OSDN Git Service

pg-rex/syncrep.git
23 years agoAdd Darwin-specific geometry test file.
Tom Lane [Mon, 11 Dec 2000 16:45:16 +0000 (16:45 +0000)]
Add Darwin-specific geometry test file.

23 years agoTweak Darwin patch to get right include order.
Tom Lane [Mon, 11 Dec 2000 16:35:59 +0000 (16:35 +0000)]
Tweak Darwin patch to get right include order.

23 years agoResolve complie error(was my fault).
Hiroshi Inoue [Mon, 11 Dec 2000 09:14:03 +0000 (09:14 +0000)]
Resolve complie error(was my fault).

23 years ago*redo: Heap move* neglects to set t_cmin for MOVED_IN tuples.
Hiroshi Inoue [Mon, 11 Dec 2000 05:25:23 +0000 (05:25 +0000)]
*redo: Heap move* neglects to set t_cmin for MOVED_IN tuples.

23 years agoFix ILIKE bug (only in multi-byte case)
Tatsuo Ishii [Mon, 11 Dec 2000 05:00:18 +0000 (05:00 +0000)]
Fix ILIKE bug (only in multi-byte case)

23 years agoUpdate TODO list.
Bruce Momjian [Mon, 11 Dec 2000 01:44:37 +0000 (01:44 +0000)]
Update TODO list.

23 years agoDarwin porting patches from Peter Bierman <bierman@apple.com>
Tom Lane [Mon, 11 Dec 2000 00:49:54 +0000 (00:49 +0000)]
Darwin porting patches from Peter Bierman <bierman@apple.com>

23 years agoHere is patch to the ODBC driver to update the version to 2.5 and allow
Bruce Momjian [Sun, 10 Dec 2000 23:54:28 +0000 (23:54 +0000)]
Here is patch to the ODBC driver to update the version to 2.5 and allow
all forms of foreign keys be exposed to SQLForeignKeys.  This patch is in
addition to the ones I mailed yesterday (forget had I changed that as
well....)

Michael Fork - CCNA - MCP - A+
Network Support - Toledo Internet Access - Toledo Ohio

23 years agoHere is a diff to info.c in interfaces/odbc that updates SQLForeignKeys to
Bruce Momjian [Sun, 10 Dec 2000 22:59:10 +0000 (22:59 +0000)]
Here is a diff to info.c in interfaces/odbc that updates SQLForeignKeys to
return foreign key information based on the pg_trigger system table.  I
have tested the patch with (what I believe) is all possible
primary/foreign key combinations -- however I may have missed some, so if
anyone feels like taking the patch for a test drive, here are some useful
links:

Michael Fork

23 years agoBackout right-hand/left-hand.
Bruce Momjian [Sun, 10 Dec 2000 22:37:27 +0000 (22:37 +0000)]
Backout right-hand/left-hand.

23 years agoin catalog.sgml
Bruce Momjian [Sun, 10 Dec 2000 22:35:56 +0000 (22:35 +0000)]
in catalog.sgml
line 1324:
"left-hand" should be "right-hand"

BTW: new document looks very good!
And the new configure/build process seems much better then before!

Thanks!

Laser

23 years agoAdd mention of \d+ to comment.
Bruce Momjian [Sun, 10 Dec 2000 21:56:00 +0000 (21:56 +0000)]
Add mention of \d+ to comment.

23 years agoCorrect one description, add one.
Peter Eisentraut [Sun, 10 Dec 2000 21:19:02 +0000 (21:19 +0000)]
Correct one description, add one.

23 years agoFix typo
Bruce Momjian [Sun, 10 Dec 2000 20:47:34 +0000 (20:47 +0000)]
Fix typo

23 years agoUpdate TODO list.
Bruce Momjian [Sat, 9 Dec 2000 23:25:52 +0000 (23:25 +0000)]
Update TODO list.

23 years agoFixes for examples from Thomas Diffenbach
Bruce Momjian [Sat, 9 Dec 2000 22:59:25 +0000 (22:59 +0000)]
Fixes for examples from Thomas Diffenbach

23 years agoPortability fix from Ryan Kirkpatrick's Alpha patches. I believe this
Tom Lane [Sat, 9 Dec 2000 20:40:57 +0000 (20:40 +0000)]
Portability fix from Ryan Kirkpatrick's Alpha patches.  I believe this
is the only diff not accounted for by fmgr rewrite...

