OSDN Git Service

pg-rex/syncrep.git
20 years agoRemove unportable // comments.
Tom Lane [Sun, 27 Jul 2003 16:32:34 +0000 (16:32 +0000)]
Remove unportable // comments.

20 years agoCleanup of timetravel compile warnings.
Bruce Momjian [Sun, 27 Jul 2003 04:59:29 +0000 (04:59 +0000)]
Cleanup of timetravel compile warnings.

20 years agoError message editing in utils/adt. Again thanks to Joe Conway for doing
Tom Lane [Sun, 27 Jul 2003 04:53:12 +0000 (04:53 +0000)]
Error message editing in utils/adt.  Again thanks to Joe Conway for doing
the bulk of the heavy lifting ...

20 years agoHere is 4 file in tgz:
Bruce Momjian [Sun, 27 Jul 2003 04:51:45 +0000 (04:51 +0000)]
Here is 4 file in tgz:
the new timetravel.c,
new timetravel.README (cut from spi/README and modified),
modified timetravel.sql.in
and modified timetravel.example.

Features:
- optionally 3 parameter for insert/update/delete user name

- work with CREATE UNIQUE INDEX ixxx on table xxx
(unique_field,time_off);
    (the  original version was work with unique index on 6.5.0-6.5.3,
and not work on 7.3.2,7.3.3)
     (before 6.5.0 and between 6.5.3 and 7.3.2 I dont know)

- get_timetravel(tablename) function for check timetravel-status.

- timetravel trigger not change  oid of the active record. (it is not a
good feature, because the  old version is automatice prevent the paralel
update with "where oid=nnn")

B?jthe Zolt?n

20 years ago> Joe Conway <mail@joeconway.com> writes:
Bruce Momjian [Sun, 27 Jul 2003 04:35:54 +0000 (04:35 +0000)]
> Joe Conway <mail@joeconway.com> writes:
>>ISTM that "source" is worth knowing.
>
> Hm, possibly.  Any other opinions?

This version has the seven fields I proposed, including "source". Here's
an example that shows why I think it's valuable:

regression=# \x
Expanded display is on.
regression=# select * from pg_settings where name = 'enable_seqscan';
-[ RECORD 1 ]-----------
name    | enable_seqscan
setting | on
context | user
vartype | bool
source  | default
min_val |
max_val |

regression=# update pg_settings set setting = 'off' where name =
'enable_seqscan';
-[ RECORD 1 ]---
set_config | off

regression=# select * from pg_settings where name = 'enable_seqscan';
-[ RECORD 1 ]-----------
name    | enable_seqscan
setting | off
context | user
vartype | bool
source  | session
min_val |
max_val |

regression=# alter user postgres set enable_seqscan to 'off';
ALTER USER

(log out and then back in again)

regression=# \x
Expanded display is on.
regression=# select * from pg_settings where name = 'enable_seqscan';
-[ RECORD 1 ]-----------
name    | enable_seqscan
setting | off
context | user
vartype | bool
source  | user
min_val |
max_val |

In the first case, enable_seqscan is set to its default value. After
setting it to off, it is obvious that the value has been changed for the
session only. In the third case, you can see that the value has been set
specifically for the user.

Joe Conway

20 years ago> Am Son, 2003-06-22 um 02.09 schrieb Joe Conway:
Bruce Momjian [Sun, 27 Jul 2003 03:51:59 +0000 (03:51 +0000)]
> Am Son, 2003-06-22 um 02.09 schrieb Joe Conway:
>>Sounds like all that's needed for your case. But to be complete, in
>>addition to changing tablefunc.c we'd have to:
>>1) come up with a new function call signature that makes sense and does
>>not cause backward compatibility problems for other people
>>2) make needed changes to tablefunc.sql.in
>>3) adjust the README.tablefunc appropriately
>>4) adjust the regression test for new functionality
>>5) be sure we don't break any of the old cases
>>
>>If you want to submit a complete patch, it would be gratefully accepted
>>-- for review at least ;-)
>
> Here's the patch, at least for steps 1-3

Nabil Sayegh
Joe Conway

20 years agoThis is a totally trivial patch for something that was a very minor nit that
Bruce Momjian [Sun, 27 Jul 2003 03:47:22 +0000 (03:47 +0000)]
This is a totally trivial patch for something that was a very minor nit that
annoyed me the other day while I was documenting my current project. It
makes pg_dump use the same layout for types as for tables, by putting "\n\t"
before the first field and "\n" before the final ");"

Can't really justify this too much except to say I had an itch and I
scratched it ;-)

Andrew Dunstan

20 years agohere are the patches for psql on Win32:
Bruce Momjian [Sun, 27 Jul 2003 03:32:26 +0000 (03:32 +0000)]
here are the patches for psql on Win32:

  psql4win32.patch  - changes in the psql source code
  psql-ref.patch    - changes in the documentation psql-ref.sgml
                      (for new builtin variable WIN32_CONSOLE)

To apply them use "patch -p 1" in the root directory of the
postgres source directory.

