]> granicus.if.org Git - postgresql/log
postgresql
22 years agoFix likely cause of rare ALTER TABLE ADD FOREIGN KEY failures ---
Tom Lane [Thu, 12 Sep 2002 21:16:42 +0000 (21:16 +0000)]
Fix likely cause of rare ALTER TABLE ADD FOREIGN KEY failures ---
don't assume relname field of a relcache entry will stay valid across
lots of operations.

22 years agoAvoid misleading error message when SET/RESET target variable name
Tom Lane [Thu, 12 Sep 2002 14:03:45 +0000 (14:03 +0000)]
Avoid misleading error message when SET/RESET target variable name
doesn't match any known variable.

22 years agoThe small context diff below corrects what seems to be an oversight in
Bruce Momjian [Thu, 12 Sep 2002 00:26:42 +0000 (00:26 +0000)]
The small context diff below corrects what seems to be an oversight in
fmgr.h - it's discouraged to access fcinfo directly but there is no
macro to get the number of arguments passed to the function. Checking
the number of arguments is often useful when you have a function which
can be called like:

 func('arg');
 func(null);
 func();

all mapping to the same C function.

the macro has a function-like appearance to match the other PG_*
macros.

Lee Kindness.

22 years agoThis is a comprehensive set of diffs (versus current CVS) that replaces those
Bruce Momjian [Thu, 12 Sep 2002 00:26:00 +0000 (00:26 +0000)]
This is a comprehensive set of diffs (versus current CVS) that replaces those
attached to the same message with the Earth Distance patches.
Recent changes include changing the subscript in one place I forgot
in the previous bugfix patch. A couple of added regression tests, which
should help catch this mistake if it reappears.
I also put in a limit of 100 dimensions in cube_large and cube_in to
prevent making it easy to create very large cubes. Changing one define
in cubedata.h will raise the limit if some needs more dimensions.

Bruno Wolff III

22 years ago> intarray and ltree both seem to be mapping their own declarations onto
Bruce Momjian [Thu, 12 Sep 2002 00:24:46 +0000 (00:24 +0000)]
> intarray and ltree both seem to be mapping their own declarations onto
> arrays using largely-similar code.  But while intarray fails its
> regression test, I find ltree still passes.  So I'm confused about what
> that code is really doing and don't want to touch it.

Please, apply attached patch, it solves the problem.

Teodor Sigaev

22 years ago> Sean Chittenden <sean@chittenden.org> writes:
Bruce Momjian [Thu, 12 Sep 2002 00:24:10 +0000 (00:24 +0000)]
> Sean Chittenden <sean@chittenden.org> writes:
>
>>::sigh:: Is it me or does it look like all
>>of pl/pgsql is schema un-aware (ie, all of the declarations).  -sc
>
>
> Yeah.  The group of routines parse_word, parse_dblword, etc that are
> called by the lexer certainly all need work.  There are some
> definitional issues to think about, too --- plpgsql presently relies on
> the number of names to give it some idea of what to look for, and those
> rules are probably all toast now.  Please come up with a sketch of what
> you think the behavior should be before you start hacking code.

Attached is a diff -c format proposal to fix this. I've also attached a short
test script. Seems to work OK and passes all regression tests.

Here's a breakdown of how I understand plpgsql's "Special word rules" -- I
think it illustrates the behavior reasonably well. New functions added by this
patch are plpgsql_parse_tripwordtype and plpgsql_parse_dblwordrowtype:

Joe Conway

22 years agoJoe Conway wrote:
Bruce Momjian [Thu, 12 Sep 2002 00:21:25 +0000 (00:21 +0000)]
Joe Conway wrote:
 > Hannu Krosing wrote:
 >
 >> It seems that my last mail on this did not get through to the list
 >> ;(
 >>
 >> Please consider renaming the new builtin function
 >> split(text,text,int)
 >>
 >> to something else, perhaps
 >>
 >> split_part(text,text,int)
 >>
 >> (like date_part)
 >>
 >> The reason for this request is that 3 most popular scripting
 >> languages (perl, python, php) all have also a function with similar
 >> signature, but returning an array instead of single element and the
 >> (optional) third argument is limit (maximum number of splits to
 >> perform)
 >>
 >> I think that it would be good to have similar function in (some
 >> future release of) postgres, but if we now let in a function with
 >> same name and arguments but returning a single string instead an
 >> array of them, then we will need to invent a new and not so easy to
 >> recognise name for the "real" split function.
 >>
 >
 > This is a good point, and I'm not opposed to changing the name, but
 > it is too bad your original email didn't get through before beta1 was
 >  rolled. The change would now require an initdb, which I know we were
 >  trying to avoid once beta started (although we could change it
 > without *requiring* an initdb I suppose).
 >
 > I guess if we do end up needing an initdb for other reasons, we
 > should make this change too. Any other opinions? Is split_part an
 > acceptable name?
 >
 > Also, if we add a todo to produce a "real" split function that
 > returns an array, similar to those languages, I'll take it for 7.4.

No one commented on the choice of name, so the attached patch changes
the name of split(text,text,int) to split_part(text,text,int) per
Hannu's recommendation above. This can be applied without an initdb if
current beta testers are advised to run:

   update pg_proc set proname = 'split_part' where proname = 'split';

in the case they want to use this function. Regression and doc fix is
also included in the patch.

Joe Conway

22 years agoNo change.
Bruce Momjian [Thu, 12 Sep 2002 00:20:04 +0000 (00:20 +0000)]
No change.

22 years ago> Now I'm testing connectby() in the /contrib/tablefunc in 7.3b1, which would
Bruce Momjian [Thu, 12 Sep 2002 00:19:44 +0000 (00:19 +0000)]
> Now I'm testing connectby()  in the /contrib/tablefunc in 7.3b1, which would
> be a useful function for many users.   However, I found the fact that
> if connectby_tree has the following data, connectby() tries to search the end
> of roots without knowing that the relations are infinite(-5-9-10-11-9-10-11-)
.
> I hope connectby() supports a check routine to find infinite relations.
>
>
> CREATE TABLE connectby_tree(keyid int, parent_keyid int);
> INSERT INTO connectby_tree VALUES(1,NULL);
> INSERT INTO connectby_tree VALUES(2,1);
> INSERT INTO connectby_tree VALUES(3,1);
> INSERT INTO connectby_tree VALUES(4,2);
> INSERT INTO connectby_tree VALUES(5,2);
> INSERT INTO connectby_tree VALUES(6,4);
> INSERT INTO connectby_tree VALUES(7,3);
> INSERT INTO connectby_tree VALUES(8,6);
> INSERT INTO connectby_tree VALUES(9,5);
>
> INSERT INTO connectby_tree VALUES(10,9);
> INSERT INTO connectby_tree VALUES(11,10);
> INSERT INTO connectby_tree VALUES(9,11);    <-- infinite
>

The attached patch fixes the infinite recursion bug in
contrib/tablefunc/tablefunc.c:connectby found by Masaru Sugawara.

test=# SELECT * FROM connectby('connectby_tree', 'keyid',
'parent_keyid', '2', 4, '~') AS t(keyid int, parent_keyid int, level
int, branch text);
  keyid | parent_keyid | level |   branch
-------+--------------+-------+-------------
      2 |              |     0 | 2
      4 |            2 |     1 | 2~4
      6 |            4 |     2 | 2~4~6
      8 |            6 |     3 | 2~4~6~8
      5 |            2 |     1 | 2~5
      9 |            5 |     2 | 2~5~9
     10 |            9 |     3 | 2~5~9~10
     11 |           10 |     4 | 2~5~9~10~11
(8 rows)

test=# SELECT * FROM connectby('connectby_tree', 'keyid',
'parent_keyid', '2', 5, '~') AS t(keyid int, parent_keyid int, level
int, branch text);
ERROR:  infinite recursion detected

I implemented it by checking the branch string for repeated keys
(whether or not the branch is returned). The performance hit was pretty
minimal -- about 1% for a moderately complex test case (220000 record
table, 9 level tree with 3800 members).

Joe Conway

22 years ago> BTW, clusterdb is not schema-aware and will surely fail in any database
Bruce Momjian [Thu, 12 Sep 2002 00:18:14 +0000 (00:18 +0000)]
> BTW, clusterdb is not schema-aware and will surely fail in any database
> where more than one schema is in use, because it doesn't trouble to
> schema-qualify table names.

