OSDN Git Service

pg-rex/syncrep.git
25 years agoWhen closure of the backend connection is detected during pqFlush,
Tom Lane [Fri, 28 May 1999 01:54:53 +0000 (01:54 +0000)]
When closure of the backend connection is detected during pqFlush,
do the right thing: look for a NOTICE message from the backend before we
close our side of the socket.  6.4 libpq did not reliably print the backend's
hara-kiri message, 'The Postmaster has informed me ...', because it only
did the right thing if connection closure was detected during a read
attempt instead of a write attempt.

25 years agoMake pg_dump dump ACL's by default, print warning on use of -z, and add
Bruce Momjian [Thu, 27 May 1999 16:29:05 +0000 (16:29 +0000)]
Make pg_dump dump ACL's by default, print warning on use of -z, and add
new -x option to skip acl dump.

25 years agoMore info is in sgml and html docs so this is now obsolete.
Thomas G. Lockhart [Thu, 27 May 1999 16:06:33 +0000 (16:06 +0000)]
More info is in sgml and html docs so this is now obsolete.

25 years agoFAQs contributed for this release.
Thomas G. Lockhart [Thu, 27 May 1999 16:05:38 +0000 (16:05 +0000)]
FAQs contributed for this release.

25 years agoMarkup fixes.
Thomas G. Lockhart [Thu, 27 May 1999 15:49:15 +0000 (15:49 +0000)]
Markup fixes.

25 years agoMarkup fixes.
Thomas G. Lockhart [Thu, 27 May 1999 15:49:08 +0000 (15:49 +0000)]
Markup fixes.
Update for v6.5 release.

25 years agoSignificant updates from Vince Vielhaber.
Thomas G. Lockhart [Thu, 27 May 1999 15:48:14 +0000 (15:48 +0000)]
Significant updates from Vince Vielhaber.

25 years agoRemove disclaimer about exact numeric types. They work now!
Thomas G. Lockhart [Thu, 27 May 1999 15:47:28 +0000 (15:47 +0000)]
Remove disclaimer about exact numeric types. They work now!

25 years agoRefresh FreeBSD info.
Thomas G. Lockhart [Thu, 27 May 1999 15:46:22 +0000 (15:46 +0000)]
Refresh FreeBSD info.

25 years agoSignificant update from Vince Vielhaber.
Thomas G. Lockhart [Thu, 27 May 1999 15:44:54 +0000 (15:44 +0000)]
Significant update from Vince Vielhaber.

25 years agoUpdate release notes for 6.5.
Bruce Momjian [Thu, 27 May 1999 15:13:43 +0000 (15:13 +0000)]
Update release notes for 6.5.

25 years agoClean up release sgml file.
Bruce Momjian [Thu, 27 May 1999 15:05:57 +0000 (15:05 +0000)]
Clean up release sgml file.

25 years agoI am not sure if libpq++ will compile with non g++ compilers,
Bruce Momjian [Thu, 27 May 1999 14:28:40 +0000 (14:28 +0000)]
I am not sure if libpq++ will compile with non g++ compilers,
but the Makefile does break non g++.

 <<mak.patch>>
Andreas

25 years agoFix for crypt memory leak, from James Thompson
Bruce Momjian [Thu, 27 May 1999 04:09:45 +0000 (04:09 +0000)]
Fix for crypt memory leak, from James Thompson

25 years agoPatch from Andreas: when CREATE TABLE is followed by CREATE INDEX
Tom Lane [Wed, 26 May 1999 22:57:39 +0000 (22:57 +0000)]
Patch from Andreas: when CREATE TABLE is followed by CREATE INDEX
before any tuples are loaded, preserve the default '1000 tuples' table
size estimate.

25 years agoFix pg_dump to use the same maximum-query-size constant as
Tom Lane [Wed, 26 May 1999 21:51:13 +0000 (21:51 +0000)]
Fix pg_dump to use the same maximum-query-size constant as
the backend does.  Remove unnecessary limitation on field size in
dumpClasses_dumpData (ie, -d or -D case).

25 years agoFix compile of plpgsql by adding 'extern int yylineno.'
Bruce Momjian [Wed, 26 May 1999 20:55:06 +0000 (20:55 +0000)]
Fix compile of plpgsql by adding 'extern int yylineno.'

