OSDN Git Service
Bruce Momjian [Sat, 3 Jul 1999 01:56:16 +0000 (01:56 +0000)]
Update tuple size check.
Bruce Momjian [Sat, 3 Jul 1999 01:47:02 +0000 (01:47 +0000)]
Fix for insertion of tuple too large.
Bruce Momjian [Sat, 3 Jul 1999 00:33:04 +0000 (00:33 +0000)]
Fix to prevent too large tuple from being created.
Bruce Momjian [Fri, 2 Jul 1999 18:09:28 +0000 (18:09 +0000)]
Fix for removal of temp tables if last transaction was aborted.
Bruce Momjian [Fri, 2 Jul 1999 03:21:37 +0000 (03:21 +0000)]
typo fix.
Bruce Momjian [Thu, 1 Jul 1999 19:47:25 +0000 (19:47 +0000)]
#ifdef around true and false.
Bruce Momjian [Thu, 1 Jul 1999 05:31:22 +0000 (05:31 +0000)]
Add files to point to TODO list.
Tom Lane [Thu, 1 Jul 1999 00:06:28 +0000 (00:06 +0000)]
make distclean missed a few symlinks created by configure.
Tom Lane [Wed, 30 Jun 1999 23:57:31 +0000 (23:57 +0000)]
Make sure symlinks for sharedlib get removed by 'make clean'.
Tom Lane [Wed, 30 Jun 1999 23:54:18 +0000 (23:54 +0000)]
Generate a full set of version-numbered symlinks when building
a shared library, not just when installing it.
Michael Meskes [Tue, 29 Jun 1999 09:25:25 +0000 (09:25 +0000)]
*** empty log message ***
Vadim B. Mikheev [Tue, 29 Jun 1999 04:54:49 +0000 (04:54 +0000)]
Avoid disk writes for read-only transactions.
Bruce Momjian [Mon, 28 Jun 1999 18:40:34 +0000 (18:40 +0000)]
Add time warning about bigtests.
Peter Mount [Sun, 27 Jun 1999 11:05:38 +0000 (11:05 +0000)]
Some minor bug fixes
Bruce Momjian [Sat, 26 Jun 1999 15:58:28 +0000 (15:58 +0000)]
Add var defines for no testandset
Bruce Momjian [Sat, 26 Jun 1999 15:54:36 +0000 (15:54 +0000)]
Remove -m486 on Openbsd.
Michael Meskes [Fri, 25 Jun 1999 10:42:06 +0000 (10:42 +0000)]
*** empty log message ***
Tom Lane [Fri, 25 Jun 1999 00:13:45 +0000 (00:13 +0000)]
Makefile.global failed to define $(LIBS), which is
unfortunate considering that several subdirectory makefiles were counting
on it to do so...
Michael Meskes [Thu, 24 Jun 1999 13:15:31 +0000 (13:15 +0000)]
*** empty log message ***
Thomas G. Lockhart [Wed, 23 Jun 1999 06:21:19 +0000 (06:21 +0000)]
Remove User's Guide entities since they were not being used.
Thomas G. Lockhart [Wed, 23 Jun 1999 06:19:36 +0000 (06:19 +0000)]
Add information on generating INSTALL and HISTORY from sgml sources.
Thomas G. Lockhart [Wed, 23 Jun 1999 06:17:51 +0000 (06:17 +0000)]
Clarify input/output timezone information.
Add detail on exact interpretation of "concatenated date"
for various lengths.
Thomas G. Lockhart [Wed, 23 Jun 1999 06:15:37 +0000 (06:15 +0000)]
Add backup/restore info to Admin Guide.
Split management chapter from start-ag.sgml to manage-ag.sgml.
Thomas G. Lockhart [Wed, 23 Jun 1999 06:12:31 +0000 (06:12 +0000)]
Minor changes to Fred Horsh's updates.
Thomas G. Lockhart [Wed, 23 Jun 1999 06:09:57 +0000 (06:09 +0000)]
Update maximum identifier length from 15 to 31 characters.
Peter Mount [Wed, 23 Jun 1999 05:56:20 +0000 (05:56 +0000)]
Some fixes supplied by Jon Nielsen jonfn@image.dk
Bruce Momjian [Mon, 21 Jun 1999 15:23:58 +0000 (15:23 +0000)]
The references to mailing lists in the resources page seem outdated.