23 years agoSuppress compiler warnings.
Tom Lane [Sat, 9 Dec 2000 20:32:44 +0000 (20:32 +0000)]
Suppress compiler warnings.

23 years agoAdd fulltextindex TODO list.
Bruce Momjian [Sat, 9 Dec 2000 15:52:11 +0000 (15:52 +0000)]
Add fulltextindex TODO list.

23 years agoUpdate FAQ_DEV.
Bruce Momjian [Sat, 9 Dec 2000 04:57:31 +0000 (04:57 +0000)]
Update FAQ_DEV.

23 years agoUpdate FAQ_DEV.
Bruce Momjian [Sat, 9 Dec 2000 04:29:43 +0000 (04:29 +0000)]
Update FAQ_DEV.

23 years agoUpdate FAQ.
Bruce Momjian [Sat, 9 Dec 2000 04:29:38 +0000 (04:29 +0000)]
Update FAQ.

23 years agoFix a bug in conversion from big5 to EUC_TW (CNS 11643-1992 Plane 3)
Tatsuo Ishii [Sat, 9 Dec 2000 04:27:36 +0000 (04:27 +0000)]
Fix a bug in conversion from big5 to EUC_TW (CNS 11643-1992 Plane 3)
Thanks Chih-Chang Hsieh <cch@cc.kmu.edu.tw> for finding the bug.

23 years agoRepair erroneous use of hashvarlena() for MACADDR, which is not a
Tom Lane [Fri, 8 Dec 2000 23:57:03 +0000 (23:57 +0000)]
Repair erroneous use of hashvarlena() for MACADDR, which is not a
varlena type.  (I did not force initdb, but you won't see the fix
unless you do one.)  Also, make sure all index support operators and
functions are careful not to leak memory for toasted inputs; I had
missed some hash and rtree support ops on this point before.

23 years agoResurrect -F switch: it controls fsyncs again, though the fsyncs are
Tom Lane [Fri, 8 Dec 2000 22:21:33 +0000 (22:21 +0000)]
Resurrect -F switch: it controls fsyncs again, though the fsyncs are
mostly just on the WAL logfile nowadays.  But if people want to disable
fsync for performance, why should we say no?

23 years agoFix Westwood/Westward, from Wessel van Norel.
Bruce Momjian [Fri, 8 Dec 2000 20:11:11 +0000 (20:11 +0000)]
Fix Westwood/Westward, from Wessel van Norel.

23 years agoAdd missing copyright and RCS identification header.
Tom Lane [Fri, 8 Dec 2000 20:10:19 +0000 (20:10 +0000)]
Add missing copyright and RCS identification header.

23 years agoRemove error check that disallowed setval() on a sequence with cache
Tom Lane [Fri, 8 Dec 2000 20:06:58 +0000 (20:06 +0000)]
Remove error check that disallowed setval() on a sequence with cache
value greater than one.  The behavior this sought to disallow doesn't
seem any less confusing than the other behaviors of cached sequences.
Improve wording of some error messages, too.
Update documentation accordingly.  Also add an explanation that
aborted transactions do not roll back their nextval() calls; this
seems to be a FAQ, so it ought to be mentioned here...

23 years agoCache invalidation for vacuum of system tables.
Hiroshi Inoue [Fri, 8 Dec 2000 06:43:44 +0000 (06:43 +0000)]
Cache invalidation for vacuum of system tables.

23 years agoREINDEX under WAL.
Hiroshi Inoue [Fri, 8 Dec 2000 06:17:58 +0000 (06:17 +0000)]
REINDEX under WAL.

23 years agoImprove error message for erroneous use of 'opaque' as plperl argument
Tom Lane [Fri, 8 Dec 2000 00:11:55 +0000 (00:11 +0000)]
Improve error message for erroneous use of 'opaque' as plperl argument
or return type.

23 years agoImprove error message for erroneous use of 'opaque' as pltcl argument
Tom Lane [Fri, 8 Dec 2000 00:09:07 +0000 (00:09 +0000)]
Improve error message for erroneous use of 'opaque' as pltcl argument
or return type.

23 years agoImprove error message for erroneous use of 'opaque' as plpgsql argument
Tom Lane [Fri, 8 Dec 2000 00:03:02 +0000 (00:03 +0000)]
Improve error message for erroneous use of 'opaque' as plpgsql argument
or return type.

