]> granicus.if.org Git - postgresql/log
postgresql
24 years agoFirst cut at unifying regular selectivity estimation with indexscan
Tom Lane [Sun, 23 Jan 2000 02:07:00 +0000 (02:07 +0000)]
First cut at unifying regular selectivity estimation with indexscan
selectivity estimation wasn't right.  This is better...

24 years agoupdated install file
Peter Eisentraut [Sun, 23 Jan 2000 01:27:39 +0000 (01:27 +0000)]
updated install file
updated date/time types doc
fixed small psql bug
removed libpq code that lower-cased db names
make notice when long identifier is truncated

24 years agoUpdate regress tests for change of pg_am/pg_amop.
Tom Lane [Sat, 22 Jan 2000 23:51:20 +0000 (23:51 +0000)]
Update regress tests for change of pg_am/pg_amop.

24 years agoRevise handling of index-type-specific indexscan cost estimation, per
Tom Lane [Sat, 22 Jan 2000 23:50:30 +0000 (23:50 +0000)]
Revise handling of index-type-specific indexscan cost estimation, per
pghackers discussion of 5-Jan-2000.  The amopselect and amopnpages
estimators are gone, and in their place is a per-AM amcostestimate
procedure (linked to from pg_am, not pg_amop).

24 years agofindoidjoins was suffering from bit rot (pginterface no
Tom Lane [Sat, 22 Jan 2000 23:05:14 +0000 (23:05 +0000)]
findoidjoins was suffering from bit rot (pginterface no
longer in expected location).

24 years agoadded ALTER TABLE DROP COLUMN, early version
Peter Eisentraut [Sat, 22 Jan 2000 14:20:56 +0000 (14:20 +0000)]
added ALTER TABLE DROP COLUMN, early version

24 years agoUpdate TODO list.
Bruce Momjian [Sat, 22 Jan 2000 04:29:46 +0000 (04:29 +0000)]
Update TODO list.

24 years agoChange a few routines into macros to improve speed of COPY IN inner loop.
Tom Lane [Sat, 22 Jan 2000 03:52:04 +0000 (03:52 +0000)]
Change a few routines into macros to improve speed of COPY IN inner loop.

24 years agoForgot to handle column length defaults in ALTER TABLE ADD COLUMN.
Tom Lane [Sat, 22 Jan 2000 01:22:48 +0000 (01:22 +0000)]
Forgot to handle column length defaults in ALTER TABLE ADD COLUMN.

24 years agoFix tab-complete so it works with old versions of readline that don't
Tom Lane [Fri, 21 Jan 2000 23:32:36 +0000 (23:32 +0000)]
Fix tab-complete so it works with old versions of readline that don't
have the rl_completion_append_character variable.  The tab completion
behavior doesn't seem to be quite perfect in that situation, but it's
better than failing to build at all...

24 years agoFix handleCopyIn's response to EOF seen mid-line, that is, input file
Tom Lane [Fri, 21 Jan 2000 04:21:12 +0000 (04:21 +0000)]
Fix handleCopyIn's response to EOF seen mid-line, that is, input file
does not end with a newline.  I don't think this explains the recent
complaints, since this bug existed in 6.5 (and probably long before).
But might as well fix it now that I see it.

24 years agoUpdate TODO list.
Bruce Momjian [Fri, 21 Jan 2000 03:55:22 +0000 (03:55 +0000)]
Update TODO list.

24 years agoSimplify match patterns a little.
Tom Lane [Fri, 21 Jan 2000 01:48:02 +0000 (01:48 +0000)]
Simplify match patterns a little.

24 years agoBruce, you forgot to rename these files...
Tom Lane [Fri, 21 Jan 2000 01:39:54 +0000 (01:39 +0000)]
Bruce, you forgot to rename these files...

24 years agoUpdate TODO list.
Bruce Momjian [Fri, 21 Jan 2000 00:45:29 +0000 (00:45 +0000)]
Update TODO list.

