjian [Sat, 8 Jul 2000 15:04:27 +0000 (15:04 +0000)]
Update TODO list.

24 years agoUpdate TODO list.
Bruce Momjian [Sat, 8 Jul 2000 13:15:14 +0000 (13:15 +0000)]
Update TODO list.

24 years agoSeems we need <errno.h> to be included explicitly now...
Tom Lane [Sat, 8 Jul 2000 05:30:33 +0000 (05:30 +0000)]
Seems we need <errno.h> to be included explicitly now...

24 years agoRemove long-dead support for invoking queries from dynamically loaded
Tom Lane [Sat, 8 Jul 2000 03:04:41 +0000 (03:04 +0000)]
Remove long-dead support for invoking queries from dynamically loaded
backend functions via backend PQexec().  The SPI interface has long
been our only documented way to do this, and the backend pqexec/portal
code is unused and suffering bit-rot.  I'm putting it out of its misery.

24 years agoPut back separate install-bin target ... I was using that,
Tom Lane [Sat, 8 Jul 2000 02:40:27 +0000 (02:40 +0000)]
Put back separate install-bin target ... I was using that,
thank you ...

24 years agoFix misuse of StrNCpy to copy and add null to non-null-terminated data.
Tom Lane [Fri, 7 Jul 2000 21:12:53 +0000 (21:12 +0000)]
Fix misuse of StrNCpy to copy and add null to non-null-terminated data.
Does not work since it fetches one byte beyond the source data, and when
the phase of the moon is wrong, the source data is smack up against the
end of backend memory and you get SIGSEGV.  Don't laugh, this is a fix
for an actual user bug report.

24 years ago- format_type function, in use by psql
Peter Eisentraut [Fri, 7 Jul 2000 19:24:43 +0000 (19:24 +0000)]
- format_type function, in use by psql
- added bigint as synonym of int8
- set typelem of varlen non-array types to 0

24 years agoRename cash_words_out to cash_words.
Bruce Momjian [Fri, 7 Jul 2000 18:49:56 +0000 (18:49 +0000)]
Rename cash_words_out to cash_words.

24 years agoRemove HTML FAQ files that are really just text files.
Bruce Momjian [Fri, 7 Jul 2000 15:41:42 +0000 (15:41 +0000)]
Remove HTML FAQ files that are really just text files.

24 years agoUpdate TODO list.
Bruce Momjian [Fri, 7 Jul 2000 03:24:09 +0000 (03:24 +0000)]
Update TODO list.

24 years agoUpdate TODO list.
Bruce Momjian [Fri, 7 Jul 2000 01:31:16 +0000 (01:31 +0000)]
Update TODO list.

24 years ago|> The Makefile.shlib changes will have to be discussed with other Linux
Bruce Momjian [Fri, 7 Jul 2000 01:23:44 +0000 (01:23 +0000)]
|> The Makefile.shlib changes will have to be discussed with other Linux
|> developers so we are sure it will work on all platforms.

The problem with the current settings is that the linker is called
directly.  This is wrong, it should always be called through the
compiler
driver (the only exception is `ld -r').  This will make sure that the
necessary libraries like libgcc are linked in.

But there is still a different problem with the setting of LDFLAGS_ODBC.
The psqlodbc module defines the functions _init and _fini which are
reserved for the shared library initialisation.  These should be changed
to constructor functions.  Then LDFLAGS_ODBC can be changed to be just
`-lm'.  Btw, why does it use -Bsymbolic?

Andreas Schwab

24 years agoMove FAQ source into main cvs tree.
Bruce Momjian [Thu, 6 Jul 2000 23:14:30 +0000 (23:14 +0000)]
Move FAQ source into main cvs tree.

24 years agoCorrect unsafe use of strcmp(). See rant of same date posted to pghackers.
Tom Lane [Thu, 6 Jul 2000 23:03:37 +0000 (23:03 +0000)]
Correct unsafe use of strcmp().  See rant of same date posted to pghackers.

24 years agoBackend makefile cleanup. make maintainer-clean is now completely
Peter Eisentraut [Thu, 6 Jul 2000 21:33:45 +0000 (21:33 +0000)]
Backend makefile cleanup. make maintainer-clean is now completely
functional.

