Bruce Momjian [Mon, 10 May 1999 04:19:30 +0000 (04:19 +0000)]
add mention of large object orphaning
Bruce Momjian [Mon, 10 May 1999 04:02:07 +0000 (04:02 +0000)]
Fix initdb problem introduced by recent patch.
Bruce Momjian [Mon, 10 May 1999 00:46:32 +0000 (00:46 +0000)]
Change error messages to oids come out as %u and not %d. Change has no
real affect now.
Tom Lane [Sun, 9 May 1999 23:43:25 +0000 (23:43 +0000)]
Correct type_sanity test so it doesn't spit up on new
definition of numeric_in.
Tom Lane [Sun, 9 May 1999 23:31:47 +0000 (23:31 +0000)]
Rearrange top-level rewrite operations so that EXPLAIN works
on queries involving UNION, EXCEPT, INTERSECT.
Tatsuo Ishii [Sun, 9 May 1999 15:00:18 +0000 (15:00 +0000)]
Running lo_read/lo_write under different memory context
cause troubles. See
Message-Id: <
199905090312.MAA00466@ext16.sra.co.jp>
for more details.
Vadim B. Mikheev [Sun, 9 May 1999 14:00:29 +0000 (14:00 +0000)]
Mistyping by me.
Fixed by Hiroshi.
Tom Lane [Sun, 9 May 1999 02:22:16 +0000 (02:22 +0000)]
Repair incorrect pg_proc entry for numeric_in.
Tom Lane [Sun, 9 May 1999 00:54:30 +0000 (00:54 +0000)]
Fix some miscellaneous places that were using raw open() or
fopen(), instead of going through fd.c ... naughty naughty.
Tom Lane [Sun, 9 May 1999 00:53:22 +0000 (00:53 +0000)]
Update hash and join routines to use fd.c's new temp-file
code, instead of not-very-bulletproof stuff they had before.
Tom Lane [Sun, 9 May 1999 00:52:08 +0000 (00:52 +0000)]
Add 'temporary file' facility to fd.c, and arrange for temp
files to be closed automatically at transaction abort or commit, should
they still be open. Also close any still-open stdio files allocated with
AllocateFile at abort/commit. This should eliminate problems with leakage
of file descriptors after an error. Also, put in some primitive buffered-IO
support so that psort.c can use virtual files without severe performance
penalties.
Bruce Momjian [Sat, 8 May 1999 19:28:20 +0000 (19:28 +0000)]
Add new netbsd file.
Michael Meskes [Fri, 7 May 1999 18:03:37 +0000 (18:03 +0000)]
*** empty log message ***
Bruce Momjian [Fri, 7 May 1999 02:46:24 +0000 (02:46 +0000)]
Remove unused Netbsd files.
Bruce Momjian [Fri, 7 May 1999 02:37:08 +0000 (02:37 +0000)]
Please apply the following patch for regress.sh to do something useful with
"SYSTEM", and unpack the files in the uuencoded .tar.gz file at the end in
src/test/regress so that the int2, int4 and geometry tests pass on NetBSD/i386.
They just fail on different wording of error messages and eg printing "0"
rather than "-0". At a guess the same will be true for the other NetBSD ports,
but I can't test them.
Cheers,
Patrick
Bruce Momjian [Fri, 7 May 1999 02:31:43 +0000 (02:31 +0000)]
Please apply the following patch for regress.sh to do something useful
with
"SYSTEM", Patrick Welche
Vadim B. Mikheev [Fri, 7 May 1999 01:23:11 +0000 (01:23 +0000)]
Fix LMGR for MVCC.
Get rid of Extend lock mode.
Tom Lane [Thu, 6 May 1999 23:09:30 +0000 (23:09 +0000)]
Update regress test for CASE to enable tests involving joins.
Tom Lane [Thu, 6 May 1999 23:07:33 +0000 (23:07 +0000)]
Fix oversights in flatten_tlistentry and replace_clause_joinvar_refs
that led to CASE expressions not working very well in joined queries.
Tom Lane [Thu, 6 May 1999 01:30:58 +0000 (01:30 +0000)]
fix_indxqual_references didn't cope with ArrayRef nodes,
meaning that this failed:
select proname,typname,prosrc from pg_proc,pg_type
where proname = 'float8' and pg_proc.proargtypes[0] = pg_type.oid;
Tom Lane [Thu, 6 May 1999 00:30:47 +0000 (00:30 +0000)]
Fix some nasty coredump bugs in hashjoin. This code was just
about certain to fail anytime it decided the relation to be hashed was
too big to fit in memory --- the code for 'batching' a series of hashjoins
had multiple errors. I've fixed the easier problems. A remaining big
problem is that you can get 'hashtable out of memory' if the code's
guesstimate about how much overflow space it will need turns out wrong.
