]> granicus.if.org Git - postgresql/log
postgresql
25 years agoMultibyte tests were all 'failing' because of change of
Tom Lane [Sun, 25 Apr 1999 19:34:44 +0000 (19:34 +0000)]
Multibyte tests were all 'failing' because of change of
wording of 'relation does not exist' error message.  Update expected files
accordingly.

25 years agoMy first cut at libpq revision didn't handle MULTIBYTE correctly,
Tom Lane [Sun, 25 Apr 1999 19:27:47 +0000 (19:27 +0000)]
My first cut at libpq revision didn't handle MULTIBYTE correctly,
but I think it's OK now...

25 years agoClean up gcc warning in MULTIBYTE code.
Tom Lane [Sun, 25 Apr 1999 18:16:47 +0000 (18:16 +0000)]
Clean up gcc warning in MULTIBYTE code.

25 years agoifdef out some unused routines to suppress gcc warnings.
Tom Lane [Sun, 25 Apr 1999 18:09:54 +0000 (18:09 +0000)]
ifdef out some unused routines to suppress gcc warnings.

25 years agoOoops, missed committing this one...
Tom Lane [Sun, 25 Apr 1999 03:27:15 +0000 (03:27 +0000)]
Ooops, missed committing this one...

25 years agoRevise backend libpq interfaces so that messages to the frontend
Tom Lane [Sun, 25 Apr 1999 03:19:27 +0000 (03:19 +0000)]
Revise backend libpq interfaces so that messages to the frontend
can be generated in a buffer and then sent to the frontend in a single
libpq call.  This solves problems with NOTICE and ERROR messages generated
in the middle of a data message or COPY OUT operation.

25 years agoRename explain's "size" to "rows".
Bruce Momjian [Fri, 23 Apr 1999 21:23:49 +0000 (21:23 +0000)]
Rename explain's "size" to "rows".

25 years agoAdd disk space message to "can not extend" message.
Bruce Momjian [Fri, 23 Apr 1999 19:37:42 +0000 (19:37 +0000)]
Add disk space message to "can not extend" message.

25 years agoRemove "Non-functional update" notices since that message has been
Thomas G. Lockhart [Fri, 23 Apr 1999 15:55:32 +0000 (15:55 +0000)]
Remove "Non-functional update" notices since that message has been
 disabled (commented-out) in the code.

25 years agoRepair missing heap_endscan() in OperatorUpd().
Tom Lane [Fri, 23 Apr 1999 00:50:57 +0000 (00:50 +0000)]
Repair missing heap_endscan() in OperatorUpd().

25 years agoFix duplicating ROOT page in concurrent updates.
Vadim B. Mikheev [Thu, 22 Apr 1999 08:19:59 +0000 (08:19 +0000)]
Fix duplicating ROOT page in concurrent updates.

25 years agoUn-break CREATE TYPE. Fix some other inconsistencies in the
Tom Lane [Tue, 20 Apr 1999 03:51:19 +0000 (03:51 +0000)]
Un-break CREATE TYPE.  Fix some other inconsistencies in the
pg_proc entries for array I/O routines besides the one detected by the
original patcher.  Tighten type_sanity regress test accordingly.

25 years agoChange elog(ERROR) to get back to main loop via a plain sigsetjmp,
Tom Lane [Tue, 20 Apr 1999 02:19:59 +0000 (02:19 +0000)]
Change elog(ERROR) to get back to main loop via a plain sigsetjmp,
instead of doing a kill(self, SIGQUIT) and expecting the signal handler
to do it.  Also, clean up inconsistent definitions of the sigjmp buffer
in the several files that already referenced it.

25 years agoPortability patches for HPUX 11 and Unixware in configure
Tom Lane [Tue, 20 Apr 1999 00:26:32 +0000 (00:26 +0000)]
Portability patches for HPUX 11 and Unixware in configure
and related files.  Also remove float.c's gratuitous redeclaration of
isinf() ... looks like there are more decls in there that ought to be
in config.h, but I'll leave well enough alone for now ...

25 years agoAdd temporary for temp.
Bruce Momjian [Mon, 19 Apr 1999 16:00:18 +0000 (16:00 +0000)]
Add temporary for temp.