Ok, the following patch should solve this concern.  It also tries to
connect as little times as possible (the previous one would connect one
time per table plus one per database; this one connects two times per
database).

Alvaro Herrera

22 years agoThe attached small patch fixes the cause of the regression test failure
Bruce Momjian [Thu, 12 Sep 2002 00:15:33 +0000 (00:15 +0000)]
The attached small patch fixes the cause of the regression test failure
for contrib/intarray.

The cause was that the library uses its own function to construct a new
array, new_intArrayType, and that function did not set the new array
struct attribute elemtype.

Joe Conway

22 years agoThe attached removes the current non-standard file
Bruce Momjian [Thu, 12 Sep 2002 00:14:40 +0000 (00:14 +0000)]
The attached removes the current non-standard file
"contrib/tablefunc/tablefunc-test.sql", and adds a standard regression
test suite to contrib/tablefunc.

Joe Conway

22 years agoAdd sprompt.obj to Win32 makefiles.
Bruce Momjian [Wed, 11 Sep 2002 17:36:13 +0000 (17:36 +0000)]
Add sprompt.obj to Win32 makefiles.

22 years agoAdd comment about sharing of sprompt.c file.
Bruce Momjian [Wed, 11 Sep 2002 17:32:37 +0000 (17:32 +0000)]
Add comment about sharing of sprompt.c file.

22 years agoTweak querytree-dependency-extraction code so that columns of tables
Tom Lane [Wed, 11 Sep 2002 14:48:55 +0000 (14:48 +0000)]
Tweak querytree-dependency-extraction code so that columns of tables
that are explicitly JOINed are not considered dependencies unless they
are actually used in the query: mere presence in the joinaliasvars
list of a JOIN RTE doesn't count as being used.  The patch touches
a number of files because I needed to generalize the API of
query_tree_walker to support an additional flag bit, but the changes
are otherwise quite small.

22 years agoPatches submitted by Kris Jurka (jurka@ejurka.com) for the following bugs:
Barry Lind [Wed, 11 Sep 2002 05:38:45 +0000 (05:38 +0000)]
Patches submitted by Kris Jurka (jurka@ejurka.com) for the following bugs:
  - Properly drop tables in jdbc regression tests with cascade for 7.3
  - problem with Statement.execute() and executeUpdate() not clearing binds
  - problem with ResultSet not correctly handling default encoding
  - changes to correctly support show transaction isolation level in 7.3
  - changed DatabaseMetaDataTest to handle differences in FK names in 7.3
  - better fix for dynamically checking server NAME data length
  (With the fixes above the jdbc regression tests pass on jdbc2 and jdbc3
   against both a 7.2 and 7.3 server)
Patchs submitted by David Wall (d.wall@computer.org):
  - problem with getBlob when largeobject oid is null
  - improvements to BlobOutputStream
Patch submitted by Haris Peco (snpe@snpe.co.yu):
  - problem with callable statement not supporting prepared statement methods

 Modified Files:
  jdbc/org/postgresql/Driver.java.in
  jdbc/org/postgresql/jdbc1/AbstractJdbc1Connection.java
  jdbc/org/postgresql/jdbc1/AbstractJdbc1DatabaseMetaData.java
  jdbc/org/postgresql/jdbc1/AbstractJdbc1Statement.java
  jdbc/org/postgresql/jdbc2/AbstractJdbc2ResultSet.java
  jdbc/org/postgresql/jdbc2/AbstractJdbc2Statement.java
  jdbc/org/postgresql/jdbc2/Jdbc2ResultSet.java
  jdbc/org/postgresql/jdbc3/Jdbc3ResultSet.java
  jdbc/org/postgresql/largeobject/BlobOutputStream.java
  jdbc/org/postgresql/largeobject/LargeObject.java
  jdbc/org/postgresql/test/TestUtil.java
  jdbc/org/postgresql/test/jdbc2/DatabaseMetaDataTest.java
  jdbc/org/postgresql/test/jdbc2/UpdateableResultTest.java
  jdbc/org/postgresql/test/jdbc2/optional/BaseDataSourceTest.java
  jdbc/org/postgresql/test/jdbc2/optional/ConnectionPoolTest.java
  jdbc/org/postgresql/test/jdbc2/optional/SimpleDataSourceTest.java

22 years agoAdd mention of config.log for configure failure debugging.
Bruce Momjian [Wed, 11 Sep 2002 04:27:48 +0000 (04:27 +0000)]
Add mention of config.log for configure failure debugging.