Handle include file installation in src/include/Makefile

genbki.sh improvements: Don't substitute anything by config.status,
instead pass in AWK and CPP through environment. Change calling
convention to support named output files, so we get to see error
messages on stderr.

Rename bootstrap template files and install them into PREFIX/share.
Update initdb to that effect and other readability improvements
in initdb.

24 years agoAdded comments about the compression algorithm as requested by Tom
Jan Wieck [Thu, 6 Jul 2000 21:02:07 +0000 (21:02 +0000)]
Added comments about the compression algorithm as requested by Tom

Jan

24 years agoPlease find attached a patch for the pg_dump directory which addresses:
Jan Wieck [Thu, 6 Jul 2000 18:39:39 +0000 (18:39 +0000)]
Please find attached a patch for the pg_dump directory which addresses:

- The problems Jan reported

- incompatibility with configure (now uses HAVE_LIBZ instead of HAVE_ZLIB)

- a problem in auto-detecting archive file format on piped archives

Philip Warner

24 years agoBugfix. If toasted tuple containted NULLs, DataFill() was handed
Jan Wieck [Thu, 6 Jul 2000 18:22:45 +0000 (18:22 +0000)]
Bugfix. If toasted tuple containted NULLs, DataFill() was handed
a wrong pointer causing the bitmap overwriting the tuple header.

Jan

24 years agoFunctions on 'text' type updated to new fmgr style. 'text' is
Tom Lane [Thu, 6 Jul 2000 05:48:31 +0000 (05:48 +0000)]
Functions on 'text' type updated to new fmgr style.  'text' is
now TOAST-able.

24 years agoUpdate textin() and textout() to new fmgr style. This is just phase
Tom Lane [Wed, 5 Jul 2000 23:12:09 +0000 (23:12 +0000)]
Update textin() and textout() to new fmgr style.  This is just phase
one of updating the whole text datatype, but there are so dang many
calls of these two routines that it seems worth a separate commit.

24 years agoMake the regression tests safe for TOAST.
Tom Lane [Wed, 5 Jul 2000 23:02:37 +0000 (23:02 +0000)]
Make the regression tests safe for TOAST.

24 years agoFix typo (extraneous semicolon) in fd.c patch to avoid excess seeks.
Tom Lane [Wed, 5 Jul 2000 21:10:05 +0000 (21:10 +0000)]
Fix typo (extraneous semicolon) in fd.c patch to avoid excess seeks.
Now it skips useless SEEK_CUR 0 calls too, as intended.

24 years agoEnable pg_rewrite to be toasted (unlimited size of views and rules).
Jan Wieck [Wed, 5 Jul 2000 19:51:03 +0000 (19:51 +0000)]
Enable pg_rewrite to be toasted (unlimited size of views and rules).

Jan

24 years agoadd check for libz compression library for pg_dump
Marc G. Fournier [Wed, 5 Jul 2000 17:44:29 +0000 (17:44 +0000)]
add check for libz compression library for pg_dump

24 years agoChanged TOAST relations to have relkind RELKIND_TOASTVALUE.
Jan Wieck [Wed, 5 Jul 2000 16:17:43 +0000 (16:17 +0000)]
Changed TOAST relations to have relkind RELKIND_TOASTVALUE.

Special handling of TOAST relations during VACUUM. TOAST relations
are vacuumed while the lock on the master table is still active.
The ANALYZE flag doesn't propagate to their vacuuming because the
toaster access routines allways use index access ignoring stats, so
why compute them at all.

Protection of TOAST relations against normal INSERT/UPDATE/DELETE
while offering SELECT for debugging purposes.

Jan

24 years agoautoconf
Bruce Momjian [Wed, 5 Jul 2000 16:11:27 +0000 (16:11 +0000)]
autoconf

24 years agoattached to this mail is a patch from a colleague that makes
Bruce Momjian [Wed, 5 Jul 2000 16:09:31 +0000 (16:09 +0000)]
attached to this mail is a patch from a colleague that makes
PostgreSQL-7.0.2 run on Linux for the Intel-IA64 architecture. It also
fixes a bug in the configure scripts that caused configure to fail on
the fcntl(F_SETLK) test.