24 years agoplperl Makefile cleanup. Doesn't work on BSDI yet.
Bruce Momjian [Thu, 20 Jan 2000 23:00:40 +0000 (23:00 +0000)]
plperl Makefile cleanup.  Doesn't work on BSDI yet.

24 years agoUpdate plperl makefile.
Bruce Momjian [Thu, 20 Jan 2000 22:31:23 +0000 (22:31 +0000)]
Update plperl makefile.

24 years agoRename 7.0.0 to 7.0 to be consistent with prior release numbering.
Bruce Momjian [Thu, 20 Jan 2000 22:21:30 +0000 (22:21 +0000)]
Rename 7.0.0 to 7.0 to be consistent with prior release numbering.

24 years agoUpdate install file for 7.0 to match new SGML version.
Bruce Momjian [Thu, 20 Jan 2000 22:17:39 +0000 (22:17 +0000)]
Update install file for 7.0 to match new SGML version.

24 years agoAdded new pg_id to fix initdb problems
Peter Eisentraut [Thu, 20 Jan 2000 21:51:09 +0000 (21:51 +0000)]
Added new pg_id to fix initdb problems
New INSTALL file
Fixed a copyright notice

24 years agoCleanup vacuum names.
Bruce Momjian [Thu, 20 Jan 2000 20:01:25 +0000 (20:01 +0000)]
Cleanup vacuum names.

24 years agoFix typo in \z.
Bruce Momjian [Thu, 20 Jan 2000 15:29:20 +0000 (15:29 +0000)]
Fix typo in \z.

24 years agoI have noticed that the geometry test's output for hppa1 and hppa2 are in
Bruce Momjian [Thu, 20 Jan 2000 15:21:23 +0000 (15:21 +0000)]
I have noticed that the geometry test's output for hppa1 and hppa2 are in
fact the same, so I suggest they could be the same file say
geometry-positive-zeros.out, as the main difference seems to be not printing
eg. (0,-0). In src/test/regress/expected, I propose

rm int2-i386-netbsd.out int4-i386-netbsd.out
mv geometry-hppa1.1.out geometry-positive-zeros.out
rm geometry-hppa2.0.out geometry-i386-netbsd.out

and the following patch to resultmap. I have only tested the netbsd results
on i386, but think that in all probability the differences will be the same
for other ports. If it turns out not to be the case, at least we might find
out.

Patrick Welche

24 years agoBruce,
Bruce Momjian [Thu, 20 Jan 2000 15:13:19 +0000 (15:13 +0000)]
Bruce,

Attached is a small fix for a stupid mistake I made in comment.c
- an attempt to drop a non-existent comment would dump core :-(.
Sometimes, I'm as sharp as a marble.

Sorry,

Mike Mascari

24 years agoClean up longstanding warnings. I think the remaining
Tom Lane [Thu, 20 Jan 2000 05:44:34 +0000 (05:44 +0000)]
Clean up longstanding warnings.  I think the remaining
'defined but not used' warnings would go away if the scanner didn't use
YY_REJECT.

24 years agoClean up longstanding gcc warnings by adding missing extern
Tom Lane [Thu, 20 Jan 2000 05:34:32 +0000 (05:34 +0000)]
Clean up longstanding gcc warnings by adding missing extern
declarations.

24 years agoClean up longstanding gcc warnings by adding missing extern
Tom Lane [Thu, 20 Jan 2000 05:26:54 +0000 (05:26 +0000)]
Clean up longstanding gcc warnings by adding missing extern
declarations.

24 years agoecpg shouldn't depend on parser/gramparse.h. Also, eliminate some
Tom Lane [Thu, 20 Jan 2000 05:23:26 +0000 (05:23 +0000)]
ecpg shouldn't depend on parser/gramparse.h.  Also, eliminate some
compiler warnings caused by lack of extern declarations in extern.h.
I believe the remaining gcc warnings here would go away if the ecpg
grammar could be tweaked so it doesn't use REJECT ...

24 years agoWell, I finally solved the linking problem
Bruce Momjian [Thu, 20 Jan 2000 05:08:58 +0000 (05:08 +0000)]
Well, I finally solved the linking problem
that kept me from making perl secure.