These patches fix the following problems of psql on Win32
(all changes only have effect #ifdef WIN32):

  a) Problem:  Static library libpq.a did not work
     Solution: Added WSAStartup() in fe-connect.c

  b) Problem:  Secret Password was echoed by psql
     Solution: Password echoing disabled in sprompt.c

  c) Problem:  8bit characters were displayed/interpreted wrong in psql
               This is due to the fact that the Win32 "console" uses a
               different encoding than the rest of the Windows system
     Solution: Introduced a new psql variable WIN32_CONSOLE
               When set with "\set WIN32_console", the function OemToChar()
               is applied after reading input and CharToOem() before
               displaying Output

Christoph Dalitz

20 years agoThis makes the initcap function compatible with Oracle 9i, it has been
Bruce Momjian [Sun, 27 Jul 2003 03:16:20 +0000 (03:16 +0000)]
This makes the initcap function compatible with Oracle 9i, it has been
tested on both redhat 8 and FreebSD.
--
Mike Nolan

20 years agoThe deferred trigger queue pushing to disk patch pointed out
Bruce Momjian [Sun, 27 Jul 2003 03:13:17 +0000 (03:13 +0000)]
The deferred trigger queue pushing to disk patch pointed out
that the regression tests for foreign keys didn't seem to test
a deferred constraint that was not satisified by a later
statement and was not made immediate by set constraints,
so here's a simple added test with a single invalid insert and
a commit.

Stephan Szabo

20 years ago>>You can alias $0, similar to the argument variables. And, I confirmed
Bruce Momjian [Sat, 26 Jul 2003 23:58:23 +0000 (23:58 +0000)]
>>You can alias $0, similar to the argument variables. And, I confirmed
>>that you cannot change the value, similar to the argument variables:
>
> Perhaps you shouldn't mark it isconst; then it would actually have some
> usefulness (you could use it directly as a temporary variable to hold
> the intended result).  I can't see much value in aliasing it if it's
> const, either.

OK; the only change in this version is "isconst = false;". Now you can
use $0 as a result placeholder if desired. E.g.:

create or replace function tmp(anyelement, anyelement) returns anyarray as '
declare
  v_ret alias for $0;
  v_el1 alias for $1;
  v_el2 alias for $2;
begin
  v_ret := ARRAY[v_el1, v_el2];
  return v_ret;
end;
' language 'plpgsql';

create table f(f1 text, f2 text, f3 int, f4 int);
insert into f values ('a','b',1,2);
insert into f values ('z','x',3,4);

select tmp(f1,f2) from f;
select tmp(f3,f4) from f;

Joe Conway

20 years agoFix for Linux startup script.
Bruce Momjian [Sat, 26 Jul 2003 20:42:55 +0000 (20:42 +0000)]
Fix for Linux startup script.

Slawomir Sudnik

20 years agoHave SSL text print only when SSL mode is enabled.
Bruce Momjian [Sat, 26 Jul 2003 15:22:22 +0000 (15:22 +0000)]
Have SSL text print only when SSL mode is enabled.

20 years agoI corecting date_trunc('quarter',...) and friends because orig version
Bruce Momjian [Sat, 26 Jul 2003 15:17:36 +0000 (15:17 +0000)]
I corecting date_trunc('quarter',...) and friends because orig version
doing '2003-07-30' -> '2003-04-01', '2003-11-30' ->'2003-07-01'

B?jthe Zolt?n

20 years agoAt long last I put together a patch to support 4 client SSL negotiation
Bruce Momjian [Sat, 26 Jul 2003 13:50:02 +0000 (13:50 +0000)]
At long last I put together a patch to support 4 client SSL negotiation
modes (and replace the requiressl boolean). The four options were first
spelled out by Magnus Hagander <mha@sollentuna.net> on 2000-08-23 in email
to pgsql-hackers, archived here:

http://archives.postgresql.org/pgsql-hackers/2000-08/msg00639.php

My original less-flexible patch and the ensuing thread are archived at:

http://dbforums.com/t623845.html

Attached is a new patch, including documentation.