This fix triggered a bug in the fcntl(F_SETLK) code of the Linux
kernel when used on unix domain sockets resulting in postmaster to
segfault immediately after startup. There is a fix available and
included in the kernel that will be on SuSE Linux 7.0, but kernels <=
2.2.16 still have this bug.

Reinhard Max

24 years agoSmall fix for ALTER TABLE xxx* ADD COLUMN ...
Jan Wieck [Wed, 5 Jul 2000 13:50:59 +0000 (13:50 +0000)]
Small fix for ALTER TABLE xxx* ADD COLUMN ...

Inherited child tables must not be open during creation of TOAST table

Jan

24 years agoAutomatically create toast tables on ALTER TABLE ... ADD COLUMN
Jan Wieck [Wed, 5 Jul 2000 13:22:25 +0000 (13:22 +0000)]
Automatically create toast tables on ALTER TABLE ... ADD COLUMN
and SELECT ... INTO ... too.

Jan

24 years agoBugfix in ALTER TABLE CREATE TOAST TABLE
Jan Wieck [Wed, 5 Jul 2000 12:45:31 +0000 (12:45 +0000)]
Bugfix in ALTER TABLE CREATE TOAST TABLE

Automatically create toast table at CREATE TABLE if new table
has toastable attributes.

Jan

24 years agoFixed lztextlen() bug in MULTIBYTE, still thinking that lztext
Jan Wieck [Wed, 5 Jul 2000 10:09:53 +0000 (10:09 +0000)]
Fixed lztextlen() bug in MULTIBYTE, still thinking that lztext
does compression itself the old way.

Jan

24 years agoU Re-add flock TODO.detail.
Bruce Momjian [Wed, 5 Jul 2000 04:39:15 +0000 (04:39 +0000)]
U Re-add flock TODO.detail.

24 years agoUpdate TODO list.
Bruce Momjian [Wed, 5 Jul 2000 04:39:07 +0000 (04:39 +0000)]
Update TODO list.

24 years agoNew resultmap for Alpha, from Ryan Kirkpatrick
Bruce Momjian [Wed, 5 Jul 2000 04:28:19 +0000 (04:28 +0000)]
New resultmap for Alpha, from Ryan Kirkpatrick

24 years agoFix broken pg_dump makefile
Peter Eisentraut [Tue, 4 Jul 2000 19:52:00 +0000 (19:52 +0000)]
Fix broken pg_dump makefile

24 years agoMoving toaster out of NO ELOG area in heap_update().
Jan Wieck [Tue, 4 Jul 2000 17:11:40 +0000 (17:11 +0000)]
Moving toaster out of NO ELOG area in heap_update().

Jan

24 years agomake TEMP -> TEMPORARY
Bruce Momjian [Tue, 4 Jul 2000 16:57:49 +0000 (16:57 +0000)]
make TEMP -> TEMPORARY

24 years agoMake pgdump_oid a temp table. Rename to pg_dump_oid.
Bruce Momjian [Tue, 4 Jul 2000 16:57:18 +0000 (16:57 +0000)]
Make pgdump_oid a temp table.  Rename to pg_dump_oid.

24 years agoExpand secondary password file feature, so that you can use these
Peter Eisentraut [Tue, 4 Jul 2000 16:32:01 +0000 (16:32 +0000)]
Expand secondary password file feature, so that you can use these
files to restrict the set of users that can connect to a database
but can still use the pg_shadow password. (You just leave off the
password field in the secondary file.)

24 years agoFix thinko in my last pg_dumpall change
Peter Eisentraut [Tue, 4 Jul 2000 16:29:59 +0000 (16:29 +0000)]
Fix thinko in my last pg_dumpall change

24 years agoMakefile cleanup from Karel.
Bruce Momjian [Tue, 4 Jul 2000 16:17:21 +0000 (16:17 +0000)]
Makefile cleanup from Karel.

24 years agoAdded CVS header
Jan Wieck [Tue, 4 Jul 2000 14:37:32 +0000 (14:37 +0000)]
Added CVS header
Jan

24 years agoNew contrib module for statistical mathematics
Jan Wieck [Tue, 4 Jul 2000 14:36:03 +0000 (14:36 +0000)]
New contrib module for statistical mathematics

Jan

24 years agopg_dump patch from Philip Warner
Bruce Momjian [Tue, 4 Jul 2000 14:25:28 +0000 (14:25 +0000)]
pg_dump patch from Philip Warner

24 years agoUpdate TODO list.
Bruce Momjian [Tue, 4 Jul 2000 14:10:37 +0000 (14:10 +0000)]
Update TODO list.