Attached is uuencoded tarball to add PL/perl
to postgresql.

Things I know don't work.
-- triggers
-- SPI

The README file has a _VERY_ short tutorial.

Mark Hollomon

24 years agoTweak includes to avoid compiler warning on HPUX.
Tom Lane [Thu, 20 Jan 2000 04:11:52 +0000 (04:11 +0000)]
Tweak includes to avoid compiler warning on HPUX.

24 years agoAssign a typmod of -1 to unadorned CHAR and NUMERIC type specs. This
Tom Lane [Thu, 20 Jan 2000 02:24:50 +0000 (02:24 +0000)]
Assign a typmod of -1 to unadorned CHAR and NUMERIC type specs.  This
allows casts without specific length requirements to continue to work
as they did before; that is, x::char will not truncate the value of x,
whereas x::char(1) will.  Likewise for NUMERIC precision/scale.
The column length defaults of char(1) and numeric(30,6) are now inserted
in analyze.c's processing of CREATE TABLE.

24 years agoFix bugs in NUMERIC ceil() and floor() functions. ceil(0) returned 1,
Tom Lane [Thu, 20 Jan 2000 02:21:44 +0000 (02:21 +0000)]
Fix bugs in NUMERIC ceil() and floor() functions.  ceil(0) returned 1,
and both would insert random junk digits if given an input that was an
exact multiple of 10.

24 years agoFix handling of NULL constraint conditions: per SQL92 spec, a NULL result
Tom Lane [Wed, 19 Jan 2000 23:55:03 +0000 (23:55 +0000)]
Fix handling of NULL constraint conditions: per SQL92 spec, a NULL result
from a constraint condition does not violate the constraint (cf. discussion
on pghackers 12/9/99).  Implemented by adding a parameter to ExecQual,
specifying whether to return TRUE or FALSE when the qual result is
really NULL in three-valued boolean logic.  Currently, ExecRelCheck is
the only caller that asks for TRUE, but if we find any other places that
have the wrong response to NULL, it'll be easy to fix them.

24 years agoUpdate vacuum comments.
Bruce Momjian [Wed, 19 Jan 2000 22:23:00 +0000 (22:23 +0000)]
Update vacuum comments.

24 years agoUpdate pg_dumpall.
Bruce Momjian [Wed, 19 Jan 2000 20:10:27 +0000 (20:10 +0000)]
Update pg_dumpall.

24 years agoanother round of user interface cleanups
Peter Eisentraut [Wed, 19 Jan 2000 20:08:36 +0000 (20:08 +0000)]
another round of user interface cleanups
removed pg_id
fixed a few bugs in the scripts

24 years agoautoconf
Bruce Momjian [Wed, 19 Jan 2000 17:27:04 +0000 (17:27 +0000)]
autoconf

24 years agoFix >& csh-ism
Bruce Momjian [Wed, 19 Jan 2000 17:18:12 +0000 (17:18 +0000)]
Fix >& csh-ism

24 years agoThe latest source does not compile on Solaris 7 due to
Bruce Momjian [Wed, 19 Jan 2000 14:01:25 +0000 (14:01 +0000)]
The latest source does not compile on Solaris 7 due to
a missing include from a modified file.

Here is a patch to fix it:-

Keith Parks.

24 years agoRemoved MBFLAGS from makefiles since it's now done in include/config.h.
Peter Eisentraut [Wed, 19 Jan 2000 02:59:03 +0000 (02:59 +0000)]
Removed MBFLAGS from makefiles since it's now done in include/config.h.

24 years agoRevert back Makefile.global
Tatsuo Ishii [Wed, 19 Jan 2000 01:42:12 +0000 (01:42 +0000)]
Revert back Makefile.global
Add check for --with-mb. It is now obsoleted.

24 years agoFix for multibyte support. Since some function declarations
Tatsuo Ishii [Wed, 19 Jan 2000 01:33:16 +0000 (01:33 +0000)]
Fix for multibyte support. Since some function declarations
moved to miscadmin.h, we have to include it now.