25 years agoDisplay numeric precision on \d.
Bruce Momjian [Wed, 26 May 1999 20:08:06 +0000 (20:08 +0000)]
Display numeric precision on \d.

25 years agoAllow GROUPs to be dumped properly.
Bruce Momjian [Wed, 26 May 1999 19:45:53 +0000 (19:45 +0000)]
Allow GROUPs to be dumped properly.

25 years agoAdd chapters on CVS access, MVCC, SQL theory to the docs.
Thomas G. Lockhart [Wed, 26 May 1999 17:30:30 +0000 (17:30 +0000)]
Add chapters on CVS access, MVCC, SQL theory to the docs.
Add an appendix with more details on date/time attributes and handling.
Update most references to Postgres version numbers to 6.5,
 *except* for the porting list which will require a report
 from a successful installation to be updated.

25 years agoChapter on multi-version concurrency control from Vadim.
Thomas G. Lockhart [Wed, 26 May 1999 17:27:39 +0000 (17:27 +0000)]
Chapter on multi-version concurrency control from Vadim.
Some wording changes from Vadim's original text doc.
Processes cleanly, but may need fixup.

25 years agoAdd new reference pages for postmaster and postgres.
Thomas G. Lockhart [Wed, 26 May 1999 17:26:32 +0000 (17:26 +0000)]
Add new reference pages for postmaster and postgres.
Normalize markup for vacuumdb; content is the same.

25 years agoFirst copy from the man pages.
Thomas G. Lockhart [Wed, 26 May 1999 17:25:38 +0000 (17:25 +0000)]
First copy from the man pages.
postgres-ref.sgml is not yet marked up.

25 years agoFix for NT from Horak Daniel
Bruce Momjian [Wed, 26 May 1999 16:19:48 +0000 (16:19 +0000)]
Fix for NT from Horak Daniel

25 years agoDisable use of -o and -d pg_dump options together. Can't set oids in
Bruce Momjian [Wed, 26 May 1999 16:06:45 +0000 (16:06 +0000)]
Disable use of -o and -d pg_dump options together.  Can't set oids in
inserts.  Change some variables to bool to be clearer.

25 years agoAdd fix for 0x7fU constants to pgindent
Bruce Momjian [Wed, 26 May 1999 15:20:04 +0000 (15:20 +0000)]
Add fix for 0x7fU constants to pgindent

25 years agoAllow pg_dump -v display proper table/sequence count display.
Bruce Momjian [Wed, 26 May 1999 14:50:38 +0000 (14:50 +0000)]
Allow pg_dump -v display proper table/sequence count display.

25 years agoMake functions static or NOT_USED as appropriate.
Bruce Momjian [Wed, 26 May 1999 12:57:23 +0000 (12:57 +0000)]
Make functions static or NOT_USED as appropriate.

25 years agoGive BEOS a chance ...
Marc G. Fournier [Wed, 26 May 1999 01:22:24 +0000 (01:22 +0000)]
Give BEOS a chance ...

25 years agoAnother pgindent run. Sorry folks.
Bruce Momjian [Tue, 25 May 1999 22:43:53 +0000 (22:43 +0000)]
Another pgindent run.  Sorry folks.

25 years agoMake 0x007f -> (unsigned)0x7f to make pgindent happy.
Bruce Momjian [Tue, 25 May 1999 22:04:56 +0000 (22:04 +0000)]
Make 0x007f -> (unsigned)0x7f to make pgindent happy.

25 years agoGet rid of page-level locking in btree-s.
Vadim B. Mikheev [Tue, 25 May 1999 18:31:28 +0000 (18:31 +0000)]
Get rid of page-level locking in btree-s.
BT_READ/BT_WRITE are BUFFER_LOCK_SHARE/BUFFER_LOCK_EXCLUSIVE now.
Also get rid of #define BT_VERSION_1 - we use version 1 as default
for near two years now.

25 years agoGet rid of page-level locking in btree-s.
Vadim B. Mikheev [Tue, 25 May 1999 18:20:31 +0000 (18:20 +0000)]
Get rid of page-level locking in btree-s.
LockBuffer is used to acquire read/write access
to index pages. Pages are released before leaving
index internals.