23 years agoChange lpad() and rpad() to behave more Oracle-compatibly when target
Tom Lane [Thu, 7 Dec 2000 23:22:56 +0000 (23:22 +0000)]
Change lpad() and rpad() to behave more Oracle-compatibly when target
length is less than original string length.

23 years agotypo correction
Peter Eisentraut [Thu, 7 Dec 2000 22:37:08 +0000 (22:37 +0000)]
typo correction

23 years agoDo not use 'ar cq' to build library archives, use 'ar cr' instead.
Peter Eisentraut [Thu, 7 Dec 2000 19:43:49 +0000 (19:43 +0000)]
Do not use 'ar cq' to build library archives, use 'ar cr' instead.

23 years agocheckretval() failed to cope with an empty SQL function body.
Tom Lane [Thu, 7 Dec 2000 19:40:56 +0000 (19:40 +0000)]
checkretval() failed to cope with an empty SQL function body.

23 years agoMake OVERLAPS operators conform to SQL92 spec regarding NULL handling.
Tom Lane [Thu, 7 Dec 2000 18:38:59 +0000 (18:38 +0000)]
Make OVERLAPS operators conform to SQL92 spec regarding NULL handling.
As I read it, the spec requires a non-null result in some cases where
one of the inputs is NULL: specifically, if the other endpoint of that
interval is between the endpoints of the other interval, then the result
is known TRUE despite the missing endpoint.  The spec could've been a
lot simpler if they did not intend this behavior.
I did not force an initdb for this change, but if you don't do one you'll
still see the old strict-function behavior.

23 years agoRecordTransactionAbort() shouldn't log XLOG_XACT_ABORT
Hiroshi Inoue [Thu, 7 Dec 2000 10:03:46 +0000 (10:03 +0000)]
RecordTransactionAbort() shouldn't log XLOG_XACT_ABORT
if the transaction has already been committed ?

23 years agoSilence compiler warning.
Tom Lane [Thu, 7 Dec 2000 02:52:27 +0000 (02:52 +0000)]
Silence compiler warning.

23 years agoSilence compiler warning.
Tom Lane [Thu, 7 Dec 2000 02:04:30 +0000 (02:04 +0000)]
Silence compiler warning.

23 years agoMake application of FOR UPDATE to a view work exactly like the parser's
Tom Lane [Thu, 7 Dec 2000 01:22:25 +0000 (01:22 +0000)]
Make application of FOR UPDATE to a view work exactly like the parser's
transformForUpdate does: it should recurse into subqueries.

23 years agotransformForUpdate() mustn't assume rowMarks list is initially empty.
Tom Lane [Thu, 7 Dec 2000 01:12:08 +0000 (01:12 +0000)]
transformForUpdate() mustn't assume rowMarks list is initially empty.
It could be recursing into a sub-query where there was already a FOR
UPDATE clause.

23 years agoClean up handling of FOR UPDATE inside views and subselects ... make it
Tom Lane [Wed, 6 Dec 2000 23:55:19 +0000 (23:55 +0000)]
Clean up handling of FOR UPDATE inside views and subselects ... make it
work where we can (given that the executor only handles it at top level)
and generate an error where we can't.  Note that while the parser has
been allowing views to say SELECT FOR UPDATE for a few weeks now, that
hasn't actually worked until just now.

23 years agoupdate VERSION to 7.1beta1..
Marc G. Fournier [Wed, 6 Dec 2000 20:26:53 +0000 (20:26 +0000)]
update VERSION to 7.1beta1..

23 years ago-L linker options are now in LDFLAGS, not LIBS. The word hadn't gotten
Peter Eisentraut [Wed, 6 Dec 2000 19:37:08 +0000 (19:37 +0000)]
-L linker options are now in LDFLAGS, not LIBS.  The word hadn't gotten
through to here yet.

23 years agoDon't include miscadmin.h in elog.h, since the former is not part of
Tom Lane [Wed, 6 Dec 2000 17:25:46 +0000 (17:25 +0000)]
Don't include miscadmin.h in elog.h, since the former is not part of
the installed header file set.

23 years agoI've just seen what happens when the MEMORY_CONTEXT_CHECKING code fires
Tom Lane [Tue, 5 Dec 2000 23:40:36 +0000 (23:40 +0000)]
I've just seen what happens when the MEMORY_CONTEXT_CHECKING code fires
an error at end of transaction ... and I did *not* like it.  Reduce ERROR
to NOTICE so that this situation doesn't cause an infinite loop.