24 years agoanother set of cleanups
Peter Eisentraut [Tue, 18 Jan 2000 23:30:24 +0000 (23:30 +0000)]
another set of cleanups

24 years agoSince this patch is not big I send it here instead. I do not have the
Bruce Momjian [Tue, 18 Jan 2000 19:08:13 +0000 (19:08 +0000)]
Since this patch is not big I send it here instead. I do not have the
complete source checked out so I cannot commit it myself.

Michael

24 years agoHi!
Bruce Momjian [Tue, 18 Jan 2000 19:05:31 +0000 (19:05 +0000)]
Hi!

Here is a patch to bring both libpq and psql to a state where it compiles on
win32 (native) again. A lot of things have changed, and I have not been able
to keep up with them all, so it has been broken for quite a while.
After this patch, at least it compiles. It also talks "basic talk" to the
server, but I have not yet tested all things. Sending queries, and using
e.g. \d or \dt works fine. The rest will have to be tested further.
It also bumps the version on libpq.dll to 7.0.

Everything should be enclosed in #ifdef WIN32, unless I have missed
something. Except for one or maybe two places where I have moved a #include
that should not be used on win32 from the "global area" into a "#ifndef
WIN32 area".

//Magnus

24 years agoBruce,
Bruce Momjian [Tue, 18 Jan 2000 18:09:02 +0000 (18:09 +0000)]
Bruce,

Attached is a patch which patches cleanly against the Sunday afternoon
snapshot. It modifies pg_dump to dump COMMENT ON statements for
user-definable descriptions. In addition, it also modifies comment.c so
that the operator behavior is as Peter E. would like: a comment on an
operator is applied to the underlying function.

Thanks,

Mike Mascari

24 years agoFix minor comple error
Tatsuo Ishii [Tue, 18 Jan 2000 13:46:10 +0000 (13:46 +0000)]
Fix minor comple error

24 years agoFix minor comping errors
Tatsuo Ishii [Tue, 18 Jan 2000 13:44:48 +0000 (13:44 +0000)]
Fix minor comping errors

24 years ago*** empty log message ***
Michael Meskes [Tue, 18 Jan 2000 13:03:49 +0000 (13:03 +0000)]
*** empty log message ***

24 years agoFix quoting bugs and incorrect trigger argument printout.
Tom Lane [Tue, 18 Jan 2000 07:29:58 +0000 (07:29 +0000)]
Fix quoting bugs and incorrect trigger argument printout.

24 years agoThis corrects an error in current gram.y for ALTER TABLE ... ADD
Bruce Momjian [Tue, 18 Jan 2000 06:12:03 +0000 (06:12 +0000)]
This corrects an error in current gram.y for ALTER TABLE ... ADD
CONSTRAINT

Oliver Elphick

24 years agoFreebsd update for sgml, from Alfred Perlstein
Bruce Momjian [Tue, 18 Jan 2000 06:10:54 +0000 (06:10 +0000)]
Freebsd update for sgml, from Alfred Perlstein

24 years agoLibpq non-blocking mode, from Alfred Perlstein
Bruce Momjian [Tue, 18 Jan 2000 06:09:24 +0000 (06:09 +0000)]
Libpq non-blocking mode, from Alfred Perlstein

24 years agoRemove compiler warnings
Tatsuo Ishii [Tue, 18 Jan 2000 05:14:24 +0000 (05:14 +0000)]
Remove compiler warnings

24 years agoShow encoding name rather than encoding id in case of psql -l.
Tatsuo Ishii [Tue, 18 Jan 2000 05:11:38 +0000 (05:11 +0000)]
Show encoding name rather than encoding id in case of psql -l.

24 years agoAdd builtin functions:
Tatsuo Ishii [Tue, 18 Jan 2000 05:10:29 +0000 (05:10 +0000)]
Add builtin functions:
pg_char_to_encoding()
pg_encoding_to_char()

24 years agonumeric_in accepts exponents; numeric to int4 rounds; float4/8 to numeric
Tom Lane [Tue, 18 Jan 2000 03:44:41 +0000 (03:44 +0000)]
numeric_in accepts exponents; numeric to int4 rounds; float4/8 to numeric
is considerably more robust and accurate than it used to be.
Also, get rid of numeric's private allocation freelist, which is no longer
a win since Jan rewrote palloc.