24 years agoNow that I look at it, is_stopword() is broken and always has been.
Tom Lane [Tue, 4 Jul 2000 06:29:27 +0000 (06:29 +0000)]
Now that I look at it, is_stopword() is broken and always has been.
Doesn't anyone remember how to program a binary search??

24 years agoResult of strcmp() is a signed int. Per bug report
Tom Lane [Tue, 4 Jul 2000 06:20:14 +0000 (06:20 +0000)]
Result of strcmp() is a signed int.  Per bug report
from Paul McGarry.

24 years agoMake toast-table creation and deletion work somewhat reliably.
Tom Lane [Tue, 4 Jul 2000 06:11:54 +0000 (06:11 +0000)]
Make toast-table creation and deletion work somewhat reliably.
Don't go through pg_exec_query_dest(), but directly to the execution
routines.  Also, extend parameter lists so that there's no need to
change the global setting of allowSystemTableMods, a hack that was
certain to cause trouble in the event of any error.

24 years agoUpdate TODO list.
Bruce Momjian [Tue, 4 Jul 2000 05:17:31 +0000 (05:17 +0000)]
Update TODO list.

24 years agoRemove unused TODO.detail files.
Bruce Momjian [Tue, 4 Jul 2000 05:17:03 +0000 (05:17 +0000)]
Remove unused TODO.detail files.

24 years agoUpdate tablespaces TODO.detail item.
Bruce Momjian [Tue, 4 Jul 2000 05:04:19 +0000 (05:04 +0000)]
Update tablespaces TODO.detail item.

24 years agoUpdate TODO list.
Bruce Momjian [Tue, 4 Jul 2000 04:45:22 +0000 (04:45 +0000)]
Update TODO list.

24 years agoDataFill() has no business resetting xact status bits
Tom Lane [Tue, 4 Jul 2000 02:40:56 +0000 (02:40 +0000)]
DataFill() has no business resetting xact status bits
in the infomask of the provided tuple.

24 years agomisc
Vadim B. Mikheev [Tue, 4 Jul 2000 01:49:44 +0000 (01:49 +0000)]
misc

24 years agounlock buffer before releasing in heap_insert
Vadim B. Mikheev [Tue, 4 Jul 2000 01:39:24 +0000 (01:39 +0000)]
unlock buffer before releasing in heap_insert
+unlock buffer in heap_fetch under !ItemIdIsUsed().

24 years agotuptoaster.h should surely not #include postgres.h.
Tom Lane [Tue, 4 Jul 2000 00:04:03 +0000 (00:04 +0000)]
tuptoaster.h should surely not #include postgres.h.

24 years agoAdd missing #include for TOAST.
Tom Lane [Mon, 3 Jul 2000 23:58:32 +0000 (23:58 +0000)]
Add missing #include for TOAST.

24 years agoTOAST changes to catalog
Jan Wieck [Mon, 3 Jul 2000 23:19:04 +0000 (23:19 +0000)]
TOAST changes to catalog

Jan

24 years agoTOAST
Jan Wieck [Mon, 3 Jul 2000 23:10:14 +0000 (23:10 +0000)]
TOAST

    WARNING: This is actually broken - we have self-deadlocks
         due to concurrent changes in buffer management.
 Vadim and me are working on it.

Jan

24 years agoMove global internal files to PGDATA/global.
Peter Eisentraut [Mon, 3 Jul 2000 20:48:46 +0000 (20:48 +0000)]
Move global internal files to PGDATA/global.

24 years agoMove PGPORT envar handling to ResetAllOptions(). Improve long options
Peter Eisentraut [Mon, 3 Jul 2000 20:46:10 +0000 (20:46 +0000)]
Move PGPORT envar handling to ResetAllOptions(). Improve long options
parsing to not clobber the optarg string -- so that we can bring
SetOptsFile() up to speed.