25 years agopgindent run over code.
Bruce Momjian [Tue, 25 May 1999 16:15:34 +0000 (16:15 +0000)]
pgindent run over code.

25 years ago*** empty log message ***
Michael Meskes [Tue, 25 May 1999 13:36:15 +0000 (13:36 +0000)]
*** empty log message ***

25 years agoBugfix - Range table entries that are unused after rewriting should
Jan Wieck [Tue, 25 May 1999 13:16:10 +0000 (13:16 +0000)]
Bugfix - Range table entries that are unused after rewriting should
not be marked inFromCl any longer. Otherwise the planner gets confused
and joins over them where in fact it does not have to.

Adjust hasSubLinks now with a recursive lookup - could be wrong in
multi action rules because parse state isn't reset correctly and all
actions in the rule are marked hasSubLinks if one of them has.

Jan

25 years agoFixed bug in rules event qualification output.
Jan Wieck [Tue, 25 May 1999 08:49:33 +0000 (08:49 +0000)]
Fixed bug in rules event qualification output.

Jan

25 years agoFIx for 0.0.0.0/0 output as 00/0.
Bruce Momjian [Tue, 25 May 1999 05:29:38 +0000 (05:29 +0000)]
FIx for 0.0.0.0/0 output as 00/0.

25 years agoDo not assign output columns to junk attributes created from
Tom Lane [Sun, 23 May 1999 21:42:09 +0000 (21:42 +0000)]
Do not assign output columns to junk attributes created from
GROUP BY or ORDER BY expressions in INSERT ... SELECT.

25 years agoDetect case of invalid use of GROUP BY when there are no
Tom Lane [Sun, 23 May 1999 21:41:14 +0000 (21:41 +0000)]
Detect case of invalid use of GROUP BY when there are no
aggregate functions, as in
select a, b from foo group by a;
The ungrouped reference to b is not kosher, but formerly we neglected to
check this unless there was an aggregate function somewhere in the query.

25 years agoRemove more -B parameters not needed.
Bruce Momjian [Sun, 23 May 1999 19:01:18 +0000 (19:01 +0000)]
Remove more -B parameters not needed.

25 years agoRemove -B from wisconsin test.
Bruce Momjian [Sun, 23 May 1999 18:53:30 +0000 (18:53 +0000)]
Remove -B from wisconsin test.

25 years agoFix tuple chain moving bug found by "Hiroshi Inoue" <Inoue@tpf.co.jp>.
Vadim B. Mikheev [Sun, 23 May 1999 09:10:24 +0000 (09:10 +0000)]
Fix tuple chain moving bug found by "Hiroshi Inoue" <Inoue@tpf.co.jp>.

25 years agoHere it is. Remove or rename the current interfaces/libpq++ and untar
Bruce Momjian [Sun, 23 May 1999 01:04:07 +0000 (01:04 +0000)]
Here it is.  Remove or rename the current interfaces/libpq++ and untar
this file in interfaces/

It will all need to be checked in.  I used the char *rcsid[] method for
cvs ids so it can be strings | grep'd to find version numbers.  The new
version for the library is 3.0.

Run configure from src/ to create the Makefile and it should be good to
go.

I did minimal documentation references in the README, I'll see if I can
get something to Tom Lockhart rather quickly.

Vince.

25 years agoUpdate commentary in sample GEQO config file.
Tom Lane [Sat, 22 May 1999 23:59:59 +0000 (23:59 +0000)]
Update commentary in sample GEQO config file.

25 years agoReduce default GEQO 'effort' setting to MEDIUM always.
Tom Lane [Sat, 22 May 1999 23:27:19 +0000 (23:27 +0000)]
Reduce default GEQO 'effort' setting to MEDIUM always.
This agrees with the documentation and seems like a more useful default
anyhow ...

25 years agoModify aset.c logic so that blocks requested from malloc get
Tom Lane [Sat, 22 May 1999 23:19:37 +0000 (23:19 +0000)]
Modify aset.c logic so that blocks requested from malloc get
bigger the more space is used in an allocset.  This reduces the malloc
overhead very substantially on queries that need lots of memory.

25 years agoImprove error message from failed LOAD command (include
Tom Lane [Sat, 22 May 1999 19:49:42 +0000 (19:49 +0000)]
Improve error message from failed LOAD command (include
kernel's error description when file is not accessible).