24 years agoFix multibyte support
Tatsuo Ishii [Tue, 18 Jan 2000 03:01:40 +0000 (03:01 +0000)]
Fix multibyte support

24 years agoAdopt for new psql
Tatsuo Ishii [Tue, 18 Jan 2000 03:00:37 +0000 (03:00 +0000)]
Adopt for new psql

24 years agoUpdated user interfaces on initdb, initlocation, pg_dump, ipcclean to a GNU-compliant...
Peter Eisentraut [Tue, 18 Jan 2000 00:03:37 +0000 (00:03 +0000)]
Updated user interfaces on initdb, initlocation, pg_dump, ipcclean to a GNU-compliant'ish state.
Made ipcclean work on Linux.

24 years agosetheapoverride() is history. Uses replaced with CommandCounterIncrement()
Tom Lane [Mon, 17 Jan 2000 23:57:48 +0000 (23:57 +0000)]
setheapoverride() is history.  Uses replaced with CommandCounterIncrement()
where necessary --- several of them didn't really need it, though.
tqual-checking macros simplified accordingly.

24 years agoCorrect minor typos.
Tom Lane [Mon, 17 Jan 2000 04:47:05 +0000 (04:47 +0000)]
Correct minor typos.

24 years agoUpdate subquery error message.
Bruce Momjian [Mon, 17 Jan 2000 04:43:56 +0000 (04:43 +0000)]
Update subquery error message.

24 years agoAdd .cvsignore so cvs update doesn't complain about derived
Tom Lane [Mon, 17 Jan 2000 04:38:49 +0000 (04:38 +0000)]
Add .cvsignore so cvs update doesn't complain about derived
files being left around.

24 years agoModify libpq's pqexpbuffer to eliminate length restriction on how much
Tom Lane [Mon, 17 Jan 2000 02:59:46 +0000 (02:59 +0000)]
Modify libpq's pqexpbuffer to eliminate length restriction on how much
data can be formatted per call.  This requires relying on vsnprintf().
On machines that haven't got vsnprintf, link in the version from
backend/port/.

24 years agoUpdate TODO list.
Bruce Momjian [Mon, 17 Jan 2000 02:43:13 +0000 (02:43 +0000)]
Update TODO list.

24 years agoPass atttypmod to CoerceTargetExpr, so that it can pass it on to
Tom Lane [Mon, 17 Jan 2000 02:04:16 +0000 (02:04 +0000)]
Pass atttypmod to CoerceTargetExpr, so that it can pass it on to
coerce_type, so that the right things happen when coercing a previously-
unknown constant to a destination data type.

24 years agoHmm, numeric array type was missing too. Added.
Tom Lane [Mon, 17 Jan 2000 01:29:07 +0000 (01:29 +0000)]
Hmm, numeric array type was missing too.  Added.
Of the standard types, only 'timestamp' seems not to have an array type;
should it be added, or are we going to remove that type for 7.0 anyway?

24 years agoFix for TODO item * spinlock stuck problem when elog(FATAL)
Hiroshi Inoue [Mon, 17 Jan 2000 01:15:19 +0000 (01:15 +0000)]
Fix for TODO item * spinlock stuck problem when elog(FATAL)
and elog(ERROR) inside bufmgr.

24 years agoUpdate unused_oids script so it works with non-GNU awk --- /* ... */
Tom Lane [Mon, 17 Jan 2000 00:53:11 +0000 (00:53 +0000)]
Update unused_oids script so it works with non-GNU awk --- /* ... */
comment style apparently isn't portable to other awks.

24 years agoApparently, no one's ever used float4abs(), because it's got incorrect
Tom Lane [Mon, 17 Jan 2000 00:40:51 +0000 (00:40 +0000)]
Apparently, no one's ever used float4abs(), because it's got incorrect
data in its pg_proc entry.  abs() doesn't require two arguments, last
I heard.