24 years agoHave pg_dumpall write CREATE USER and CREATE GROUP commands instead of raw
Peter Eisentraut [Mon, 3 Jul 2000 16:35:39 +0000 (16:35 +0000)]
Have pg_dumpall write CREATE USER and CREATE GROUP commands instead of raw
COPYs of pg_shadow and pg_group.

It also turns out that pg_dumpall was all but broken for multiple servers
running at non-standard port numbers. You might get the users and groups
from one server and the databases from another. Fixed that.

A little user interface and code cleanup along with that. This also takes
care of the portability bug discussed in "[BUGS] pg_dumpall" in March 2000.

24 years ago Very small changes in the contrib's pg_dumplo in the attache.
Bruce Momjian [Mon, 3 Jul 2000 16:03:27 +0000 (16:03 +0000)]
 Very small changes in the contrib's pg_dumplo in the attache.

 And:

 Note, Bruce I found in the contrib tree any files that we forget
remove during contrib cleaning. Please remove these files:

contrib/lo/test.sql
contrib/pg_dumplo/Makefile.out
contrib/pgbench/pgbench_jis.doc
contrib/spi/new_example.example
contrib/spi/README.MAX

Thanks.

                                Karel

24 years agoIn the attache is new (correct) version. I add TZ (timezone) support
Bruce Momjian [Mon, 3 Jul 2000 16:01:33 +0000 (16:01 +0000)]
In the attache is new (correct) version. I add TZ (timezone) support
for
to_char() too.

Karel

24 years agoAdd a check of HEAP_MARKED_FOR_UPDATE in HeapTupleSatisfiesDirty().
Hiroshi Inoue [Mon, 3 Jul 2000 04:45:09 +0000 (04:45 +0000)]
Add a check of HEAP_MARKED_FOR_UPDATE in HeapTupleSatisfiesDirty().
This enables unique constraint check in case of SELECT FOR UPDATEd.

24 years agoClean up memory-context stuff, other minor infelicities.
Tom Lane [Mon, 3 Jul 2000 03:57:03 +0000 (03:57 +0000)]
Clean up memory-context stuff, other minor infelicities.

24 years agoRemove unused variable. (Hey Peter, you use gcc, why didn't you
Tom Lane [Mon, 3 Jul 2000 03:27:31 +0000 (03:27 +0000)]
Remove unused variable.  (Hey Peter, you use gcc, why didn't you
notice this warning?)

24 years agoheap' logging
Vadim B. Mikheev [Mon, 3 Jul 2000 02:54:21 +0000 (02:54 +0000)]
heap' logging

24 years agoAttached is a new patch which addresses this problem. (oids in
Bruce Momjian [Sun, 2 Jul 2000 22:01:27 +0000 (22:01 +0000)]
Attached is a new patch which addresses this problem. (oids in
regression tests).

Chris Bitmead

24 years agoVersion number now set in configure, available through Makefile.global
Peter Eisentraut [Sun, 2 Jul 2000 15:21:27 +0000 (15:21 +0000)]
Version number now set in configure, available through Makefile.global
and config.h. Adjusted all referring code.

Scrapped pg_version and changed initdb accordingly. Integrated
src/utils/version.c into src/backend/utils/init/miscinit.c. Changed all
callers.

Set version number to `7.1devel'. (Non-numeric version suffixes now allowed.)

24 years agoDon't run in CacheMemoryContext more than necessary.
Tom Lane [Sun, 2 Jul 2000 05:38:40 +0000 (05:38 +0000)]
Don't run in CacheMemoryContext more than necessary.

24 years agoDon't run in CacheMemoryContext longer than necessary.
Tom Lane [Sun, 2 Jul 2000 04:46:09 +0000 (04:46 +0000)]
Don't run in CacheMemoryContext longer than necessary.

24 years agoRemove memory leak from VACUUM parsing.
Tom Lane [Sun, 2 Jul 2000 04:04:09 +0000 (04:04 +0000)]
Remove memory leak from VACUUM parsing.

24 years agoNew memmgr logic in xact.c failed if AbortTransaction() is called when
Tom Lane [Sun, 2 Jul 2000 02:28:38 +0000 (02:28 +0000)]
New memmgr logic in xact.c failed if AbortTransaction() is called when
there is no open transaction.

24 years agoBack out Karels to_char changes.
Bruce Momjian [Sat, 1 Jul 2000 21:27:14 +0000 (21:27 +0000)]
Back out Karels to_char changes.