25 years agoFix typo in multi-byte encodings (EHC_CN --> EUC_CN)
Tatsuo Ishii [Mon, 19 Apr 1999 08:46:34 +0000 (08:46 +0000)]
Fix typo in multi-byte encodings (EHC_CN --> EUC_CN)

25 years agoFix problems seen when result of a subselect was used in an
Tom Lane [Mon, 19 Apr 1999 04:17:11 +0000 (04:17 +0000)]
Fix problems seen when result of a subselect was used in an
expression context (ie, not at the top level of a WHERE clause).  Examples
like this one work now:
SELECT name, value FROM t1 as touter WHERE
(value/(SELECT AVG(value) FROM t1 WHERE name = touter.name)) > 0.75;

25 years agoRepair some problems in planner's handling of HAVING clauses.
Tom Lane [Mon, 19 Apr 1999 01:43:12 +0000 (01:43 +0000)]
Repair some problems in planner's handling of HAVING clauses.
This fixes a few of the problems Hiroshi Inoue complained of, but
I have not touched the rewrite-related issues.

25 years agoAfter transforming a CASE expr with a default argument,
Tom Lane [Sun, 18 Apr 1999 17:35:51 +0000 (17:35 +0000)]
After transforming a CASE expr with a default argument,
delete the default argument from the node.  This prevents the executor
from spitting up on the untransformed argument expression.  Typical
failure was:
select (case f1 when 'val' then 'subst' else f1 end) from t1;
ERROR:  copyObject: don't know how to copy 704

25 years agoUpdate CREATE FUNCTION's on-line help in psql.
Tom Lane [Sun, 18 Apr 1999 03:01:49 +0000 (03:01 +0000)]
Update CREATE FUNCTION's on-line help in psql.

25 years agoAllow CREATE FUNCTION xyz AS '' LANGUAGE 'internal' to
Tom Lane [Sun, 18 Apr 1999 02:57:22 +0000 (02:57 +0000)]
Allow CREATE FUNCTION xyz AS '' LANGUAGE 'internal' to
work the way it used to (ie, assume xyz is the name of the builtin
function to call).  Complain if an unknown builtin function name is
referenced.

25 years agoAdd res checking to libpq examples, from Dan Merillat.
Bruce Momjian [Sat, 17 Apr 1999 17:18:41 +0000 (17:18 +0000)]
Add res checking to libpq examples, from Dan Merillat.

25 years agoAdd missing function prototypes to stifle gcc warnings.
Tom Lane [Fri, 16 Apr 1999 21:27:23 +0000 (21:27 +0000)]
Add missing function prototypes to stifle gcc warnings.

25 years ago*** empty log message ***
Michael Meskes [Fri, 16 Apr 1999 12:26:49 +0000 (12:26 +0000)]
*** empty log message ***

25 years agoFix kill() call in elog() so that it gets its own pid by calling getpid().
Tatsuo Ishii [Fri, 16 Apr 1999 06:38:17 +0000 (06:38 +0000)]
Fix kill() call in elog() so that it gets its own pid by calling getpid().
MyProcPid global variable is set to 0 when postgres starts as a command
(not as a backend daemon). This leads issuing SIGQUIT to the process group,
not the process itself. As a result, parent sh gets core dumped in the
Wisconsin benchmark test.

25 years agoOverhaul Wisconsin Benchmark test suit
Tatsuo Ishii [Fri, 16 Apr 1999 06:31:13 +0000 (06:31 +0000)]
Overhaul Wisconsin Benchmark test suit
- change temp -> temp_bench ("temp" is now a reserved word)
- fix bugs in queries
- add -B 256 option to run the postgres command
  (without this, postgres seems to fail with hashjoin)

25 years agoImprove error messages when a connection is rejected.
Tom Lane [Fri, 16 Apr 1999 04:59:03 +0000 (04:59 +0000)]
Improve error messages when a connection is rejected.

25 years agoFix max(int8) result by making sure int8larger() copies its result
Thomas G. Lockhart [Thu, 15 Apr 1999 13:34:45 +0000 (13:34 +0000)]
Fix max(int8) result by making sure int8larger() copies its result
 rather than reusing the input storage.
Also made the same fix to int8smaller(), though there wasn't a symptom,
 and went through and verified that other pass-by-reference data types
 do the same thing. Not an issue for the by-value types.

25 years agoHello!
Tatsuo Ishii [Thu, 15 Apr 1999 07:25:05 +0000 (07:25 +0000)]
Hello!