24 years agoUpdate strings test to reflect the fact that casting to char() will
Tom Lane [Mon, 17 Jan 2000 00:16:41 +0000 (00:16 +0000)]
Update strings test to reflect the fact that casting to char() will
now truncate or pad to the specified length.

24 years agoCreate a new parsetree node type, TypeCast, so that transformation of
Tom Lane [Mon, 17 Jan 2000 00:14:49 +0000 (00:14 +0000)]
Create a new parsetree node type, TypeCast, so that transformation of
SQL cast constructs can be performed during expression transformation
instead of during parsing.  This allows constructs like x::numeric(9,2)
and x::int2::float8 to behave as one would expect.

24 years agoRearrange coding in COPY so that expansible string buffer for data being
Tom Lane [Sun, 16 Jan 2000 21:37:50 +0000 (21:37 +0000)]
Rearrange coding in COPY so that expansible string buffer for data being
read is reused for successive attributes, instead of being deleted and
recreated from scratch for each value read in.  This reduces palloc/pfree
overhead a lot.  COPY IN still seems to be noticeably slower than it was
in 6.5 --- we need to figure out why.  This change takes care of the only
major performance loss I can see in copy.c itself, so the performance
problem is at a lower level somewhere.

24 years agoSigh, I'm an idiot ... I broke the async startup logic a couple days ago,
Tom Lane [Sun, 16 Jan 2000 21:18:52 +0000 (21:18 +0000)]
Sigh, I'm an idiot ... I broke the async startup logic a couple days ago,
by creating a race condition.  It wasn't waiting for select() to say
write-ready immediately after connect, which meant that you might get
an unhelpful 'broken pipe' error message if connect failed, rather than
the intended error message.

24 years agoPut back change to 'connection failed' message formatting that someone
Tom Lane [Sun, 16 Jan 2000 20:34:54 +0000 (20:34 +0000)]
Put back change to 'connection failed' message formatting that someone
overwrote.

24 years agoRemoved lextest, because lex'ed files are now in the distribution.
Peter Eisentraut [Sun, 16 Jan 2000 20:08:45 +0000 (20:08 +0000)]
Removed lextest, because lex'ed files are now in the distribution.

24 years agoIncluded all yacc and lex files into the distribution.
Peter Eisentraut [Sun, 16 Jan 2000 20:05:00 +0000 (20:05 +0000)]
Included all yacc and lex files into the distribution.

24 years agoAdd check that inherited constraints and defaults work.
Tom Lane [Sun, 16 Jan 2000 19:57:48 +0000 (19:57 +0000)]
Add check that inherited constraints and defaults work.

24 years agoRepair breakage of inherited constraint expressions --- needed a
Tom Lane [Sun, 16 Jan 2000 19:57:00 +0000 (19:57 +0000)]
Repair breakage of inherited constraint expressions --- needed a
CommandCounterIncrement to make new relation visible before trying to
parse/deparse the expressions.  Also, eliminate unnecessary
setheapoverride calls in AddNewAttributeTuples.

24 years agoFix broken FOR UPDATE error message.
Tom Lane [Sun, 16 Jan 2000 08:21:59 +0000 (08:21 +0000)]
Fix broken FOR UPDATE error message.

24 years agoUpdate TODO list.
Bruce Momjian [Sun, 16 Jan 2000 07:05:35 +0000 (07:05 +0000)]
Update TODO list.

24 years agoFix passing of atttypmod that Tom found.
Bruce Momjian [Sun, 16 Jan 2000 05:18:19 +0000 (05:18 +0000)]
Fix passing of atttypmod that Tom found.

24 years agoUpdate TODO list.
Bruce Momjian [Sun, 16 Jan 2000 05:18:07 +0000 (05:18 +0000)]
Update TODO list.