That will require much more extensive revisions to fix, so I'm committing
these fixes now before I start on that problem.
Bruce Momjian [Wed, 5 May 1999 21:38:40 +0000 (21:38 +0000)]
I have two patches for 6.5.0:
arrayfuncs.patch fixes a small bug in my previous patches for
arrays
array-regress.patch adds _bpchar and _varchar to regression tests
--
Massimo Dal Zotto
Bruce Momjian [Wed, 5 May 1999 13:51:38 +0000 (13:51 +0000)]
small error message improvement from Dmitry Samersoff
Tom Lane [Tue, 4 May 1999 23:39:20 +0000 (23:39 +0000)]
Repair incorrectly-figured snprintf length restriction.
Thomas G. Lockhart [Tue, 4 May 1999 15:50:24 +0000 (15:50 +0000)]
Use sprintf() to convert float8 to a string during conversion to numeric.
Original code used float8out(), but the resulting exponential notation
was not handled (e.g. '3E9' was decoded as '3').
Thomas G. Lockhart [Tue, 4 May 1999 15:47:35 +0000 (15:47 +0000)]
Allow -t tablename to preserve case if specified as "tablename".
Thomas G. Lockhart [Tue, 4 May 1999 02:57:13 +0000 (02:57 +0000)]
Fix up examples of sh/ksh PATH environment variable definition.
Thomas G. Lockhart [Tue, 4 May 1999 02:43:55 +0000 (02:43 +0000)]
Clean up markup for first useful version.
Thomas G. Lockhart [Tue, 4 May 1999 02:26:06 +0000 (02:26 +0000)]
Include new introductory chapter on SQL from Stefan S.
Should this be in the tutorial instead?
Thomas G. Lockhart [Tue, 4 May 1999 02:24:26 +0000 (02:24 +0000)]
Indent markup using emacs.
Thomas G. Lockhart [Tue, 4 May 1999 02:22:13 +0000 (02:22 +0000)]
Include specific info on available timezones.
Document date/time input parsing procedure.
Thomas G. Lockhart [Tue, 4 May 1999 02:19:20 +0000 (02:19 +0000)]
Include chapters on security and an intro to SQL.
Thomas G. Lockhart [Tue, 4 May 1999 02:18:34 +0000 (02:18 +0000)]
First version of markup which can run through jade for html.
Probably not acceptable yet. Need to check some math symbols and markup.
Thomas G. Lockhart [Tue, 4 May 1999 02:16:57 +0000 (02:16 +0000)]
Include chapter on security.
Tom Lane [Tue, 4 May 1999 00:00:20 +0000 (00:00 +0000)]
Make sure targetlist generated for subplan does not share
nodes with HAVING qualifier of upper plan. Have not seen any failures,
just being a little bit paranoid...
Tom Lane [Mon, 3 May 1999 23:48:26 +0000 (23:48 +0000)]
Correct declaration of array_map() so that it doesn't make
gcc quite so unhappy.
Bruce Momjian [Mon, 3 May 1999 19:10:48 +0000 (19:10 +0000)]
here are some patches for 6.5.0 which I already submitted but have never
been applied. The patches are in the .tar.gz attachment at the end:
varchar-array.patch this patch adds support for arrays of bpchar() and
varchar(), which where always missing from postgres.
These datatypes can be used to replace the _char4,
_char8, etc., which were dropped some time ago.
block-size.patch this patch fixes many errors in the parser and other
program which happen with very large query statements
(> 8K) when using a page size larger than 8192.
This patch is needed if you want to submit queries
larger than 8K. Postgres supports tuples up to 32K
but you can't insert them because you can't submit
queries larger than 8K. My patch fixes this problem.
The patch also replaces all the occurrences of `8192'
and `1<<13' in the sources with the proper constants
defined in include files. You should now never find
8192 hardwired in C code, just to make code clearer.
--
Massimo Dal Zotto
Tom Lane [Mon, 3 May 1999 00:38:44 +0000 (00:38 +0000)]
Revise union_planner and associated routines to clean up breakage
from EXCEPT/HAVING patch. Cases involving nontrivial GROUP BY expressions
now work again. Also, the code is at least somewhat better documented...