   Attached is new README.locale. Should go into .../doc. Includes
instructions on how to install and use Multibyte support. Primarily
oriented toward Russian (Cyrillic) users.

Oleg.
----
    Oleg Broytmann     http://members.xoom.com/phd2/     phd2@earthling.net
           Programmers don't die, they just GOSUB without RETURN.

25 years agoReplaced <literal></literal> with "" in ProgramListing sections
Peter Mount [Thu, 15 Apr 1999 06:00:16 +0000 (06:00 +0000)]
Replaced <literal></literal> with "" in ProgramListing sections

25 years agoInitialize reltuples = 1000, relpages = 10 in a newly created
Tom Lane [Thu, 15 Apr 1999 04:08:07 +0000 (04:08 +0000)]
Initialize reltuples = 1000, relpages = 10 in a newly created
relation, rather than zeroes.  This prevents the optimizer from making
foolish choices (ie, using nested-loop plans) on never-yet-vacuumed tables.
This is a hack, of course.  Keeping accurate track of these statistics
would be a cleaner solution, but it's far from clear that it'd be worth
the cost of doing so.  In any case we're not going to do that for 6.5.
In the meantime, this quick hack provides a useful performance improvement
in the regression tests and in many real-world scenarios.

25 years agoAdd ORDER BY clauses to some select-from-view operations
Tom Lane [Thu, 15 Apr 1999 03:40:50 +0000 (03:40 +0000)]
Add ORDER BY clauses to some select-from-view operations
in rules regression test, in order to eliminate bogus test 'failures'
that occur due to platform-dependent and join-implementation-dependent
ordering of tuples.  I'm not sure that I got all of the SELECTs that need
ordering clauses --- we may need some more.  But this takes care of the
diffs between my platform and Jan's.

25 years agopsql did the wrong thing with COPY FROM STDIN inside a file
Tom Lane [Thu, 15 Apr 1999 02:24:41 +0000 (02:24 +0000)]
psql did the wrong thing with COPY FROM STDIN inside a file
sourced with \i (tried to read data from the terminal, rather than from
the source file; this breaks pg_dump scripts read with \i).  Also, \o file
followed by COPY TO STDOUT wrote to terminal not designated file.
All better now.

25 years agoMove some useful date/time test macros to here to allow
Thomas G. Lockhart [Thu, 15 Apr 1999 02:24:23 +0000 (02:24 +0000)]
Move some useful date/time test macros to here to allow
 sharing across files.

25 years agoDeclare hashint8().
Thomas G. Lockhart [Thu, 15 Apr 1999 02:23:37 +0000 (02:23 +0000)]
Declare hashint8().

25 years agoFix code to check legal dates *before* calling localtime() to get the
Thomas G. Lockhart [Thu, 15 Apr 1999 02:22:39 +0000 (02:22 +0000)]
Fix code to check legal dates *before* calling localtime() to get the
 time zone.
Previously, localtime() rotated a date with a day of month field which
 exceeded the actual range into the next months, masking the fact that
 a bad date had been specified.
Regression tests pass.

25 years agoFix boolean assignment of return values to use "FALSE" rather than the
Thomas G. Lockhart [Thu, 15 Apr 1999 02:20:50 +0000 (02:20 +0000)]
Fix boolean assignment of return values to use "FALSE" rather than the
 mis-copied "NULL", which happens to have the same binary value.
Previously, gcc gave non-fatal warnings.

25 years agoCode modified to reject out of range day of month.
Thomas G. Lockhart [Thu, 15 Apr 1999 02:19:02 +0000 (02:19 +0000)]
Code modified to reject out of range day of month.
Previously, dates falling within Unix system time range were run through
 a call to localtime() to get the time zone, if it was not specified.
This had the effect that dates with DOMs which were larger than would be
 valid for that month were "rotated" into the following months.

25 years agoFix error message to match that returned by new regression test reference
Thomas G. Lockhart [Thu, 15 Apr 1999 02:15:36 +0000 (02:15 +0000)]
Fix error message to match that returned by new regression test reference
 machine (linux-2.0.36 RH5.2 with RH5.2 patches).

25 years agoCorrect documentation of CREATE OPERATOR.
Tom Lane [Thu, 15 Apr 1999 00:09:00 +0000 (00:09 +0000)]
Correct documentation of CREATE OPERATOR.