23 years agoIn SELECT FOR UPDATE, silently ignore null CTIDs, rather than generating
Tom Lane [Tue, 5 Dec 2000 22:03:57 +0000 (22:03 +0000)]
In SELECT FOR UPDATE, silently ignore null CTIDs, rather than generating
an error as we used to.  In an OUTER JOIN scenario, retrieving a null
CTID from one of the input relations is entirely expected.  We still
want to lock the input rows from the other relations, so just ignore
the null and keep going.

23 years agoFrom Stephan Szabo:
Tom Lane [Tue, 5 Dec 2000 19:57:56 +0000 (19:57 +0000)]
From Stephan Szabo:
I believe this should fix the issue that Philip Warner
noticed about the check for unique constraints meeting the
referenced keys of a foreign key constraint allowing the
specification of a subset of a foreign key instead of
rejecting it.  I also added tests for a base case of
this to the foreign key and alter table tests and patches
for expected output.

23 years agoAdd regress test case for INSERT ... SELECT in rules.
Tom Lane [Tue, 5 Dec 2000 19:15:49 +0000 (19:15 +0000)]
Add regress test case for INSERT ... SELECT in rules.

23 years agoRepair breakage of rules containing INSERT ... SELECT actions, per bug
Tom Lane [Tue, 5 Dec 2000 19:15:10 +0000 (19:15 +0000)]
Repair breakage of rules containing INSERT ... SELECT actions, per bug
report from Joel Burton.  Turns out that my simple idea of turning the
SELECT into a subquery does not interact well *at all* with the way the
rule rewriter works.  Really what we need to make INSERT ... SELECT work
cleanly is to decouple targetlists from rangetables: an INSERT ... SELECT
wants to have two levels of targetlist but only one rangetable.  No time
for that for 7.1, however, so I've inserted some ugly hacks to make the
rewriter know explicitly about the structure of INSERT ... SELECT queries.
Ugh :-(

23 years agoFix bogus makefiles ... these didn't build on platforms that are sticky
Tom Lane [Mon, 4 Dec 2000 01:32:19 +0000 (01:32 +0000)]
Fix bogus makefiles ... these didn't build on platforms that are sticky
about being given accurate references to referenced libraries ...

23 years agoSeems to have been missed in GetPgUserName updates.
Tom Lane [Mon, 4 Dec 2000 01:25:35 +0000 (01:25 +0000)]
Seems to have been missed in GetPgUserName updates.

23 years agoEliminate some of the more blatant platform-dependencies ... it builds here now,...
Tom Lane [Mon, 4 Dec 2000 01:20:38 +0000 (01:20 +0000)]
Eliminate some of the more blatant platform-dependencies ... it builds here now, anyway ...

23 years agoSeems like make clean or make distclean should clean out contrib
Tom Lane [Mon, 4 Dec 2000 00:34:40 +0000 (00:34 +0000)]
Seems like make clean or make distclean should clean out contrib
directories too, not only src and doc.

23 years agocorrections
Peter Eisentraut [Sun, 3 Dec 2000 21:12:19 +0000 (21:12 +0000)]
corrections

23 years agoEnsure that all uses of <ctype.h> functions are applied to unsigned-char
Tom Lane [Sun, 3 Dec 2000 20:45:40 +0000 (20:45 +0000)]
Ensure that all uses of <ctype.h> functions are applied to unsigned-char
values, whether the local char type is signed or not.  This is necessary
for portability.  Per discussion on pghackers around 9/16/00.

23 years agoUpdate for new tests. These are extrapolations and will need to be confirmed
Tom Lane [Sun, 3 Dec 2000 17:53:26 +0000 (17:53 +0000)]
Update for new tests.  These are extrapolations and will need to be confirmed
correct on the relevant platforms.

23 years agoUpdate for new tests.
Tom Lane [Sun, 3 Dec 2000 17:37:26 +0000 (17:37 +0000)]
Update for new tests.

23 years agoDon't use 'private' as a parameter name in visible headers ... makes C++
Tom Lane [Sun, 3 Dec 2000 17:18:10 +0000 (17:18 +0000)]
Don't use 'private' as a parameter name in visible headers ... makes C++
very unhappy ...