22 years agoFix atan() description.
Bruce Momjian [Wed, 11 Sep 2002 02:56:46 +0000 (02:56 +0000)]
Fix atan() description.

Bruno Wolff III

22 years agoFix portability problem (size_t != int).
Tom Lane [Tue, 10 Sep 2002 18:25:13 +0000 (18:25 +0000)]
Fix portability problem (size_t != int).

22 years agoPowerup defaults for LC_foo GUC variables should match what main.c does.
Tom Lane [Tue, 10 Sep 2002 16:09:02 +0000 (16:09 +0000)]
Powerup defaults for LC_foo GUC variables should match what main.c does.

22 years agoUse different sed separator for configure arguments.
Peter Eisentraut [Mon, 9 Sep 2002 18:35:04 +0000 (18:35 +0000)]
Use different sed separator for configure arguments.

22 years agoRemove more references to pgaccess as a build target in docs.
Bruce Momjian [Sun, 8 Sep 2002 02:33:08 +0000 (02:33 +0000)]
Remove more references to pgaccess as a build target in docs.

22 years agoFixed DatabaseMetaData to correctly handle NAME size of 64
Barry Lind [Sun, 8 Sep 2002 00:15:29 +0000 (00:15 +0000)]
Fixed DatabaseMetaData to correctly handle NAME size of 64
Fixed Statement to correctly DEALLOCATE any prepared statements

 Modified Files:
  jdbc/org/postgresql/PGStatement.java
  jdbc/org/postgresql/jdbc1/AbstractJdbc1DatabaseMetaData.java
  jdbc/org/postgresql/jdbc1/AbstractJdbc1Statement.java

22 years agoChanges to documentation and the regression tests for the default
Bruce Momjian [Sat, 7 Sep 2002 18:39:05 +0000 (18:39 +0000)]
Changes to documentation and the regression tests for the default
NAMEDATALEN of 64.

Kris Jurka

22 years agoDone as far as possible:
Bruce Momjian [Sat, 7 Sep 2002 18:37:15 +0000 (18:37 +0000)]
Done as far as possible:

< * Add documentation to lock shared memory into RAM for each OS, if possible

22 years agoRemove pgaccess from docs; not shipping it anymore.
Bruce Momjian [Sat, 7 Sep 2002 16:49:50 +0000 (16:49 +0000)]
Remove pgaccess from docs; not shipping it anymore.

22 years agoAllow pg_dumpall to work with previous releases again. Don't pass the -c
Peter Eisentraut [Sat, 7 Sep 2002 16:14:33 +0000 (16:14 +0000)]
Allow pg_dumpall to work with previous releases again.  Don't pass the -c
option down to pg_dump, where it's useless, and clarify the meaning of -c
in the documentation.

22 years agoFix help output.
Peter Eisentraut [Sat, 7 Sep 2002 16:12:27 +0000 (16:12 +0000)]
Fix help output.

22 years agoMake sure the pg_dump tar archiver can handle members larger than 2 GB, but
Peter Eisentraut [Fri, 6 Sep 2002 21:58:36 +0000 (21:58 +0000)]
Make sure the pg_dump tar archiver can handle members larger than 2 GB, but
does not create members larger than allowed by the tar format.  Also, fix
the generation of the tar header to conform to POSIX.

22 years agoAdjust the tarball splitting scheme to the new reality.
Peter Eisentraut [Fri, 6 Sep 2002 21:57:11 +0000 (21:57 +0000)]
Adjust the tarball splitting scheme to the new reality.

22 years agoRun pgjindent for Java folks.
Bruce Momjian [Fri, 6 Sep 2002 21:23:06 +0000 (21:23 +0000)]
Run pgjindent for Java folks.