Patch attached.
Fred Horch
Bruce Momjian [Mon, 21 Jun 1999 15:12:19 +0000 (15:12 +0000)]
Some suggestions to clarify the current status of PostgreSQL. Patch
attached.
Fred Horch
Tom Lane [Mon, 21 Jun 1999 01:26:56 +0000 (01:26 +0000)]
Replace rewriter's checkQueryHasAggs and checkQueryHasSubLink
with expression_tree_walker-based code. The former failed to cope with
expressions containing SubLinks, and the latter returned TRUE for both
SubLinks and Aggrefs (cut-and-paste bug?). There is a lot more scope for
using expression_tree_walker in this module, but I'll restrain myself
until the 6.6 split occurs from touching not-demonstrably-broken code.
Tom Lane [Mon, 21 Jun 1999 01:20:57 +0000 (01:20 +0000)]
Clean up problems with sublinks + grouping in planner. Not
sure if they are all fixed, because rewriter is now the stumbling block,
but at least some cases work that did not work before.
Tom Lane [Mon, 21 Jun 1999 01:18:02 +0000 (01:18 +0000)]
On second thought, expression_tree_walker should handle bare
SubLink nodes after all ...
Bruce Momjian [Sat, 19 Jun 1999 05:05:52 +0000 (05:05 +0000)]
I have a small patch for 6.5.
aclchk.c: heap_close() is not called after calling heap_openr().
Atsushi Ogawa
Bruce Momjian [Sat, 19 Jun 1999 05:00:30 +0000 (05:00 +0000)]
Rename to vararg_format().
Bruce Momjian [Sat, 19 Jun 1999 04:54:23 +0000 (04:54 +0000)]
Change form() to varargform() to prevent portability problems.
Tom Lane [Sat, 19 Jun 1999 03:48:31 +0000 (03:48 +0000)]
My first chosen victim for expression_tree_walker conversion
is parse_aggs.c. This fixes its failure to cope with (at least) CaseExpr
and ArrayRef nodes, which is the reason why both of these fail in 6.5:
select coalesce(f1,0) from int4_tbl group by f1;
ERROR: Illegal use of aggregates or non-group column in target list
select sentence.words[0] from sentence group by sentence.words[0];
ERROR: Illegal use of aggregates or non-group column in target list
The array case still fails, but at least it's not parse_agg's fault
anymore ... considering that we now support CASE officially, I think
it's important to fix the first example ...
Tom Lane [Sat, 19 Jun 1999 03:41:45 +0000 (03:41 +0000)]
Create a generic expression-tree-walker subroutine, which
will gradually replace all of the boilerplate tree-walk-recursion code that
currently exists in O(N) slightly different forms in N subroutines.
I've had it with adding missing cases to these subroutines...
Tom Lane [Sat, 19 Jun 1999 00:44:44 +0000 (00:44 +0000)]
Temporarily disable error checks for missing selectivity
functions, in order to work around oversight in 6.5 release: rtree
index functions haven't got any. Mea culpa ...
Bruce Momjian [Fri, 18 Jun 1999 16:47:23 +0000 (16:47 +0000)]
Patch to allow vacuum on multi-segment tables, from Hiroshi Inoue
Thomas G. Lockhart [Fri, 18 Jun 1999 05:03:48 +0000 (05:03 +0000)]
Incorporate fixes from Date and Darwen, 1997.
Did not check the function declarations as carefully as the other parts,
though all of the function names *do* match up with D&D.
Bruce Momjian [Fri, 18 Jun 1999 00:46:11 +0000 (00:46 +0000)]
linux_ppc now use -O0 to fix problem in optimization of fmgr calls.
Tom Lane [Thu, 17 Jun 1999 23:45:32 +0000 (23:45 +0000)]
Explain didn't handle inheritance correctly (it didn't
manipulate rtable the same way executor does).
Tom Lane [Thu, 17 Jun 1999 22:21:41 +0000 (22:21 +0000)]
Defend against function calls with more than 8 arguments (code
used to overrun its fixed-size arrays before detecting error; not cool).
Also, replace uses of magic constant '8' with 'MAXFARGS'.