25 years agoCorrect psql's online help for CREATE OPERATOR, which did
Tom Lane [Wed, 14 Apr 1999 23:48:07 +0000 (23:48 +0000)]
Correct psql's online help for CREATE OPERATOR, which did
not match what the backend actually accepts.

25 years agopg_dump was trying to use an incorrect (or, perhaps, only obsolete?)
Tom Lane [Wed, 14 Apr 1999 23:47:19 +0000 (23:47 +0000)]
pg_dump was trying to use an incorrect (or, perhaps, only obsolete?)
syntax for CREATE OPERATOR with SORT parameters.  Fixed.
It is now actually possible to dump and reload a database containing
fully specified user-definable operators ...

25 years ago*** empty log message ***
Michael Meskes [Wed, 14 Apr 1999 18:51:37 +0000 (18:51 +0000)]
*** empty log message ***

25 years agouse new config.guess output for system specific expected files ..
Marc G. Fournier [Wed, 14 Apr 1999 14:04:41 +0000 (14:04 +0000)]
use new config.guess output for system specific expected files ..

25 years agouse config.guess instead of uname -s to figure out system, so that we can
Marc G. Fournier [Wed, 14 Apr 1999 14:02:00 +0000 (14:02 +0000)]
use config.guess instead of uname -s to figure out system, so that we can
include platform spcific changes ...

thanks to Mark Hollomon <mhh@nortelnetworks.com> for the awk script used

25 years agoFix adding columns to a temp table.
Bruce Momjian [Tue, 13 Apr 1999 19:04:33 +0000 (19:04 +0000)]
Fix adding columns to a temp table.

25 years agoAdd ARM32 support by Andrew McMurry
Bruce Momjian [Tue, 13 Apr 1999 17:42:26 +0000 (17:42 +0000)]
Add ARM32 support by Andrew McMurry

25 years agoarray_in is defined in the system catalog as taking two arguments while it
Bruce Momjian [Tue, 13 Apr 1999 17:28:35 +0000 (17:28 +0000)]
array_in is defined in the system catalog as taking two arguments while it
actually takes three. Please apply the following patch.

Massimo

25 years agoThere are some bugs about backward scanning using
Bruce Momjian [Tue, 13 Apr 1999 17:18:29 +0000 (17:18 +0000)]
There are some bugs about backward scanning using
indexes.

1. Index Scan using plural indexids never scan backward
   as to the order of indexids.
2. The cursor using Index scan is not usable after moving
   past the end.

This patch solves above bugs.
Moreover the change of _bt_first() would be useful to extend
ORDER BY patch by Jan Wieck for all descending order cases.

Hiroshi Inoue

25 years ago*** empty log message ***
Michael Meskes [Tue, 13 Apr 1999 12:36:38 +0000 (12:36 +0000)]
*** empty log message ***

25 years agoAdd -E option
Tatsuo Ishii [Tue, 13 Apr 1999 09:46:58 +0000 (09:46 +0000)]
Add -E option

25 years agoDocument updation for 6.5
Tatsuo Ishii [Tue, 13 Apr 1999 09:29:46 +0000 (09:29 +0000)]
Document updation for 6.5

25 years agoFix HEAP_MOVED_IN handling in HeapTupleSatisfiesDirty().
Vadim B. Mikheev [Mon, 12 Apr 1999 16:57:27 +0000 (16:57 +0000)]
Fix HEAP_MOVED_IN handling in HeapTupleSatisfiesDirty().

25 years agoSome comments...
Vadim B. Mikheev [Mon, 12 Apr 1999 16:56:36 +0000 (16:56 +0000)]
Some comments...

25 years agoUnique btree-s:
Vadim B. Mikheev [Mon, 12 Apr 1999 16:56:08 +0000 (16:56 +0000)]
Unique btree-s:
/*
 * Have to check is inserted heap tuple deleted one
 * (i.e. just moved to another place by vacuum)!
 */

25 years agoadd mipsel-unknown-linux-gnu support. config.guess seems slightly changed
Tatsuo Ishii [Mon, 12 Apr 1999 04:04:54 +0000 (04:04 +0000)]
add mipsel-unknown-linux-gnu support. config.guess seems slightly changed
since April 1?