22 years agoFix another typo.
Bruce Momjian [Fri, 6 Sep 2002 20:26:00 +0000 (20:26 +0000)]
Fix another typo.

22 years agoFix typo.
Bruce Momjian [Fri, 6 Sep 2002 20:08:06 +0000 (20:08 +0000)]
Fix typo.

22 years agoUpdate for new IP.
Bruce Momjian [Fri, 6 Sep 2002 18:46:23 +0000 (18:46 +0000)]
Update for new IP.

22 years agoAdd JAVA_HOME test for Ant.
Bruce Momjian [Fri, 6 Sep 2002 14:31:16 +0000 (14:31 +0000)]
Add JAVA_HOME test for Ant.

22 years agoFix printf() quote handling and improper exit(), per Tom.
Bruce Momjian [Fri, 6 Sep 2002 02:33:47 +0000 (02:33 +0000)]
Fix printf() quote handling and improper exit(), per Tom.

22 years agoRemove heap_mark4update from AlterTableCreateToastTable. This has
Tom Lane [Fri, 6 Sep 2002 00:01:53 +0000 (00:01 +0000)]
Remove heap_mark4update from AlterTableCreateToastTable.  This has
never been the correct procedure for locking a relation, and the
recently-found ALTER TABLE bug with adding a constraint and a toast
table in the same command shows why it's a bad idea.

22 years agoFix some operator-precedence problems. New constructs IS DISTINCT FRM
Tom Lane [Thu, 5 Sep 2002 22:52:48 +0000 (22:52 +0000)]
Fix some operator-precedence problems.  New constructs IS DISTINCT FRM
and IS [NOT] OF were not being parsed consistently with other IS forms.
Also, make the world a little safer for functions named LEFT, RIGHT, etc.

22 years agoFix compile error.
Bruce Momjian [Thu, 5 Sep 2002 22:24:23 +0000 (22:24 +0000)]
Fix compile error.

22 years agoMissed mention of PGPASSWORDFILE.
Bruce Momjian [Thu, 5 Sep 2002 22:09:42 +0000 (22:09 +0000)]
Missed mention of PGPASSWORDFILE.

22 years agoUpdate HISTORY for PGPASSWORDFILE change.
Bruce Momjian [Thu, 5 Sep 2002 22:08:55 +0000 (22:08 +0000)]
Update HISTORY for PGPASSWORDFILE change.

22 years agoRemove PGPASSWORDFILE and default to always trying $HOME/.pgpass.
Bruce Momjian [Thu, 5 Sep 2002 22:05:50 +0000 (22:05 +0000)]
Remove PGPASSWORDFILE and default to always trying $HOME/.pgpass.

Cleanup up memory allocation for $HOME in related psql places.

Update mention of $HOME/.pgpass in the docs;  add mention in pg_dumpall.

22 years agoUpdate:
Bruce Momjian [Thu, 5 Sep 2002 22:03:02 +0000 (22:03 +0000)]
Update:

> * -Add ~/.pgpass to store passwords with user/host/password combinations

22 years agoFill in section on table modification.
Peter Eisentraut [Thu, 5 Sep 2002 21:32:23 +0000 (21:32 +0000)]
Fill in section on table modification.

22 years agoFix compile warning.
Tom Lane [Thu, 5 Sep 2002 21:19:13 +0000 (21:19 +0000)]
Fix compile warning.

22 years agoFix bit-rotted reference to GetUserName() ...
Tom Lane [Thu, 5 Sep 2002 21:13:03 +0000 (21:13 +0000)]
Fix bit-rotted reference to GetUserName() ...
it's GetUserNameFromId() now.

22 years agoFix breakage introduced by careless snprintf patching.
Tom Lane [Thu, 5 Sep 2002 21:09:54 +0000 (21:09 +0000)]
Fix breakage introduced by careless snprintf patching.

22 years agoFix unsafe macro definitions (which were producing incorrect code,
Tom Lane [Thu, 5 Sep 2002 21:08:26 +0000 (21:08 +0000)]
Fix unsafe macro definitions (which were producing incorrect code,
leading to compile warnings).

22 years agoRemove compile warnings, ensure consistent build environment for
Tom Lane [Thu, 5 Sep 2002 21:01:16 +0000 (21:01 +0000)]
Remove compile warnings, ensure consistent build environment for
largefile usage.