Bruce Momjian [Thu, 17 Jun 1999 15:16:09 +0000 (15:16 +0000)]
Remove QUERY_LIMIT and documenation on same. Change _ALIGN to TYPEALIGN
for Irix.
Thomas G. Lockhart [Thu, 17 Jun 1999 14:19:17 +0000 (14:19 +0000)]
Add CLI required header and examples from SQL3/SQL98
August 1994 draft standard.
Use the ecpg support libraries to write the CLI interface?
Date and Darwen claim that CLI is a more modern and flexible approach...
Bruce Momjian [Thu, 17 Jun 1999 13:44:43 +0000 (13:44 +0000)]
Make linux_ppc use only -O, not -O2.
Michael Meskes [Wed, 16 Jun 1999 18:35:51 +0000 (18:35 +0000)]
*** empty log message ***
Michael Meskes [Wed, 16 Jun 1999 18:25:50 +0000 (18:25 +0000)]
*** empty log message ***
Bruce Momjian [Wed, 16 Jun 1999 18:09:59 +0000 (18:09 +0000)]
Add NT readme file.
Bruce Momjian [Wed, 16 Jun 1999 11:01:17 +0000 (11:01 +0000)]
Cleanup
Tom Lane [Mon, 14 Jun 1999 17:49:06 +0000 (17:49 +0000)]
Undo Jan's typo that broke regress.sh's detection of system
type name.
Thomas G. Lockhart [Mon, 14 Jun 1999 16:33:46 +0000 (16:33 +0000)]
Fix cross-reference to large objects.
Thomas G. Lockhart [Mon, 14 Jun 1999 16:32:41 +0000 (16:32 +0000)]
Fix cross-reference to large objects in the CREATE TYPE section.
Thomas G. Lockhart [Mon, 14 Jun 1999 15:57:24 +0000 (15:57 +0000)]
Release notes from sgml.
Came out with max 79 columns from Applix by defining the page width at
10 inches and the right margin at 9.5 inches, then exporting
as "ASCII Layout".
Thomas G. Lockhart [Mon, 14 Jun 1999 07:41:04 +0000 (07:41 +0000)]
Put in CREATE TABLE AS and SELECT INTO reference pages.
Left them out after splitting them
from the CREATE TABLE and SELECT sources.
Thomas G. Lockhart [Mon, 14 Jun 1999 07:37:05 +0000 (07:37 +0000)]
Minor updates for release.
Split reference pages for CREATE TABLE AS and SELECT INTO to allow psgml
(the emacs parser) to handle parsing.
Thomas G. Lockhart [Mon, 14 Jun 1999 07:36:12 +0000 (07:36 +0000)]
Minor updates for release.
Thomas G. Lockhart [Mon, 14 Jun 1999 07:24:32 +0000 (07:24 +0000)]
Cut down to 80 columns per Bruce. Mostly hacked at the porting table.
Thomas G. Lockhart [Mon, 14 Jun 1999 06:59:34 +0000 (06:59 +0000)]
HTML versions of docs for v6.5 release.
Thomas G. Lockhart [Mon, 14 Jun 1999 06:56:31 +0000 (06:56 +0000)]
Fix URL for Norm Walsh's Style Sheets.
Thomas G. Lockhart [Mon, 14 Jun 1999 06:55:45 +0000 (06:55 +0000)]
Fix typos and wording in v6.5 release summary per Fred Horsh.
Thomas G. Lockhart [Mon, 14 Jun 1999 06:25:03 +0000 (06:25 +0000)]
User's Guide for v6.5.
Tatsuo Ishii [Sun, 13 Jun 1999 00:07:43 +0000 (00:07 +0000)]
change #if defined(__mc68000__) to:
#if defined(__mc68000__) && defined(__linux__)
so that other m68k systems(such as NetBSD) will not be affected.
Tom Lane [Sat, 12 Jun 1999 22:17:24 +0000 (22:17 +0000)]
Move default NBuffers setting into config.h, and rename it
to DEF_NBUFFERS for readability. Make sure the default value is OK
according to postmaster.c's new sanity check for -B values.
Tom Lane [Sat, 12 Jun 1999 20:41:25 +0000 (20:41 +0000)]
Fix critical error noticed by Massimo: copy.c used to have a
special hack to ensure it would close its output file even after failure
due to elog(ERROR) partway through the copy. This is now unnecessary
because fd.c takes care of cleaning up open files at transaction abort;
worse, after fd.c closed the file copy.c would try to do so *again* at
the start of the next COPY command. This would result in havoc in most
implementations of stdio library.