25 years agoAllow GEQO effort to be specified numerically, as well as
Tom Lane [Sat, 22 May 1999 19:29:01 +0000 (19:29 +0000)]
Allow GEQO effort to be specified numerically, as well as
symbolic LOW/MEDIUM/HIGH values --- needed for experiments with other
effort levels ...

25 years agoModify backend switch parsing to prevent 'insecure' switches
Tom Lane [Sat, 22 May 1999 17:47:54 +0000 (17:47 +0000)]
Modify backend switch parsing to prevent 'insecure' switches
from being accepted when they are passed from client connection request.
Get rid of a couple that no longer do anything (like -P).

25 years agoFix for select 1;select 2 without trailing semi.
Bruce Momjian [Sat, 22 May 1999 05:06:43 +0000 (05:06 +0000)]
Fix for select 1;select 2 without trailing semi.

25 years agoFix for DEFAULT ''.
Bruce Momjian [Sat, 22 May 1999 04:12:29 +0000 (04:12 +0000)]
Fix for DEFAULT ''.

25 years agoMake postgres prompt backend>, and remove PARSEDEBUG.
Bruce Momjian [Sat, 22 May 1999 02:55:58 +0000 (02:55 +0000)]
Make postgres prompt backend>, and remove PARSEDEBUG.

25 years agoFinish initial markup of cvs.sgml, and include it in the programmer's guide
Thomas G. Lockhart [Sat, 22 May 1999 02:27:25 +0000 (02:27 +0000)]
Finish initial markup of cvs.sgml, and include it in the programmer's guide
 and the integrated doc. Clean up other markup.

25 years agoChange perl Makefile test.
Bruce Momjian [Fri, 21 May 1999 19:03:48 +0000 (19:03 +0000)]
Change perl Makefile test.

25 years agoDisable fix. Didn't work.
Bruce Momjian [Fri, 21 May 1999 18:33:12 +0000 (18:33 +0000)]
Disable fix.  Didn't work.

25 years agoFix typo and attempt default fix.
Bruce Momjian [Fri, 21 May 1999 18:31:06 +0000 (18:31 +0000)]
Fix typo and attempt default fix.

25 years ago*** empty log message ***
Michael Meskes [Fri, 21 May 1999 16:36:27 +0000 (16:36 +0000)]
*** empty log message ***

25 years agoTreat {} as special regex too.
Bruce Momjian [Fri, 21 May 1999 15:47:13 +0000 (15:47 +0000)]
Treat {} as special regex too.

25 years agoadd retest, a regex testing program
Tatsuo Ishii [Fri, 21 May 1999 06:27:54 +0000 (06:27 +0000)]
add retest, a regex testing program

25 years agoFix problem with | in ~ comparison using index.
Bruce Momjian [Fri, 21 May 1999 04:40:04 +0000 (04:40 +0000)]
Fix problem with | in ~ comparison using index.

25 years agoAdd material about postmaster startup failure messages;
Tom Lane [Fri, 21 May 1999 02:42:59 +0000 (02:42 +0000)]
Add material about postmaster startup failure messages;
update material about client connection failure messages.

25 years agoReport strerror() rather than errno in low-level backend libpq
Tom Lane [Fri, 21 May 1999 01:25:06 +0000 (01:25 +0000)]
Report strerror() rather than errno in low-level backend libpq
failure messages.

25 years agoAdded a long section about proper use of the optimizer-hint
Tom Lane [Fri, 21 May 1999 00:38:33 +0000 (00:38 +0000)]
Added a long section about proper use of the optimizer-hint
clauses in CREATE OPERATOR.  Needs markup work.

25 years agoMinor updates to libpq documentation.
Tom Lane [Fri, 21 May 1999 00:36:46 +0000 (00:36 +0000)]
Minor updates to libpq documentation.

25 years agoRemoved the automatic installation of built procedural languages
Jan Wieck [Thu, 20 May 1999 16:50:08 +0000 (16:50 +0000)]
Removed the automatic installation of built procedural languages
from initdb again.

Added two new commands, createlang and destroylang to bin. These
hopefully end this damned mklang.sql discussion.