25 years agoImplement UpdateCount
Peter Mount [Sun, 11 Apr 1999 18:03:00 +0000 (18:03 +0000)]
Implement UpdateCount

25 years agoRepair problems seen when CREATE OPERATOR mentions a
Tom Lane [Sun, 11 Apr 1999 02:30:59 +0000 (02:30 +0000)]
Repair problems seen when CREATE OPERATOR mentions a
not-yet-defined operator in commutator, negator, etc links.  This is
necessary in order to ensure that a pg_dump dump of user-defined operators
can be reloaded.  There may still be a bug lurking here, because it's
provoking a 'Buffer Leak' notice message in one case.  See my mail to
pgsql-hackers.

25 years agoFix another batch of bogosities in pg_operator table.
Tom Lane [Sat, 10 Apr 1999 23:53:00 +0000 (23:53 +0000)]
Fix another batch of bogosities in pg_operator table.
These were bogus selectivity-estimator links, like a '>' operator
pointing to intltsel when it should use intgtsel.

25 years agoTweak create_operator regress test so it doesn't illustrate
Tom Lane [Sat, 10 Apr 1999 23:51:16 +0000 (23:51 +0000)]
Tweak create_operator regress test so it doesn't illustrate
incorrect usage of commutator link.

25 years agovacuumlo deals with large objects not referenced by any tables and removes them.
Peter Mount [Sat, 10 Apr 1999 16:48:05 +0000 (16:48 +0000)]
vacuumlo deals with large objects not referenced by any tables and removes them.

25 years agoFix CREATE OPERATOR ... LANGUAGE 'internal', which I broke while
Tom Lane [Fri, 9 Apr 1999 22:35:43 +0000 (22:35 +0000)]
Fix CREATE OPERATOR ... LANGUAGE 'internal', which I broke while
making prosrc instead of proname be the link to the actual internal function.

25 years agoAdd section from Tom Lane on hashjoin characteristics of operators.
Thomas G. Lockhart [Thu, 8 Apr 1999 13:29:08 +0000 (13:29 +0000)]
Add section from Tom Lane on hashjoin characteristics of operators.
Add emacs editor hints to bottom of file.

25 years agoAdd emacs editor hints to bottom of file.
Thomas G. Lockhart [Thu, 8 Apr 1999 13:28:22 +0000 (13:28 +0000)]
Add emacs editor hints to bottom of file.

25 years agoFix some more hashjoin-related bugs in pg_operator. Fix
Tom Lane [Wed, 7 Apr 1999 23:33:33 +0000 (23:33 +0000)]
Fix some more hashjoin-related bugs in pg_operator.  Fix
hashjoin's hashFunc() so that it does the right thing with pass-by-value
data types (the old code would always return 0 for int2 or char values,
which would work but would slow things down a lot).  Extend opr_sanity
regress test to catch more kinds of errors.

25 years agoPrepare multi-byte support document for 6.5
Tatsuo Ishii [Wed, 7 Apr 1999 05:14:51 +0000 (05:14 +0000)]
Prepare multi-byte support document for 6.5

25 years agoUnmark 'hashable' operators that can't really be used for
Tom Lane [Wed, 7 Apr 1999 04:21:11 +0000 (04:21 +0000)]
Unmark 'hashable' operators that can't really be used for
hashjoins.  Extend opr_sanity regress test to help detect similar mistakes.

25 years agoFix markup to include all closing tags.
Thomas G. Lockhart [Tue, 6 Apr 1999 15:41:20 +0000 (15:41 +0000)]
Fix markup to include all closing tags.
Required by newest versions of DocBook and style sheets.

25 years agoStart indenting markup. No text changes.
Thomas G. Lockhart [Tue, 6 Apr 1999 15:38:10 +0000 (15:38 +0000)]
Start indenting markup. No text changes.

25 years agoFix XREF markup and typo.
Thomas G. Lockhart [Tue, 6 Apr 1999 15:36:56 +0000 (15:36 +0000)]
Fix XREF markup and typo.

25 years agoVersion string was truncating the last character.
Thomas G. Lockhart [Tue, 6 Apr 1999 15:35:36 +0000 (15:35 +0000)]
Version string was truncating the last character.
Use memcpy() rather than StrNCpy() which had forced a string termination
 character at the end.