22 years agoFix compile warning.
Tom Lane [Thu, 5 Sep 2002 20:57:00 +0000 (20:57 +0000)]
Fix compile warning.

22 years agofindoidjoins and tsearch are not broken anymore.
Tom Lane [Thu, 5 Sep 2002 20:53:45 +0000 (20:53 +0000)]
findoidjoins and tsearch are not broken anymore.

22 years agoFix compile errors.
Tom Lane [Thu, 5 Sep 2002 20:51:39 +0000 (20:51 +0000)]
Fix compile errors.

22 years agoImprove opr_sanity regression test to check oprltcmpop and opgtcmpop
Tom Lane [Thu, 5 Sep 2002 20:23:19 +0000 (20:23 +0000)]
Improve opr_sanity regression test to check oprltcmpop and opgtcmpop
mergejoin links.

22 years agoUpdate oidjoins regression test for 7.3 catalogs.
Tom Lane [Thu, 5 Sep 2002 19:58:14 +0000 (19:58 +0000)]
Update oidjoins regression test for 7.3 catalogs.

22 years agofindoidjoins is updated for schemas, does not use libpgeasy.
Tom Lane [Thu, 5 Sep 2002 19:57:32 +0000 (19:57 +0000)]
findoidjoins is updated for schemas, does not use libpgeasy.
From Joe Conway.

22 years agoSeems like a good idea for template1 to contain ANALYZE stats for the
Tom Lane [Thu, 5 Sep 2002 19:56:57 +0000 (19:56 +0000)]
Seems like a good idea for template1 to contain ANALYZE stats for the
system tables.

22 years agoCommenting out doesn't work, so move the broken modules out of the list.
Peter Eisentraut [Thu, 5 Sep 2002 18:40:33 +0000 (18:40 +0000)]
Commenting out doesn't work, so move the broken modules out of the list.

22 years agoautoconf
Peter Eisentraut [Thu, 5 Sep 2002 18:39:11 +0000 (18:39 +0000)]
autoconf

22 years agoAssorted fixes for Cygwin:
Peter Eisentraut [Thu, 5 Sep 2002 18:28:46 +0000 (18:28 +0000)]
Assorted fixes for Cygwin:

Eliminate the mysterious games that the Cygwin build plays with the linker
flag variables.  DLLLIBS is gone, use SHLIB_LINK like everyone else.
Detect cygipc in configure, after the linker flags are set up, otherwise
configure might not work at all.

Make sure everything is covered by make clean.

Fix the build of the new conversion procedure modules.

Add new DLLIMPORT markers where required.

Finally, the compiler complains if we use an explicit
-I/usr/local/include, so don't do that.  Curiously, -L/usr/local/lib is
still necessary.

22 years agoDon't use gethostbyname2(). It's not portable and we don't claim to
Peter Eisentraut [Thu, 5 Sep 2002 18:27:13 +0000 (18:27 +0000)]
Don't use gethostbyname2().  It's not portable and we don't claim to
support IPv6 anyway.

22 years agoFix compilation warning. (Cygwin has char* as second argument of
Peter Eisentraut [Thu, 5 Sep 2002 18:26:18 +0000 (18:26 +0000)]
Fix compilation warning.  (Cygwin has char* as second argument of
recvfrom(), not void*.)

22 years agoRemove:
Bruce Momjian [Thu, 5 Sep 2002 16:40:18 +0000 (16:40 +0000)]
Remove:

< * Make pg_trigger.tgargs refer to columns by number, not name

22 years agoUpdate:
Bruce Momjian [Thu, 5 Sep 2002 05:11:11 +0000 (05:11 +0000)]
Update:

> * Make pg_trigger.tgargs refer to columns by number, not name

22 years agoDone:
Bruce Momjian [Thu, 5 Sep 2002 04:58:28 +0000 (04:58 +0000)]
Done:

> * -Make triggers refer to columns by number, not name

22 years agoremove pgaccess from GNUmakefile.in if we want the beta to build :)
Marc G. Fournier [Thu, 5 Sep 2002 02:05:30 +0000 (02:05 +0000)]
remove pgaccess from GNUmakefile.in if we want the beta to build :)

22 years agoRemove include of libpq-int.h --- dblink.c should not be (and was not)
Tom Lane [Thu, 5 Sep 2002 00:56:35 +0000 (00:56 +0000)]
Remove include of libpq-int.h --- dblink.c should not be (and was not)
depending on libpq internals.