24 years agoWe don't need config.status substituting anything into Gen_fmgrtab.sh, we
Peter Eisentraut [Sat, 1 Jul 2000 21:16:44 +0000 (21:16 +0000)]
We don't need config.status substituting anything into Gen_fmgrtab.sh, we
just pass in CPP and AWK via the environment from the Makefile.

24 years agoSynchronized integrated documentation with latest admin guide change.
Peter Eisentraut [Sat, 1 Jul 2000 15:05:47 +0000 (15:05 +0000)]
Synchronized integrated documentation with latest admin guide change.

24 years agoA little bit of clean-up/bug fix in Tcl build. Handle gracefully cases
Peter Eisentraut [Sat, 1 Jul 2000 15:02:31 +0000 (15:02 +0000)]
A little bit of clean-up/bug fix in Tcl build. Handle gracefully cases
where someone wants to build these but hasn't configured for Tcl.

24 years ago In the attache is new to_date()/to_timestamp() version.
Bruce Momjian [Sat, 1 Jul 2000 14:10:32 +0000 (14:10 +0000)]
 In the attache is new to_date()/to_timestamp() version.

 Now the to_timestamp() support WW,W,J,SSSS,DDD conversion from strings and
the am/pm bug is fixed, the to_char() use week-of-year (WW) full compatible
with Oracle.

 This patch update relevant regress-tests and docs too.

                                        Karel
~
~

24 years agoFix select '1--2'; for PEter.
Bruce Momjian [Fri, 30 Jun 2000 18:03:40 +0000 (18:03 +0000)]
Fix select '1--2'; for PEter.

24 years agoNew and revised material for Admin guide, re backup & restore and database
Peter Eisentraut [Fri, 30 Jun 2000 16:14:21 +0000 (16:14 +0000)]
New and revised material for Admin guide, re backup & restore and database
management

24 years agoFixes for Solaris/cc suggested by <pgsql-hackers@thewrittenword.com>
Peter Eisentraut [Fri, 30 Jun 2000 16:11:02 +0000 (16:11 +0000)]
Fixes for Solaris/cc suggested by <pgsql-hackers@thewrittenword.com>

Don't use DISABLE_COMPLEX_MACRO on Solaris. Don't define the
replacement function in the header file. Use -KPIC, not -K PIC.
Use CC to link C++ libraries, not ld/ar.

Eliminate file not found warnings in tcl build code.

24 years agoRemoveView was doing things the hard way...
Tom Lane [Fri, 30 Jun 2000 07:06:05 +0000 (07:06 +0000)]
RemoveView was doing things the hard way...

24 years agoUse a private memory context to store rule information in each relcache
Tom Lane [Fri, 30 Jun 2000 07:04:23 +0000 (07:04 +0000)]
Use a private memory context to store rule information in each relcache
entry that has rules.  This allows us to release the rule parsetrees
on relcache flush without needing a working freeObject() routine.
Formerly, the rule trees were leaked permanently at relcache flush.
Also, clean up handling of rule creation and deletion --- there was
not sufficient locking of the relation being modified, and there was
no reliable notification of other backends that a relcache reload
was needed.  Also, clean up relcache.c code so that scans of system
tables needed to load a relcache entry are done in the caller's
memory context, not in CacheMemoryContext.  This prevents any
un-pfreed memory from those scans from becoming a permanent memory
leak.

24 years agoFix quotes in /* */ comments in psql.
Bruce Momjian [Thu, 29 Jun 2000 16:27:57 +0000 (16:27 +0000)]
Fix quotes in /* */ comments in psql.

24 years agoAdd test code to copy all parse/plan trees. Repair essential omissions
Tom Lane [Thu, 29 Jun 2000 07:35:57 +0000 (07:35 +0000)]
Add test code to copy all parse/plan trees.  Repair essential omissions
in copyfuncs and equalfuncs exposed by regression tests.  We still have
some work to do: these modules really ought to handle most or all of
the utility statement node types.  But it's better than it was.

24 years agoRearrange config.h entries into a more sensible order, and add some
Tom Lane [Thu, 29 Jun 2000 05:50:29 +0000 (05:50 +0000)]
Rearrange config.h entries into a more sensible order, and add some
documentation.  Let's try to keep this file a bit neater in future,
hmm?  Also (to get back to the original point) update info about
FUNC_MAX_ARGS, and add additional config symbols for debugging
new memory management changes.