25 years agoOn reflection, filesize limit ought to be an exact power
Tom Lane [Tue, 6 Apr 1999 03:04:22 +0000 (03:04 +0000)]
On reflection, filesize limit ought to be an exact power
of 2 to save a few cycles in md.c.  So, make it 2^30 not 10^9.

25 years agoReduce default file size limit to 1Gb, and move the
Tom Lane [Mon, 5 Apr 1999 22:25:11 +0000 (22:25 +0000)]
Reduce default file size limit to 1Gb, and move the
configuration constant to config.h.

25 years agoFix HEAP_MOVED_OFF handling in HeapTupleSatisfiesDirty
Vadim B. Mikheev [Mon, 5 Apr 1999 10:55:49 +0000 (10:55 +0000)]
Fix HEAP_MOVED_OFF handling in HeapTupleSatisfiesDirty
(may be used while vacuuming unique btree-s).

25 years agoFix potential overflow problems when relation size exceeds
Tom Lane [Mon, 5 Apr 1999 02:07:07 +0000 (02:07 +0000)]
Fix potential overflow problems when relation size exceeds
2gig.  Fix failure to reliably put the smaller relation on the inside of
a hashjoin.

25 years agoWe use CFLAGS = -O2 on every other gcc platform, so why not
Tom Lane [Sun, 4 Apr 1999 20:16:50 +0000 (20:16 +0000)]
We use CFLAGS = -O2 on every other gcc platform, so why not
hpux_gcc too?

25 years agoUse MAXALIGN value found by configure instead of a
Tom Lane [Sun, 4 Apr 1999 20:10:12 +0000 (20:10 +0000)]
Use MAXALIGN value found by configure instead of a
hardwired assumption.

25 years agoGive a better error message if an SQL-language function is
Tom Lane [Sat, 3 Apr 1999 22:57:29 +0000 (22:57 +0000)]
Give a better error message if an SQL-language function is
called through fmgr.  Someday we should try to actually execute the function,
but that looks like it might be a major feature addition.
Not something to try during beta phase.

25 years agoFix bogus pg_amop entries for int8 hash, cidr btree,
Tom Lane [Sat, 3 Apr 1999 18:07:47 +0000 (18:07 +0000)]
Fix bogus pg_amop entries for int8 hash, cidr btree,
polygon rtree, circle rtree indexes.

25 years agoFix nasty bug in optimization of multiway joins: optimizer
Tom Lane [Sat, 3 Apr 1999 00:18:28 +0000 (00:18 +0000)]
Fix nasty bug in optimization of multiway joins: optimizer
would sometimes generate a plan that omitted a sort step before merge.

25 years agoRepair problems with omitted password and VALID UNTIL
Tom Lane [Fri, 2 Apr 1999 06:16:36 +0000 (06:16 +0000)]
Repair problems with omitted password and VALID UNTIL
parameters in CREATE USER.

25 years agoRemove overly presumptuous use of __STDC__ in c.h, replacing
Tom Lane [Fri, 2 Apr 1999 05:10:16 +0000 (05:10 +0000)]
Remove overly presumptuous use of __STDC__ in c.h, replacing
it with configure-script tests to see whether const, inline, volatile, etc
work or not.  (Curiously, configure was already doing the work to see if
const and inline were OK, but the results were not getting plugged into
config.h :-(.)

25 years agoClean up compile errors and warnings, cf Billy Allie's
Tom Lane [Fri, 2 Apr 1999 04:51:05 +0000 (04:51 +0000)]
Clean up compile errors and warnings, cf Billy Allie's
complaints (and some of my own).

25 years agoUpdate config.guess and config.sub to versions from
Tom Lane [Fri, 2 Apr 1999 03:37:17 +0000 (03:37 +0000)]
Update config.guess and config.sub to versions from
autoconf 2.13.

25 years agoFix up small markup problems. Force omit-tags to nil so we have tag
Thomas G. Lockhart [Tue, 30 Mar 1999 15:25:56 +0000 (15:25 +0000)]
Fix up small markup problems. Force omit-tags to nil so we have tag
 completion as required by the newest DocBook conventions.

25 years agoRepair cut-and-paste typo which described createdb rather than destroydb.
Thomas G. Lockhart [Tue, 30 Mar 1999 15:24:49 +0000 (15:24 +0000)]
Repair cut-and-paste typo which described createdb rather than destroydb.