23 years agoFix counting of lines in scripts with 'copy from stdin'.
Peter Eisentraut [Sun, 3 Dec 2000 15:39:38 +0000 (15:39 +0000)]
Fix counting of lines in scripts with 'copy from stdin'.

23 years agoRepair usage of the OVERLAPS operator.
Thomas G. Lockhart [Sun, 3 Dec 2000 14:51:12 +0000 (14:51 +0000)]
Repair usage of the OVERLAPS operator.
Allow some operator-like tokens to be used as function names.
Flesh out support for time, timetz, and interval operators
 and interactions.
Regression tests pass, but non-reference-platform horology test results
 will need to be updated.

23 years agoClarify the allowed length of the text data type.
Thomas G. Lockhart [Sun, 3 Dec 2000 14:47:18 +0000 (14:47 +0000)]
Clarify the allowed length of the text data type.

23 years agoConvert files from DOS format to normal text.
Thomas G. Lockhart [Sun, 3 Dec 2000 14:43:59 +0000 (14:43 +0000)]
Convert files from DOS format to normal text.

23 years agoSupport IBM S/390. Patches from Neale Ferguson@softwareAG-usa.com.
Thomas G. Lockhart [Sun, 3 Dec 2000 14:41:47 +0000 (14:41 +0000)]
Support IBM S/390. Patches from Neale Ferguson@softwareAG-usa.com.

23 years agoFinal(?) GUC clean-up. Update psql tab completion.
Peter Eisentraut [Sun, 3 Dec 2000 14:36:47 +0000 (14:36 +0000)]
Final(?) GUC clean-up.  Update psql tab completion.

23 years agoDisable elog(ERROR|FATAL) in signal handlers in
Vadim B. Mikheev [Sun, 3 Dec 2000 10:27:29 +0000 (10:27 +0000)]
Disable elog(ERROR|FATAL) in signal handlers in
critical sections of code.

23 years agoAvoid memory leakage during regular COPY when outputting toasted values.
Tom Lane [Sat, 2 Dec 2000 20:49:24 +0000 (20:49 +0000)]
Avoid memory leakage during regular COPY when outputting toasted values.
COPY BINARY is still broken for toasted data, however.

23 years agoAvoid repeated detoasting (and possible memory leaks) when processing
Tom Lane [Sat, 2 Dec 2000 19:38:34 +0000 (19:38 +0000)]
Avoid repeated detoasting (and possible memory leaks) when processing
a toasted datum in VACUUM ANALYZE.

23 years agoDon't cause --enable-cassert to define COPY_PARSE_PLAN_TREES automatically
Tom Lane [Sat, 2 Dec 2000 18:16:40 +0000 (18:16 +0000)]
Don't cause --enable-cassert to define COPY_PARSE_PLAN_TREES automatically
anymore.  That won't teach us anything new for the rest of this release
cycle, so it seems better to keep the --assert environment more like the
non-assert environment for beta.
I'm going to leave CLOBBER_FREED_MEMORY and MEMORY_CONTEXT_CHECKING
turned on by --enable-cassert for now, however.

23 years agoMake tuple receive/print routines TOAST-aware. Formerly, printtup would
Tom Lane [Fri, 1 Dec 2000 22:10:31 +0000 (22:10 +0000)]
Make tuple receive/print routines TOAST-aware.  Formerly, printtup would
leak memory when printing a toasted attribute, and printtup_internal
didn't work at all...

23 years agoFix inadequate tree-walking code in exec_eval_clear_fcache.
Tom Lane [Fri, 1 Dec 2000 20:43:59 +0000 (20:43 +0000)]
Fix inadequate tree-walking code in exec_eval_clear_fcache.

23 years agoMake elog() switch to ErrorContext while invoking libpq output routines,
Tom Lane [Fri, 1 Dec 2000 19:52:04 +0000 (19:52 +0000)]
Make elog() switch to ErrorContext while invoking libpq output routines,
since those routines may do palloc's.  We want to be fairly sure we can
send the error message to the client even under low-memory conditions.
That's what we stashed away 8K in ErrorContext for, after all ...

23 years agoPursuant to a pghackers discussion back around 11-Jul-00, get rid of aset.c's
Tom Lane [Fri, 1 Dec 2000 18:14:29 +0000 (18:14 +0000)]
Pursuant to a pghackers discussion back around 11-Jul-00, get rid of aset.c's
not-very-good handling of mid-size allocation requests.  Do everything via
either the "small" case (chunk size rounded up to power of 2) or the "large"
case (pass it straight off to malloc()).  Increase the number of freelists
a little to set the breakpoint between these behaviors at 8K.