Tom Lane [Sat, 12 Jun 1999 19:38:30 +0000 (19:38 +0000)]
Remove query_planner's overhasty rejection of cases where
tlist and qual are NULL. It ought to handle these the same as the cases
where tlist contains only constant expressions, ie, be willing to generate
a Result-node plan. This did not use to matter, but it does now because
union_planner will flatten the tlist when aggregates are present. Thus,
'select count(1) from table' now causes query_planner to be given a null
tlist, and to duplicate 6.4's behavior we need it to give back a Result
plan rather than refusing the query. 6.4 was arguably doing the Wrong
Thing for this query, but I'm not going to open a semantics issue right
before 6.5 release ... can revisit that problem later.
Tom Lane [Sat, 12 Jun 1999 19:27:41 +0000 (19:27 +0000)]
Plug hole in dike: planner would coredump if query_planner
returned NULL, which it will do in some cases where an elog(ERROR) would
probably be more appropriate. For the moment, generate a not-very-
informative error message rather than proceeding to certain coredump.
Probably ought to think about making query_planner elog instead of
returning NULL, but this is at least a safe change for now.
Tom Lane [Sat, 12 Jun 1999 19:22:40 +0000 (19:22 +0000)]
When targetlist is NULL, ExecTargetList was passing back a
pointer to palloc'd but uninitialized memory. This is not cool; anyone looking
at the returned 'tuple' would at best coredump and at worst behave in a
bizarre and irreproducible way. Fix it to return a predictable value,
namely a correctly-set-up palloc'd tuple containing zero attributes.
I believe this fix is both safe and critical.
Bruce Momjian [Sat, 12 Jun 1999 14:07:33 +0000 (14:07 +0000)]
Reversed out Massimo patch.
Bruce Momjian [Sat, 12 Jun 1999 14:05:41 +0000 (14:05 +0000)]
I don't like last minute patches before the final freeze, but I believe that
this one could be useful for people experiencing out-of-memory crashes while
executing queries which retrieve or use a very large number of tuples.
The problem happens when storage is allocated for functions results used in
a large query, for example:
select upper(name) from big_table;
select big_table.array[1] from big_table;
select count(upper(name)) from big_table;
This patch is a dirty hack that fixes the out-of-memory problem for the most
common cases, like the above ones. It is not the final solution for the
problem but it can work for some people, so I'm posting it.
The patch should be safe because all changes are under #ifdef. Furthermore
the feature can be enabled or disabled at runtime by the `free_tuple_memory'
options in the pg_options file. The option is disabled by default and must
be explicitly enabled at runtime to have any effect.
To enable the patch add the follwing line to Makefile.custom:
CUSTOM_COPT += -DFREE_TUPLE_MEMORY
To enable the option at runtime add the following line to pg_option:
free_tuple_memory=1
Massimo
Bruce Momjian [Fri, 11 Jun 1999 15:01:29 +0000 (15:01 +0000)]
Hi,
please apply the included patch. It corrects the headers in src/win32 -
there are some missing #endif.
Dan
Bruce Momjian [Fri, 11 Jun 1999 14:34:02 +0000 (14:34 +0000)]
add m86 similar entries.
Vadim B. Mikheev [Fri, 11 Jun 1999 09:35:08 +0000 (09:35 +0000)]
Change Assert(Ptp.t_data->t_xmax == tp.t_data->t_xmin) to :
/*
* Read above about cases when !ItemIdIsUsed(Citemid)
* (child item is removed)... Due to the fact that
* at the moment we don't remove unuseful part of
* update-chain, it's possible to get too old
* parent row here. Like as in the case which
* caused this problem, we stop shrinking here.
* I could try to find real parent row but want
* not to do it because of real solution will
* be implemented anyway, latter, and we are too
* close to 6.5 release. - vadim 06/11/99
*/
if (Ptp.t_data->t_xmax != tp.t_data->t_xmin)
...
Vadim B. Mikheev [Fri, 11 Jun 1999 09:00:02 +0000 (09:00 +0000)]
Removed bad Assert(!buf->ri_lock) when unlocking exclusively
locked buffer.