25 years agoInclude libpq+.sgml in the doc set.
Thomas G. Lockhart [Tue, 30 Mar 1999 15:23:03 +0000 (15:23 +0000)]
Include libpq+.sgml in the doc set.
Had been omitted from the integrated doc for no good reason.

25 years agoUpdates from Vince Vielhaber are the first since 1997.
Thomas G. Lockhart [Tue, 30 Mar 1999 15:21:42 +0000 (15:21 +0000)]
Updates from Vince Vielhaber are the first since 1997.

25 years agoMarkup fixes from Vince Vielhaber.
Thomas G. Lockhart [Tue, 30 Mar 1999 15:20:43 +0000 (15:20 +0000)]
Markup fixes from Vince Vielhaber.

25 years agoAdd -E option
Tatsuo Ishii [Tue, 30 Mar 1999 05:14:03 +0000 (05:14 +0000)]
Add -E option

25 years agoBug fixes and enhances to psql submitted by Masaaki Sakaida
Tatsuo Ishii [Tue, 30 Mar 1999 05:00:42 +0000 (05:00 +0000)]
Bug fixes and enhances to psql submitted by Masaaki Sakaida

1. Fix problems of PAGER and \? command
2. Add -E option that shows actual queries sent by \dt and friends
3. Add version number in startup banners for psql

25 years agoSmall cleanups.
Bruce Momjian [Tue, 30 Mar 1999 01:37:28 +0000 (01:37 +0000)]
Small cleanups.

25 years agoAddition to test/locale submitted by Oleg Broytmann. Comments from him:
Tatsuo Ishii [Mon, 29 Mar 1999 09:00:19 +0000 (09:00 +0000)]
Addition to test/locale submitted by Oleg Broytmann. Comments from him:

There are two subdirectories (ISO8859-7 and koi8-to-win1251) containing
tests for Greek locale and server<=>client recoding feature (recently
submitted by Tatsuo Ishii <t-ishii@sra.co.jp>; we've debugged his patches
together in the field of Cyrillic support).

25 years agochange comparison char* and NULL to char* and '\0'. This should be
Tatsuo Ishii [Mon, 29 Mar 1999 08:19:36 +0000 (08:19 +0000)]
change comparison char* and NULL to char* and '\0'. This should be
more portable way.

25 years agoModify fmgr so that internal name (compiler name) of a built-in
Tom Lane [Mon, 29 Mar 1999 01:30:45 +0000 (01:30 +0000)]
Modify fmgr so that internal name (compiler name) of a built-in
function is found in prosrc field of pg_proc, not proname.  This allows
multiple aliases of a built-in to all be implemented as direct builtins,
without needing a level of indirection through an SQL function.  Replace
existing SQL alias functions with builtin entries accordingly.
Save a few K by not storing string names of builtin functions in fmgr's
internal table (if you really want 'em, get 'em from pg_proc...).
Update opr_sanity with a few more cross-checks.

25 years ago1. Vacuum is updated for MVCC.
Vadim B. Mikheev [Sun, 28 Mar 1999 20:32:42 +0000 (20:32 +0000)]
1. Vacuum is updated for MVCC.
2. Much faster btree tuples deletion in the case when first on page
   index tuple is deleted (no movement to the left page(s)).
3. Remember blkno of new root page in BTPageOpaque of
   left/right siblings when root page is splitted.

25 years agoExpanded opr_sanity test to look at pg_proc and other
Tom Lane [Sun, 28 Mar 1999 02:07:58 +0000 (02:07 +0000)]
Expanded opr_sanity test to look at pg_proc and other
related tables.

25 years agoNew regression test to cross-check pg_type, pg_class,
Tom Lane [Sun, 28 Mar 1999 02:06:23 +0000 (02:06 +0000)]
New regression test to cross-check pg_type, pg_class,
and related tables.

25 years agoClean up various minor irregularities detected by type_sanity
Tom Lane [Sun, 28 Mar 1999 02:01:39 +0000 (02:01 +0000)]
Clean up various minor irregularities detected by type_sanity
and newly expanded opr_sanity tests.

25 years agoRemove pg_attribute_check.sql: these checks merged into new
Tom Lane [Sun, 28 Mar 1999 01:57:32 +0000 (01:57 +0000)]
Remove pg_attribute_check.sql: these checks merged into new
regress test 'type_sanity'.