Tom Lane [Sat, 1 May 1999 19:47:42 +0000 (19:47 +0000)]
Clean up cost_sort some more: most callers were double-counting
the cost of reading the source data.
Tom Lane [Sat, 1 May 1999 19:09:46 +0000 (19:09 +0000)]
Arrange for VACUUM to delete the init file that relcache.c uses
to save a little bit of backend startup time. This way, the first
backend started after a VACUUM will rebuild the init file with up-to-date
statistics for the critical system indexes.
Tom Lane [Sat, 1 May 1999 17:16:25 +0000 (17:16 +0000)]
-T was omitted from getopt() call.
Tom Lane [Sat, 1 May 1999 17:14:56 +0000 (17:14 +0000)]
Fix compile failures in dt.c --- line broken in middle of
an identifier :-(. Sloppy transmission of a patch, likely.
Vadim B. Mikheev [Sat, 1 May 1999 16:09:45 +0000 (16:09 +0000)]
Patch from "Hiroshi Inoue" <Inoue@tpf.co.jp> for
FATAL 1:btree: BTP_CHAIN flag was expected
Vadim B. Mikheev [Sat, 1 May 1999 15:04:46 +0000 (15:04 +0000)]
Use page-level ExtendLock lock instead of table-level -
should be faster.
Bruce Momjian [Fri, 30 Apr 1999 17:03:04 +0000 (17:03 +0000)]
cleanup
Bruce Momjian [Fri, 30 Apr 1999 16:22:46 +0000 (16:22 +0000)]
Clean up typo causing compile failure.
Tom Lane [Fri, 30 Apr 1999 04:04:27 +0000 (04:04 +0000)]
Fill in reasonable-looking cost estimates in inserted nodes.
This makes no difference to the optimizer, which has already decided what
it's gonna do, but it makes the output of EXPLAIN much more plausible.
Tom Lane [Fri, 30 Apr 1999 04:01:44 +0000 (04:01 +0000)]
Clean up some bogosities in path cost estimation, like
sometimes estimating an index scan of a table to be cheaper than a
sequential scan of the same tuples...
Tom Lane [Fri, 30 Apr 1999 03:59:06 +0000 (03:59 +0000)]
Fix nasty little typo that prevented get_cheapest_path_for_joinkeys
from ever returning a path. This put a bit of a crimp in the system's
ability to generate intelligent merge-join plans...
Bruce Momjian [Fri, 30 Apr 1999 02:04:51 +0000 (02:04 +0000)]
Prevent priority inversion in locking by checking for existing locks
before going into queue behind person with higher piority.
Jan Wieck [Thu, 29 Apr 1999 15:52:01 +0000 (15:52 +0000)]
Bugfix (bug by me in 1.4) in backparsing INSERT ... SELECT
Jan
Tom Lane [Thu, 29 Apr 1999 03:01:50 +0000 (03:01 +0000)]
Defend against 'update oid'. Someday we might want to support
that, but it'd be a New Feature, wouldn't it ... in the meantime,
avoiding a backend crash seems worthwhile.
Tom Lane [Thu, 29 Apr 1999 01:13:13 +0000 (01:13 +0000)]
Aggregate functions didn't work on subscripted array references.
Things are better now.
Tom Lane [Thu, 29 Apr 1999 00:20:27 +0000 (00:20 +0000)]
Several routines in setrefs.c would crash on array refs
due to lack of check for recursing into a null subexpression.
Tom Lane [Wed, 28 Apr 1999 22:17:58 +0000 (22:17 +0000)]
Fix stupid typo that broke empty-query response... odd that
this is not revealed by any of our regression tests...
Jan Wieck [Tue, 27 Apr 1999 13:33:43 +0000 (13:33 +0000)]
Fixed DECIMAL data type to handle specified precision in atttypmod
Jan
Tatsuo Ishii [Tue, 27 Apr 1999 09:49:36 +0000 (09:49 +0000)]
Fix for _copyUnique() suggested by Hiroshi Inoue
Marc G. Fournier [Mon, 26 Apr 1999 13:37:14 +0000 (13:37 +0000)]
These are no longer used ...
Marc G. Fournier [Mon, 26 Apr 1999 13:33:19 +0000 (13:33 +0000)]
provides string to be used for expected/* files for system specific output...