Jan

25 years agoGenerate distinct error messages for trigger function not found
Tom Lane [Thu, 20 May 1999 14:39:49 +0000 (14:39 +0000)]
Generate distinct error messages for trigger function not found
and trigger function found but returns wrong type.

25 years agoFixed shift/reduce conflict
Jan Wieck [Thu, 20 May 1999 12:12:55 +0000 (12:12 +0000)]
Fixed shift/reduce conflict

SelectStmt and CursorStmt tried to parse FOR UPDATE ... / FOR READ ONLY.
Cursor now checks that it is read only by looking at forUpdate of Query.
SelectStmt handles FOR READ ONLY too.

Jan

25 years agooverwriting a large object now works
Tatsuo Ishii [Thu, 20 May 1999 09:30:36 +0000 (09:30 +0000)]
overwriting a large object now works

25 years agoAdd reference pages on user interface applications.
Thomas G. Lockhart [Thu, 20 May 1999 05:40:27 +0000 (05:40 +0000)]
Add reference pages on user interface applications.

25 years agoRearrange and consolidate the Admin Guide.
Thomas G. Lockhart [Thu, 20 May 1999 05:39:29 +0000 (05:39 +0000)]
Rearrange and consolidate the Admin Guide.
Add reference pages for utilities and remove standalone chapters for same.
Add material for an appendix on date/time properties, but not yet
 integrated with the User's Guide.
Break up the former chapter on pg_options
 into Admin and Programmer's Guides.

25 years agoVery minor improvements in CREATE OPERATOR docs.
Tom Lane [Thu, 20 May 1999 03:22:01 +0000 (03:22 +0000)]
Very minor improvements in CREATE OPERATOR docs.

25 years agoUpdate CREATE FUNCTION docs: mention use of AS clause with
Tom Lane [Thu, 20 May 1999 02:57:15 +0000 (02:57 +0000)]
Update CREATE FUNCTION docs: mention use of AS clause with
INTERNAL functions, add a warning about trying to overload function names
for dynamically loaded C functions (from old man page).

25 years agoUpdate SGML docs with some changes I made to test/regress/README
Tom Lane [Thu, 20 May 1999 02:46:40 +0000 (02:46 +0000)]
Update SGML docs with some changes I made to test/regress/README
many moons ago, but forgot to put here.

25 years agoDoco updates for change to handling of INTERNAL function
Tom Lane [Thu, 20 May 1999 02:44:53 +0000 (02:44 +0000)]
Doco updates for change to handling of INTERNAL function
entries (prosrc is now name of C-level function).

25 years agoMinor improvements to postmaster/backend man pages.
Tom Lane [Wed, 19 May 1999 23:30:43 +0000 (23:30 +0000)]
Minor improvements to postmaster/backend man pages.

25 years agoDec ALpha patches
Bruce Momjian [Wed, 19 May 1999 18:06:43 +0000 (18:06 +0000)]
Dec ALpha patches

25 years agoDec alpha patches.
Bruce Momjian [Wed, 19 May 1999 18:04:51 +0000 (18:04 +0000)]
Dec alpha patches.

25 years agoRemove 4096 string limited key on block size
Bruce Momjian [Wed, 19 May 1999 17:53:12 +0000 (17:53 +0000)]
Remove 4096 string limited key on block size

25 years agoUpgrade to PyGreSQL (2.4)
Bruce Momjian [Wed, 19 May 1999 16:46:12 +0000 (16:46 +0000)]
Upgrade to PyGreSQL (2.4)

25 years agoBring python up to date ...
Marc G. Fournier [Wed, 19 May 1999 14:46:54 +0000 (14:46 +0000)]
Bring python up to date ...