23 years agoIt seems appropriate that the extended MEMORY_CONTEXT_CHECKING code
Tom Lane [Fri, 1 Dec 2000 05:17:19 +0000 (05:17 +0000)]
It seems appropriate that the extended MEMORY_CONTEXT_CHECKING code
immediately uncovered three of Karel's own bugs, including a routine
that scribbled on its input (naughty naughty!)

23 years agoClean up MEMORY_CONTEXT_CHECKING code, and apply it more thoroughly. Also,
Tom Lane [Fri, 1 Dec 2000 05:16:45 +0000 (05:16 +0000)]
Clean up MEMORY_CONTEXT_CHECKING code, and apply it more thoroughly.  Also,
apply Karel Zak's patch to recycle residual space in an exhausted allocation
block.  (Bet you thought I'd forgot about that, Karel?)

23 years agoRepair residual sillinesses from UUNET virtual host/socket path patch.
Tom Lane [Thu, 30 Nov 2000 23:20:51 +0000 (23:20 +0000)]
Repair residual sillinesses from UUNET virtual host/socket path patch.
I hope all the dust has settled out now ...

23 years agoMake default socket directory location configurable from config.h.
Tom Lane [Thu, 30 Nov 2000 23:19:04 +0000 (23:19 +0000)]
Make default socket directory location configurable from config.h.
If we're going to let it be run-time configurable, might as well allow
this too...

23 years agoCorrect obsolete entry for strpos().
Tom Lane [Thu, 30 Nov 2000 23:02:04 +0000 (23:02 +0000)]
Correct obsolete entry for strpos().

23 years agoMinor tweaks in installation instructions, regenerate INSTALL file.
Peter Eisentraut [Thu, 30 Nov 2000 21:44:07 +0000 (21:44 +0000)]
Minor tweaks in installation instructions, regenerate INSTALL file.

23 years agoRemove DISABLE_COMPLEX_MACRO definitions, since people seem to be getting
Peter Eisentraut [Thu, 30 Nov 2000 21:16:12 +0000 (21:16 +0000)]
Remove DISABLE_COMPLEX_MACRO definitions, since people seem to be getting
by without them.

Don't check for preprocessor symbols from system header files in port
include files, since those header files aren't included at this point.

23 years agoRemove old regression test drivers.
Peter Eisentraut [Thu, 30 Nov 2000 20:43:31 +0000 (20:43 +0000)]
Remove old regression test drivers.

23 years agoMake all commands that link a program look like
Peter Eisentraut [Thu, 30 Nov 2000 20:36:13 +0000 (20:36 +0000)]
Make all commands that link a program look like

$(CC) $(CFLAGS) $(LDFLAGS) <object files> <extra-libraries> $(LIBS) -o $@

This form seemed to be the most portable, readable, and logical, but in any
case it's better than having a dozen different ones in the tree.

23 years agomisc
Vadim B. Mikheev [Thu, 30 Nov 2000 19:06:37 +0000 (19:06 +0000)]
misc

23 years agoHope that this is valid localbuf.c version
Vadim B. Mikheev [Thu, 30 Nov 2000 19:03:26 +0000 (19:03 +0000)]
Hope that this is valid localbuf.c version

23 years agoRemove obsolete claim that char(n) is more efficient than other text types.
Tom Lane [Thu, 30 Nov 2000 18:46:42 +0000 (18:46 +0000)]
Remove obsolete claim that char(n) is more efficient than other text types.

23 years agoRemove VARLENA_FIXED_SIZE hack, which is irreversibly broken now that
Tom Lane [Thu, 30 Nov 2000 18:38:47 +0000 (18:38 +0000)]
Remove VARLENA_FIXED_SIZE hack, which is irreversibly broken now that
both MULTIBYTE and TOAST prevent char(n) from being truly fixed-size.
Simplify and speed up fastgetattr() and index_getattr() macros by
eliminating special cases for attnum=1.  It's just as fast to handle
the first attribute by presetting its attcacheoff to zero; so do that
instead when loading the tupledesc in relcache.c.