To sum up, there's a new client parameter "sslmode" and environment
variable "PGSSLMODE", with these options:

sslmode   description
-------   -----------
disable   Unencrypted non-SSL only
allow     Negotiate, prefer non-SSL
prefer    Negotiate, prefer SSL (default)
require   Require SSL

The only change to the server is a new pg_hba.conf line type,
"hostnossl", for specifying connections that are not allowed to use SSL
(for example, to prevent servers on a local network from accidentally
using SSL and wasting cycles). Thus the 3 pg_hba.conf line types are:

pg_hba.conf line types
----------------------
host       applies to either SSL or regular connections
hostssl    applies only to SSL connections
hostnossl  applies only to regular connections

These client and server options, the postgresql.conf ssl = false option,
and finally the possibility of compiling with no SSL support at all,
make quite a range of combinations to test. I threw together a test
script to try many of them out. It's in a separate tarball with its
config files, a patch to psql so it'll announce SSL connections even in
absence of a tty, and the test output. The test is especially informative
when run on the same tty the postmaster was started on, so the FATAL:
errors during negotiation are interleaved with the psql client output.

I saw Tom write that new submissions for 7.4 have to be in before midnight
local time, and since I'm on the east coast in the US, this just makes it
in before the bell. :)

Jon Jensen

20 years agoUpdate to match error editing.
Tom Lane [Sat, 26 Jul 2003 00:02:02 +0000 (00:02 +0000)]
Update to match error editing.

20 years agoError message editing in src/pl. The plpython module could use another
Tom Lane [Fri, 25 Jul 2003 23:37:31 +0000 (23:37 +0000)]
Error message editing in src/pl.  The plpython module could use another
look ... I'm not real certain which errors are strictly internal and which
are likely to be provoked by users.

20 years agoDone:
Bruce Momjian [Fri, 25 Jul 2003 22:08:36 +0000 (22:08 +0000)]
Done:

> * -Allow SSL-enabled clients to turn off SSL transfers

20 years agoAdd mention of new SSL capabilities.
Bruce Momjian [Fri, 25 Jul 2003 22:07:44 +0000 (22:07 +0000)]
Add mention of new SSL capabilities.

20 years agoUpdate copyrights to 2003.
Bruce Momjian [Fri, 25 Jul 2003 21:48:45 +0000 (21:48 +0000)]
Update copyrights to 2003.

print.c: Add one more line to pager calculation to account for the prompt.
help.c: Call PageOutput with correct number of lines within slashUsage
        Add one to line count in helpSQL to account for "Available help:" line.
        Make copyright match COPYRIGHT file. (Just "1994")

Greg Sabino Mullane

20 years ago> Having read the list, and noticed the message about table inheritance I
Bruce Momjian [Fri, 25 Jul 2003 21:42:26 +0000 (21:42 +0000)]
> Having read the list, and noticed the message about table inheritance I
> thought that I would see if I could come up with a simple solution, and
> have my first delve into the code for PostgreSQL.
>
> Attached is a diff against 7.3.3 source, of changes to describe.c for
> psql. This should print out a list of parent tables in a similar style
> to that of the index listing. I have done some testing on my side and it
> all seems fine, can some other people have a quick look? What do people
> think? Useful?

Nick Barr

20 years agoRecent patch to dump nondefault attstorage settings broke pg_dump for
Tom Lane [Fri, 25 Jul 2003 21:02:52 +0000 (21:02 +0000)]
Recent patch to dump nondefault attstorage settings broke pg_dump for
dropped columns.  Fix by using LEFT JOIN rather than straight join
between pg_attribute and pg_type.  Also, use pg_type.oid as input to
format_type, so that we don't get a failure on deleted types of deleted
columns (this may be a change we ought to backpatch to 7.3....).

20 years agoError message editing in backend/utils (except /adt).
Tom Lane [Fri, 25 Jul 2003 20:18:01 +0000 (20:18 +0000)]
Error message editing in backend/utils (except /adt).

20 years agoSeems my check constraint change did break stuff.
Bruce Momjian [Fri, 25 Jul 2003 19:37:21 +0000 (19:37 +0000)]
Seems my check constraint change did break stuff.

Alias the appropriate columns back to their original name.