From: D'Arcy J.M. Cain <darcy@druid.net>

25 years agoAdd Aggref and ArrayRef to the set of node types that transformExpr
Tom Lane [Tue, 18 May 1999 23:40:05 +0000 (23:40 +0000)]
Add Aggref and ArrayRef to the set of node types that transformExpr
will pass through rather than spitting up.  This is necessary to handle
cases where coerce_type causes a subexpression to be retransformed, as in
SELECT count(*) + 1.0 FROM table

25 years agoForgot the CHANGELOG
Peter Mount [Tue, 18 May 1999 23:22:08 +0000 (23:22 +0000)]
Forgot the CHANGELOG

25 years agoInternationalisation of error messages
Peter Mount [Tue, 18 May 1999 23:17:46 +0000 (23:17 +0000)]
Internationalisation of error messages

25 years agoNow that hashjoin is reliable for large joins (knock on wood),
Tom Lane [Tue, 18 May 1999 21:36:10 +0000 (21:36 +0000)]
Now that hashjoin is reliable for large joins (knock on wood),
remove optimizer's arbitrary limit on how large a join it will use hashing
for.  (The limit was too large to prevent the problems we'd been seeing,
anyway...)

25 years agoRemove no-longer-used fields in Hash and HashJoin nodes.
Tom Lane [Tue, 18 May 1999 21:34:29 +0000 (21:34 +0000)]
Remove no-longer-used fields in Hash and HashJoin nodes.

25 years agoRewrite hash join to use simple linked lists instead of a
Tom Lane [Tue, 18 May 1999 21:33:06 +0000 (21:33 +0000)]
Rewrite hash join to use simple linked lists instead of a
fixed-size hashtable.  This should prevent 'hashtable out of memory' errors,
unless you really do run out of memory.  Note: target size for hashtable
is now taken from -S postmaster switch, not -B, since it is local memory
in the backend rather than shared memory.

25 years agoTransactions in ImageViewer
Peter Mount [Tue, 18 May 1999 06:07:25 +0000 (06:07 +0000)]
Transactions in ImageViewer

25 years ago*** empty log message ***
Michael Meskes [Tue, 18 May 1999 05:20:17 +0000 (05:20 +0000)]
*** empty log message ***

25 years agoFixed Internationalization of error messages.
Peter Mount [Mon, 17 May 1999 22:58:19 +0000 (22:58 +0000)]
Fixed Internationalization of error messages.

25 years agoMinor bug fixes. Replaced DateStyle support with ISO.
Peter Mount [Mon, 17 May 1999 22:43:30 +0000 (22:43 +0000)]
Minor bug fixes. Replaced DateStyle support with ISO.

25 years agoAll works on linux now by my tests and regression(with patch below).
Bruce Momjian [Mon, 17 May 1999 18:24:48 +0000 (18:24 +0000)]
All works on linux now by my tests and regression(with patch below).

ALTER TABLE RENAME with extents.
Ole Gjerde

25 years agoSkip junk nodes when comparing UNION target list lengths.
Bruce Momjian [Mon, 17 May 1999 18:22:19 +0000 (18:22 +0000)]
Skip junk nodes when comparing UNION target list lengths.

25 years agoChange resjunk to a boolean.
Bruce Momjian [Mon, 17 May 1999 17:03:51 +0000 (17:03 +0000)]
Change resjunk to a boolean.

25 years agoFixed latest regression diff's by omitting viewowner in SELECT from pg_views.
Jan Wieck [Mon, 17 May 1999 09:03:26 +0000 (09:03 +0000)]
Fixed latest regression diff's by omitting viewowner in SELECT from pg_views.

Jan

25 years agoThis is actually more of a fundamental problem with mdtruncate. It
Bruce Momjian [Mon, 17 May 1999 06:38:41 +0000 (06:38 +0000)]
This is actually more of a fundamental problem with mdtruncate.  It
looks
like someone just didn't add support for multiple segments for
truncation.

The following patch seems to do the right thing, for me at least.
It passed my tests, my data looks right(no data that shouldn't be in
there) and regression is ok.

Ole Gjerde

25 years agoAdd python
Bruce Momjian [Mon, 17 May 1999 06:27:07 +0000 (06:27 +0000)]
Add python

25 years agoRe-add python.
Bruce Momjian [Mon, 17 May 1999 06:15:31 +0000 (06:15 +0000)]
Re-add python.

25 years agoUpdates for 6.5.
Bruce Momjian [Mon, 17 May 1999 06:06:35 +0000 (06:06 +0000)]
Updates for 6.5.

25 years agoUpdate pointers to where the register/mailing list pages are found on the
Marc G. Fournier [Mon, 17 May 1999 04:58:19 +0000 (04:58 +0000)]
Update pointers to where the register/mailing list pages are found on the
wweb site ..