23 years agoSort out things after fe-connect.c cleanup after unix socket patch.
Peter Eisentraut [Thu, 30 Nov 2000 18:34:36 +0000 (18:34 +0000)]
Sort out things after fe-connect.c cleanup after unix socket patch.

23 years agoRemove remaining unixsocket-setting interfaces, since the host parameter
Peter Eisentraut [Thu, 30 Nov 2000 18:32:52 +0000 (18:32 +0000)]
Remove remaining unixsocket-setting interfaces, since the host parameter
does that.  Disable URL-style connection parameter specification code,
which doesn't work.

23 years agoNo more #ifdef XLOG.
Vadim B. Mikheev [Thu, 30 Nov 2000 08:46:26 +0000 (08:46 +0000)]
No more #ifdef XLOG.

23 years agoIt seems some platforms declare kill(2) in signal.h not unistd.h.
Tom Lane [Thu, 30 Nov 2000 03:11:24 +0000 (03:11 +0000)]
It seems some platforms declare kill(2) in signal.h not unistd.h.

23 years agoXLOG stuff for sequences.
Vadim B. Mikheev [Thu, 30 Nov 2000 01:47:33 +0000 (01:47 +0000)]
XLOG stuff for sequences.
CommitDelay in guc.c

23 years agoRearrange bufmgr header files so that buf_internals.h need not be
Tom Lane [Thu, 30 Nov 2000 01:39:08 +0000 (01:39 +0000)]
Rearrange bufmgr header files so that buf_internals.h need not be
included by everything that includes bufmgr.h --- it's supposed to be
internals, after all, not part of the API!  This fixes the conflict
against FreeBSD headers reported by Rosenman, by making it unnecessary
for s_lock.h to be included by plperl.c.

23 years agoall options are allowed if not under postmaster:
Vadim B. Mikheev [Thu, 30 Nov 2000 01:27:19 +0000 (01:27 +0000)]
all options are allowed if not under postmaster:

   SetConfigOption(name, value,
        (IsUnderPostmaster) ? PGC_BACKEND : PGC_POSTMASTER);

23 years agoJust noticed that with -S switch, MyProcPid is permanently wrong in
Tom Lane [Wed, 29 Nov 2000 22:04:04 +0000 (22:04 +0000)]
Just noticed that with -S switch, MyProcPid is permanently wrong in
postmaster, because it isn't updated after forking away from the terminal.
Apparently it's not used anyplace in the postmaster ... but seems best
to make it show the correct PID ...

23 years agoGet rid of not-very-portable fcntl(F_SETLK) mechanism for locking the Unix
Tom Lane [Wed, 29 Nov 2000 20:59:54 +0000 (20:59 +0000)]
Get rid of not-very-portable fcntl(F_SETLK) mechanism for locking the Unix
socket file, in favor of having an ordinary lockfile beside the socket file.
Clean up a few robustness problems in the lockfile code.  If postmaster is
going to reject a connection request based on database state, it will now
tell you so before authentication exchange not after.  (Of course, a failure
after is still possible if conditions change meanwhile, but this makes life
easier for a yet-to-be-written pg_ping utility.)

23 years agoRewrite system catalog documentation man page in DocBook with more recent
Peter Eisentraut [Wed, 29 Nov 2000 20:15:59 +0000 (20:15 +0000)]
Rewrite system catalog documentation man page in DocBook with more recent
information.

23 years agoaclitemout() shouldn't coredump when it finds an ACL item
Tom Lane [Tue, 28 Nov 2000 23:42:31 +0000 (23:42 +0000)]
aclitemout() shouldn't coredump when it finds an ACL item
for a now-vanished group.  Instead, display the numeric group ID, same
as it does for vanished users.

23 years agoSignificant cleanups in SysV IPC handling (shared mem and semaphores).
Tom Lane [Tue, 28 Nov 2000 23:27:57 +0000 (23:27 +0000)]
Significant cleanups in SysV IPC handling (shared mem and semaphores).
IPC key assignment will now work correctly even when multiple postmasters
are using same logical port number (which is possible given -k switch).
There is only one shared-mem segment per postmaster now, not 3.
Rip out broken code for non-TAS case in bufmgr and xlog, substitute a
complete S_LOCK emulation using semaphores in spin.c.  TAS and non-TAS
logic is now exactly the same.
When deadlock is detected, "Deadlock detected" is now the elog(ERROR)
message, rather than a NOTICE that comes out before an unhelpful ERROR.