24 years agoClean up pg_dump coredumps caused by change of output formatting for
Tom Lane [Sun, 16 Jan 2000 03:54:58 +0000 (03:54 +0000)]
Clean up pg_dump coredumps caused by change of output formatting for
oidvector/int2vector.  pg_dump code was assuming that it would see
exactly FUNC_MAX_ARGS integers in the string returned by the backend.
That's no longer true.  (Perhaps that change wasn't such a good idea
after all --- will it break any other applications??)

24 years agoUpdate TODO list.
Bruce Momjian [Sun, 16 Jan 2000 02:35:30 +0000 (02:35 +0000)]
Update TODO list.

24 years agoUpdate TODO list.
Bruce Momjian [Sun, 16 Jan 2000 02:32:13 +0000 (02:32 +0000)]
Update TODO list.

24 years agoAdd some examples to numeric regress test to verify that recently-fixed
Tom Lane [Sat, 15 Jan 2000 23:44:17 +0000 (23:44 +0000)]
Add some examples to numeric regress test to verify that recently-fixed
problems are indeed fixed.

24 years agoClean up problems with rounding/overflow code in NUMERIC, particularly
Tom Lane [Sat, 15 Jan 2000 23:42:49 +0000 (23:42 +0000)]
Clean up problems with rounding/overflow code in NUMERIC, particularly
the case wherein zero was rejected for a field like NUMERIC(4,4).
Miscellaneous other code beautification efforts.

24 years agoFix a passel of problems with incorrect calls to typinput and typoutput
Tom Lane [Sat, 15 Jan 2000 22:43:25 +0000 (22:43 +0000)]
Fix a passel of problems with incorrect calls to typinput and typoutput
functions, which would lead to trouble with datatypes that paid attention
to the typelem or typmod parameters to these functions.  In particular,
incorrect code in pg_aggregate.c explains the platform-specific failures
that have been reported in NUMERIC avg().

24 years agoUpdate TODO list.
Bruce Momjian [Sat, 15 Jan 2000 19:32:06 +0000 (19:32 +0000)]
Update TODO list.

24 years agoNow that new psql is fflush()'ing properly, it emerges that several
Tom Lane [Sat, 15 Jan 2000 19:18:24 +0000 (19:18 +0000)]
Now that new psql is fflush()'ing properly, it emerges that several
regress test expected outputs were committed with NOTICEs appearing out
of order.  Update to correct results.

24 years agoUpdate arrays regress test to reflect fact that several things
Tom Lane [Sat, 15 Jan 2000 19:11:40 +0000 (19:11 +0000)]
Update arrays regress test to reflect fact that several things
work now that did not work in 6.5.

24 years ago- Allow array on int8
Peter Eisentraut [Sat, 15 Jan 2000 18:30:35 +0000 (18:30 +0000)]
- Allow array on int8
- Prevent permissions on indexes
- Instituted --enable-multibyte option and tweaked the MB build process where necessary
- initdb prompts for superuser password

24 years agoAdd pgench: a TPC-B like benchmarking tool
Tatsuo Ishii [Sat, 15 Jan 2000 12:38:09 +0000 (12:38 +0000)]
Add pgench: a TPC-B like benchmarking tool

24 years agoAdapt to the changes of libpq(eliminateing using putenv()).
Tatsuo Ishii [Sat, 15 Jan 2000 05:38:50 +0000 (05:38 +0000)]
Adapt to the changes of libpq(eliminateing using putenv()).

24 years agoEliminate using putenv().
Tatsuo Ishii [Sat, 15 Jan 2000 05:37:21 +0000 (05:37 +0000)]
Eliminate using putenv().

24 years agoPrepare for new psql
Tatsuo Ishii [Sat, 15 Jan 2000 05:17:45 +0000 (05:17 +0000)]
Prepare for new psql

24 years agoFixed all elog related warnings, as well as a few others.
Peter Eisentraut [Sat, 15 Jan 2000 02:59:43 +0000 (02:59 +0000)]
Fixed all elog related warnings, as well as a few others.

24 years agoFixed psql variables vs array syntax, as well as minor psql enhancements
Peter Eisentraut [Fri, 14 Jan 2000 22:18:03 +0000 (22:18 +0000)]
Fixed psql variables vs array syntax, as well as minor psql enhancements