Vadim B. Mikheev [Fri, 11 Jun 1999 05:40:18 +0000 (05:40 +0000)]
More about chained mode and isolation.
Vadim B. Mikheev [Fri, 11 Jun 1999 05:19:53 +0000 (05:19 +0000)]
More about chained mode and isolation levels.
Bruce Momjian [Fri, 11 Jun 1999 02:39:43 +0000 (02:39 +0000)]
Change mdtruncate to truncate and not unlink.
Hiroshi Inoue
Bruce Momjian [Thu, 10 Jun 1999 22:59:22 +0000 (22:59 +0000)]
This patch should enable 6.5 to build on Motorola 68000 architecture.
It comes from Roman Hodek <Roman.Hodek@informatik.uni-erlangen.de>.
Michael Meskes [Thu, 10 Jun 1999 19:11:33 +0000 (19:11 +0000)]
*** empty log message ***
Jan Wieck [Thu, 10 Jun 1999 17:49:32 +0000 (17:49 +0000)]
Added 2 regression tests for NUMERIC data type.
1. Using 100 digits after decimal point on the default
make runtest.
2. Using 1000 digits after decimal point in a new target
make bigtest.
At the end of 'make runtest', a hint about the new bigtest is
printed.
Jan
Vadim B. Mikheev [Thu, 10 Jun 1999 14:17:12 +0000 (14:17 +0000)]
1. Fix for elog(ERROR, "EvalPlanQual: t_xmin is uncommitted ?!")
and possibly for other cases too:
DO NOT cache status of transaction in unknown state
(i.e. non-committed and non-aborted ones)
Example:
T1 reads row updated/inserted by running T2 and cache T2 status.
T2 commits.
Now T1 reads a row updated by T2 and with HEAP_XMAX_COMMITTED
in t_infomask (so cached T2 status is not changed).
Now T1 EvalPlanQual gets updated row version without HEAP_XMIN_COMMITTED
-> TransactionIdDidCommit(t_xmin) and TransactionIdDidAbort(t_xmin)
return FALSE and T2 decides that t_xmin is not committed and gets
ERROR above.
It's too late to find more smart way to handle such cases and so
I just changed xact status caching and got rid TransactionIdFlushCache()
from code.
Changed: transam.c, xact.c, lmgr.c and transam.h - last three
just because of TransactionIdFlushCache() is removed.
2. heapam.c:
T1 marked a row for update. T2 waits for T1 commit/abort.
T1 commits. T3 updates the row before T2 locks row page.
Now T2 sees that new row t_xmax is different from xact id (T1)
T2 was waiting for. Old code did Assert here. New one goes to
HeapTupleSatisfiesUpdate. Obvious changes too.
3. Added Assert to vacuum.c
4. bufmgr.c: break
Assert(buf->r_locks == 0 && !buf->ri_lock)
into two Asserts.
Tom Lane [Thu, 10 Jun 1999 06:55:40 +0000 (06:55 +0000)]
Fix errors in SELECT ... GROUP BY ... UNION SELECT ...
ye proverbial one-line patch (not counting five lines of comment so's
maybe it won't happen again)
Thomas G. Lockhart [Wed, 9 Jun 1999 13:44:59 +0000 (13:44 +0000)]
Add mention of hot backups to the release notes.
Add id tags to chapter and section headings to ensure good labels
on output html files.
Thomas G. Lockhart [Wed, 9 Jun 1999 13:43:42 +0000 (13:43 +0000)]
Make minor changes in wording.
Adjust tags to get a clean build.
Bruce Momjian [Wed, 9 Jun 1999 12:39:46 +0000 (12:39 +0000)]
Remove -O2 for linux/alpha, make -O.
Vadim B. Mikheev [Wed, 9 Jun 1999 12:23:42 +0000 (12:23 +0000)]
Reset evaluation plan tuple table next free slot counter to 0
after ExecEndNode. It must be done! Or we'll be out of free
tuple slots very soon, though slots are freed by ExecEndNode
and ready for reusing.
We didn't see this problem before because of
int nSlots = ExecCountSlotsNode(plan);
TupleTable tupleTable = ExecCreateTupleTable(nSlots + 10);
/* why add ten? - jolly */
code in InitPlan - i.e. extra 10 slots. Simple select uses
3 slots and so it was possible to re-use evaluation plan
3 additional times and didn't get
elog(NOTICE, "Plan requires more slots than are available");
elog(ERROR, "send mail to your local executor guru to fix this");
Changes are obvious and shouldn't be problems with them.