22 years agoBe careful to include postgres.h *before* any system headers, to ensure
Tom Lane [Thu, 5 Sep 2002 00:43:07 +0000 (00:43 +0000)]
Be careful to include postgres.h *before* any system headers, to ensure
that the right flavors of largefile-related definitions are seen.
Most of these changes are probably unnecessary, but better safe than
sorry.

22 years agoGuard against send-lots-and-lots-of-data DoS attack from unauthenticated
Tom Lane [Wed, 4 Sep 2002 23:31:35 +0000 (23:31 +0000)]
Guard against send-lots-and-lots-of-data DoS attack from unauthenticated
users, by limiting the length of string we will accept for a password.
Patch by Serguei Mokhov, some editorializing by Tom Lane.

22 years agoAdd change by Neil.
Bruce Momjian [Wed, 4 Sep 2002 23:14:22 +0000 (23:14 +0000)]
Add change by Neil.

22 years agoRemove leftovers from subproject removals. Fixes for Python and Kerberos
Peter Eisentraut [Wed, 4 Sep 2002 22:54:18 +0000 (22:54 +0000)]
Remove leftovers from subproject removals.  Fixes for Python and Kerberos
configuration.

22 years agoFix compile warning.
Peter Eisentraut [Wed, 4 Sep 2002 22:51:23 +0000 (22:51 +0000)]
Fix compile warning.

22 years agoDisable findoidjoins while it doesn't compile.
Peter Eisentraut [Wed, 4 Sep 2002 22:50:43 +0000 (22:50 +0000)]
Disable findoidjoins while it doesn't compile.

22 years agoFix includes for plperl: ensure postgres.h is included first,
Tom Lane [Wed, 4 Sep 2002 22:49:37 +0000 (22:49 +0000)]
Fix includes for plperl: ensure postgres.h is included first,
remove unnecessary inclusions.

22 years agoAdd:
Bruce Momjian [Wed, 4 Sep 2002 22:09:47 +0000 (22:09 +0000)]
Add:

> * Remove Cyrillic recode support

22 years agoDone:
Bruce Momjian [Wed, 4 Sep 2002 21:07:41 +0000 (21:07 +0000)]
Done:

>       o -Add SHOW command to see locale

22 years agoAdd:
Bruce Momjian [Wed, 4 Sep 2002 21:05:40 +0000 (21:05 +0000)]
Add:

>  o -Add SHOW command to see locale

22 years agopgindent run.
Bruce Momjian [Wed, 4 Sep 2002 20:31:48 +0000 (20:31 +0000)]
pgindent run.

22 years agoUpdate:
Bruce Momjian [Wed, 4 Sep 2002 20:26:08 +0000 (20:26 +0000)]
Update:

Improvements to /contrib/intarray (Oleg, Teodor Sigaev, Andrey
Oktyabrski)

22 years agoAdd Joe's table functions description.
Bruce Momjian [Wed, 4 Sep 2002 20:06:23 +0000 (20:06 +0000)]
Add Joe's table functions description.

22 years agoUpdate symbols for 7.3.
Bruce Momjian [Wed, 4 Sep 2002 19:11:06 +0000 (19:11 +0000)]
Update symbols for 7.3.

22 years agoUpdate wording for Tom.
Bruce Momjian [Wed, 4 Sep 2002 19:07:43 +0000 (19:07 +0000)]
Update wording for Tom.

22 years agoUpdate for 7.3 typedefs.
Bruce Momjian [Wed, 4 Sep 2002 19:00:01 +0000 (19:00 +0000)]
Update for 7.3 typedefs.

22 years agoUpdate to reflect Tom's suggestions.
Bruce Momjian [Wed, 4 Sep 2002 18:45:52 +0000 (18:45 +0000)]
Update to reflect Tom's suggestions.