24 years agoRemove freefuncs.c, which hasn't been used in a long time and is not
Tom Lane [Thu, 29 Jun 2000 02:26:23 +0000 (02:26 +0000)]
Remove freefuncs.c, which hasn't been used in a long time and is not
worth the effort to continue to maintain.  Since freeObject() is not
capable of coping with cases like multiple links to a node, it's
unlikely that it ever will be useful again.  We now have memory
context management that offers a faster and more reliable way of
getting rid of arbitrary node trees (at the cost of having to know
in advance of building the tree that you'll want to get rid of it).

24 years agoAdd SIGFPE to list of signals never to block. #ifdef all signals in
Tom Lane [Thu, 29 Jun 2000 02:17:42 +0000 (02:17 +0000)]
Add SIGFPE to list of signals never to block.  #ifdef all signals in
that list to cope with platforms that don't have them all.

24 years agoAdd #ifdef for LC_MESSAGES, because it is not ANSI C standard, and BSDI
Bruce Momjian [Thu, 29 Jun 2000 01:19:36 +0000 (01:19 +0000)]
Add #ifdef for LC_MESSAGES, because it is not ANSI C standard, and BSDI
does not have it.

24 years agoMakefile cleanup for interface tree. Now essentially with all the
Peter Eisentraut [Wed, 28 Jun 2000 18:30:16 +0000 (18:30 +0000)]
Makefile cleanup for interface tree. Now essentially with all the
standard targets and behaviour. Replaced Makefile.in's with
Makefile's and declared the respective variables in Makefile.global.

maintainer-clean target now available at top level, although it does
not work in the backend tree yet.

Cleanup pass over Makefile.shlib, renamed some targets and variables.
The shared library symlink tests are now done by make, not the shell.

ecpg: Remove one warning in sloppy flex output.

PL/Perl and Perl interface: the MakeMaker documentation is confusing,
the realclean target *does* "delete derived files", but it also
uninstalls them. Don't use that.

The submake targets in the various bin directories that update libpq
should `make all', not `make libpq.a'. That is a) unportable, and
b) doesn't build the shared library.

24 years agoSuppress memory leak in COPY WITH OIDS TO wherever.
Tom Lane [Wed, 28 Jun 2000 06:05:36 +0000 (06:05 +0000)]
Suppress memory leak in COPY WITH OIDS TO wherever.

24 years agoRepair incorrect assumption that all versions of 'install' know
Tom Lane [Wed, 28 Jun 2000 05:09:37 +0000 (05:09 +0000)]
Repair incorrect assumption that all versions of 'install' know
how to install multiple files in one invocation.

24 years agoCall me silly, but I think it would be a good idea if initdb and other
Tom Lane [Wed, 28 Jun 2000 04:58:36 +0000 (04:58 +0000)]
Call me silly, but I think it would be a good idea if initdb and other
scripts are executable after being installed.  Latest changes got the
permissions wrong (non executable).

24 years agoFirst phase of memory management rewrite (see backend/utils/mmgr/README
Tom Lane [Wed, 28 Jun 2000 03:33:33 +0000 (03:33 +0000)]
First phase of memory management rewrite (see backend/utils/mmgr/README
for details).  It doesn't really do that much yet, since there are no
short-term memory contexts in the executor, but the infrastructure is
in place and long-term contexts are handled reasonably.  A few long-
standing bugs have been fixed, such as 'VACUUM; anything' in a single
query string crashing.  Also, out-of-memory is now considered a
recoverable ERROR, not FATAL.
Eliminate a large amount of crufty, now-dead code in and around
memory management.
Fix problem with holding off SIGTRAP, SIGSEGV, etc in postmaster and
backend startup.

24 years agoTamotsu Nakagawa has posted a fix for this to a local mail list in
Bruce Momjian [Tue, 27 Jun 2000 18:08:49 +0000 (18:08 +0000)]
Tamotsu Nakagawa has posted a fix for this to a local mail list in
Japan. Can someone comment on this? According to him, with the patch
now only the geometry test fails.

 Tatsuo Ishii