Though, I added Assert(epqstate->es_tupleTable->next == 0)
before EvalPlanQual():ExecInitNode and we'll notice if
something is still wrong. Is it better to change Assert
to elog(ERROR) ?
Vadim B. Mikheev [Wed, 9 Jun 1999 03:51:40 +0000 (03:51 +0000)]
MVCC updates.
Vadim B. Mikheev [Tue, 8 Jun 1999 20:01:01 +0000 (20:01 +0000)]
MVCC updation.
Bruce Momjian [Tue, 8 Jun 1999 19:19:51 +0000 (19:19 +0000)]
Linux-alpha gets -O
Bruce Momjian [Tue, 8 Jun 1999 17:34:43 +0000 (17:34 +0000)]
Remove optimization on linux/alpha.
Bruce Momjian [Tue, 8 Jun 1999 16:42:55 +0000 (16:42 +0000)]
Fix for duplicate LC_COLLATE
Bruce Momjian [Tue, 8 Jun 1999 16:42:10 +0000 (16:42 +0000)]
Add mention of new pg_dump feature.
Bruce Momjian [Tue, 8 Jun 1999 16:27:26 +0000 (16:27 +0000)]
Install proper Irix FAQ.
Bruce Momjian [Tue, 8 Jun 1999 01:37:55 +0000 (01:37 +0000)]
Cleanup
Bruce Momjian [Tue, 8 Jun 1999 01:32:05 +0000 (01:32 +0000)]
Disable pg_upgrade for 6.5.
Bruce Momjian [Mon, 7 Jun 1999 22:58:31 +0000 (22:58 +0000)]
I attach a patch for pg_upgrade. This does two things:
1. check whether the program is being executed in $PGDATA/.. This is
necessary if the data tree is not in the standard place, as is the
case with the Debian distribution (because of Debian policy).
2. give a clearer error message if the dumped data structure fails to
be loaded.
Oliver Elphick
Bruce Momjian [Mon, 7 Jun 1999 17:20:24 +0000 (17:20 +0000)]
On AIX 4.3.2 the third line in template/aix_42:
SHARED_LIB:
needs to be changed to:
SHARED_LIB:-lc
I think this was also needed on AIX 4.2. Comments Please !!
If nobody objects, I suggest to make this change, since it cannot
break AIX 4.2 and is necessary on AIX 4.3
Andreas
Vadim B. Mikheev [Mon, 7 Jun 1999 15:14:54 +0000 (15:14 +0000)]
Concurrency... Highest one...
DO NOT EVEN TRY TO DO PageGetMaxOffsetNumber BEFORE
LockBuffer!
-:)
Bruce Momjian [Mon, 7 Jun 1999 14:29:20 +0000 (14:29 +0000)]
> Here is a small patch that should only affect win32 building
> (native win32, not cygnus).
> It does the following:
> Patches two win32.mak files to DEFINE HAVE_VSNPRINTF and
> HAVE_STRDUP. This is required to build at all.
> Bumps the version number on libpq.dll from 6.4 to 6.5.
> Required for install programs to work.
> Adds defintions for BLCKSZ and MAXIMUM_ALIGN to "win32.h" in
> the client-side libpiq directory.
>
> All these files are only used when building on native win32,
> so it should be safe I think.
>
> Again, really sorry to throw this in so late, but I would
> hate to do the same thing as with 6.4 (which required 6.4.1
> to at all compile on Win32).
>
> Thanks,
>
> //Magnus
Tom Lane [Mon, 7 Jun 1999 14:28:26 +0000 (14:28 +0000)]
Repair recently-introduced error in makeIndexable for LIKE:
a non-leading % would be put into the >=/<= patterns. Also, repair
longstanding confusion about whether %% means a literal %%. The SQL92
doesn't say any such thing, and textlike() knows that, but gram.y didn't.
Vadim B. Mikheev [Mon, 7 Jun 1999 14:28:22 +0000 (14:28 +0000)]
Have to release meta page before reading root one!
< 6.5 versions were just not affected by this bug due to locking.