22 years agoRemove sh -x option that snuck in somehow.
Peter Eisentraut [Wed, 4 Sep 2002 18:04:57 +0000 (18:04 +0000)]
Remove sh -x option that snuck in somehow.

22 years agoReorder items.
Bruce Momjian [Wed, 4 Sep 2002 17:27:13 +0000 (17:27 +0000)]
Reorder items.

22 years agoMore HISTORY improvements from Joe.
Bruce Momjian [Wed, 4 Sep 2002 17:25:08 +0000 (17:25 +0000)]
More HISTORY improvements from Joe.

22 years agoUpdate HISTORY wording.
Bruce Momjian [Wed, 4 Sep 2002 17:11:47 +0000 (17:11 +0000)]
Update HISTORY wording.

22 years agoAvoid multiple scans of utils/mb/conversion_procs/ subdirectories during
Tom Lane [Wed, 4 Sep 2002 15:45:50 +0000 (15:45 +0000)]
Avoid multiple scans of utils/mb/conversion_procs/ subdirectories during
'make install'; there are enough of 'em that this slowed down the make
noticeably.  Ensure that 'all' is the default make target in all these
directories (defaulting to 'make install' is surprising and dangerous
IMHO).  Fix a couple small typos.

22 years agoStamp configure.in with 7.3b1, not just 7.3. Seems Marc does that
Bruce Momjian [Wed, 4 Sep 2002 08:08:29 +0000 (08:08 +0000)]
Stamp configure.in with 7.3b1, not just 7.3.   Seems Marc does that
usually anyway.

22 years agoUpdate based on TODO file contents.
Bruce Momjian [Wed, 4 Sep 2002 07:42:24 +0000 (07:42 +0000)]
Update based on TODO file contents.

22 years agoFile list cleanup.
Bruce Momjian [Wed, 4 Sep 2002 07:31:59 +0000 (07:31 +0000)]
File list cleanup.

22 years agoSpellcheck of HISTORY.
Bruce Momjian [Wed, 4 Sep 2002 07:30:33 +0000 (07:30 +0000)]
Spellcheck of HISTORY.

22 years agoUpdate files to be changed.
Bruce Momjian [Wed, 4 Sep 2002 07:26:37 +0000 (07:26 +0000)]
Update files to be changed.

22 years agoBrand 7.3. Ready for beta!
Bruce Momjian [Wed, 4 Sep 2002 07:23:04 +0000 (07:23 +0000)]
Brand 7.3.  Ready for beta!

22 years agoUpdate new HISTORY file for 7.3!
Bruce Momjian [Wed, 4 Sep 2002 07:16:32 +0000 (07:16 +0000)]
Update new HISTORY file for 7.3!

22 years agoRefrect changes made by Tom Lane
Tatsuo Ishii [Wed, 4 Sep 2002 02:54:59 +0000 (02:54 +0000)]
Refrect changes made by Tom Lane

22 years agoRefrect the changes to src/test/regress/sql/conversion.sql By Tom.
Tatsuo Ishii [Wed, 4 Sep 2002 02:42:34 +0000 (02:42 +0000)]
Refrect the changes to src/test/regress/sql/conversion.sql By Tom.

22 years agoAdd:
Bruce Momjian [Tue, 3 Sep 2002 23:39:04 +0000 (23:39 +0000)]
Add:

> * Gavin Sherry <swm@linuxworld.com.au>

22 years agoEXTRACT(EPOCH FROM timestamp) gave wrong answers in the int64-timestamp
Tom Lane [Tue, 3 Sep 2002 22:55:54 +0000 (22:55 +0000)]
EXTRACT(EPOCH FROM timestamp) gave wrong answers in the int64-timestamp
case for timestamptz input, and differently wrong answers in the float-
timestamp case for timestamp input.

22 years agoArrange for the default permissions on a database to allow temp table
Tom Lane [Tue, 3 Sep 2002 22:17:35 +0000 (22:17 +0000)]
Arrange for the default permissions on a database to allow temp table
creation to world, but disallow temp table creation in template1.  Per
latest round of pghackers discussion.
I did not force initdb, but the permissions lockdown on template1 will
not take effect unless you do one (or manually REVOKE TEMP ON DATABASE template1 FROM public).