Michael Meskes [Mon, 26 Apr 1999 05:28:48 +0000 (05:28 +0000)]
*** empty log message ***
Tatsuo Ishii [Mon, 26 Apr 1999 04:42:49 +0000 (04:42 +0000)]
Fix from Yutaka Tanida <yutaka@marin.or.jp> for Cygwin32 support.
Tatsuo Ishii [Mon, 26 Apr 1999 04:40:17 +0000 (04:40 +0000)]
Fix from Yutaka Tanida <yutaka@marin.or.jp>
Tom Lane [Mon, 26 Apr 1999 00:37:46 +0000 (00:37 +0000)]
Insert a test for missing targetlist entry in replace_agg_clause.
Tom Lane [Sun, 25 Apr 1999 23:30:31 +0000 (23:30 +0000)]
It looks like the multibyte regress tests only work if both
the database encoding and the client encoding match the encoding expected
by the test. So, force both of them to be set from the MULTIBYTE
environment var. This allows regress tests to be run successfully in
multibyte environments other than the compiled-in default.
Tom Lane [Sun, 25 Apr 1999 23:16:23 +0000 (23:16 +0000)]
More multibyte tests with obsolete spelling of error message...
Tom Lane [Sun, 25 Apr 1999 23:10:36 +0000 (23:10 +0000)]
psql's recognition of comments didn't work right in MULTIBYTE
environments; it was being careless about character lengths.
Tom Lane [Sun, 25 Apr 1999 21:56:19 +0000 (21:56 +0000)]
Explicitly set PGCLIENTENCODING during regression tests,
since multibyte tests fail if it's different from database...
Tom Lane [Sun, 25 Apr 1999 21:50:58 +0000 (21:50 +0000)]
Still had a few MULTIBYTE problems when client encoding was
different from database's ...
Tom Lane [Sun, 25 Apr 1999 20:35:51 +0000 (20:35 +0000)]
Correct potential infinite loop in pg_utf2wchar_with_len;
it failed to cover the case where high bits of char are 100 or 101.
Not sure if fix is right, but it agrees with pg_utf_mblen ... and it
doesn't lock up ...
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.
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...
Tom Lane [Sun, 25 Apr 1999 18:16:47 +0000 (18:16 +0000)]
Clean up gcc warning in MULTIBYTE code.
Tom Lane [Sun, 25 Apr 1999 18:09:54 +0000 (18:09 +0000)]
ifdef out some unused routines to suppress gcc warnings.
Tom Lane [Sun, 25 Apr 1999 03:27:15 +0000 (03:27 +0000)]
Ooops, missed committing this one...
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.
Bruce Momjian [Fri, 23 Apr 1999 21:23:49 +0000 (21:23 +0000)]
Rename explain's "size" to "rows".
Bruce Momjian [Fri, 23 Apr 1999 19:37:42 +0000 (19:37 +0000)]
Add disk space message to "can not extend" message.
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.
Tom Lane [Fri, 23 Apr 1999 00:50:57 +0000 (00:50 +0000)]
Repair missing heap_endscan() in OperatorUpd().
Vadim B. Mikheev [Thu, 22 Apr 1999 08:19:59 +0000 (08:19 +0000)]
Fix duplicating ROOT page in concurrent updates.
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.
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.
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 ...
Bruce Momjian [Mon, 19 Apr 1999 16:00:18 +0000 (16:00 +0000)]
Add temporary for temp.
Tatsuo Ishii [Mon, 19 Apr 1999 08:46:34 +0000 (08:46 +0000)]
Fix typo in multi-byte encodings (EHC_CN --> EUC_CN)
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;
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.
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
Tom Lane [Sun, 18 Apr 1999 03:01:49 +0000 (03:01 +0000)]
Update CREATE FUNCTION's on-line help in psql.
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.
Bruce Momjian [Sat, 17 Apr 1999 17:18:41 +0000 (17:18 +0000)]
Add res checking to libpq examples, from Dan Merillat.
Tom Lane [Fri, 16 Apr 1999 21:27:23 +0000 (21:27 +0000)]
Add missing function prototypes to stifle gcc warnings.
Michael Meskes [Fri, 16 Apr 1999 12:26:49 +0000 (12:26 +0000)]
*** empty log message ***
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.
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)
Tom Lane [Fri, 16 Apr 1999 04:59:03 +0000 (04:59 +0000)]
Improve error messages when a connection is rejected.
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.
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.
Peter Mount [Thu, 15 Apr 1999 06:00:16 +0000 (06:00 +0000)]
Replaced <literal></literal> with "" in ProgramListing sections
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.