Fixed formatting of a few other places as I went along (indenting)
--
Rod Taylor <rbt@rbt.ca>

20 years ago> Rod Taylor <rbt@rbt.ca> writes:
Bruce Momjian [Fri, 25 Jul 2003 19:27:06 +0000 (19:27 +0000)]
> Rod Taylor <rbt@rbt.ca> writes:
> > It seems that readline() on my system (FreeBSD 4.8) isn't declared to
> > take the prompt as a const.  Thus, remove const from gets_interactive()
> > to remove the warning.
>
> I think it would be a lot cleaner to just put a cast to char * into the
> readline call (with a note about why).

Ok.. that works.

I must say it's a little strange being able to take a constant and say
its no longer constant anymore -- but I suppose it's no different than
defining then undefining pre-processor constants.

Rod Taylor <rbt@rbt.ca>

20 years agoAdded explicit casts for date/interval/timestamp.
Michael Meskes [Fri, 25 Jul 2003 16:10:26 +0000 (16:10 +0000)]
Added explicit casts for date/interval/timestamp.

20 years agoBegin improving markup of (past) release notes.
Peter Eisentraut [Fri, 25 Jul 2003 09:00:17 +0000 (09:00 +0000)]
Begin improving markup of (past) release notes.

20 years agoApplied Peter's patch to use yyless instead of my string_unput function.
Michael Meskes [Fri, 25 Jul 2003 05:42:27 +0000 (05:42 +0000)]
Applied Peter's patch to use yyless instead of my string_unput function.

20 years agoNo need for Static.Entries, use -rHEAD.
Bruce Momjian [Fri, 25 Jul 2003 02:11:17 +0000 (02:11 +0000)]
No need for Static.Entries, use -rHEAD.

20 years agoAdd example of cvs log pulls _just_ from HEAD.
Bruce Momjian [Fri, 25 Jul 2003 01:40:02 +0000 (01:40 +0000)]
Add example of cvs log pulls _just_ from HEAD.

20 years agoError message editing in backend/optimizer, backend/rewrite.
Tom Lane [Fri, 25 Jul 2003 00:01:09 +0000 (00:01 +0000)]
Error message editing in backend/optimizer, backend/rewrite.

20 years agoError message editing in backend/storage.
Tom Lane [Thu, 24 Jul 2003 22:04:15 +0000 (22:04 +0000)]
Error message editing in backend/storage.

20 years agoFix a few portability issues, adjust error messages.
Tom Lane [Thu, 24 Jul 2003 17:59:56 +0000 (17:59 +0000)]
Fix a few portability issues, adjust error messages.

20 years agoError message editing in contrib (mostly by Joe Conway --- thanks Joe!)
Tom Lane [Thu, 24 Jul 2003 17:52:50 +0000 (17:52 +0000)]
Error message editing in contrib (mostly by Joe Conway --- thanks Joe!)

20 years agoAdd tsearch2.
Tom Lane [Thu, 24 Jul 2003 16:54:58 +0000 (16:54 +0000)]
Add tsearch2.

20 years agoFix grant option dumping and related cross-version compatibility issues.
Peter Eisentraut [Thu, 24 Jul 2003 15:52:53 +0000 (15:52 +0000)]
Fix grant option dumping and related cross-version compatibility issues.

20 years agoFixed mdy functions to use correct offset.
Michael Meskes [Thu, 24 Jul 2003 08:41:07 +0000 (08:41 +0000)]
Fixed mdy functions to use correct offset.

20 years agoFix timestamp_date for HAVE_INT64_TIMESTAMP case.
Tom Lane [Thu, 24 Jul 2003 04:38:19 +0000 (04:38 +0000)]
Fix timestamp_date for HAVE_INT64_TIMESTAMP case.

21 years agoAdd to 7.3.4 release notes:
Bruce Momjian [Thu, 24 Jul 2003 00:43:53 +0000 (00:43 +0000)]
Add to 7.3.4 release notes:

* Repair breakage in timestamp-to-date conversion for dates before 2000

21 years agoFixes additional sql injection vulnerabilities reported by Oliver Jowett
Barry Lind [Thu, 24 Jul 2003 00:30:39 +0000 (00:30 +0000)]
Fixes additional sql injection vulnerabilities reported by Oliver Jowett
and Dmitry Tkach.  Specifically the previous fix still allowed the statement termination character through in unquoted places in the sql statement, and the driver never correctly handled someone passing a value of \0 in a string which under the v2 protocol would end the statement causing the following text to possibly
be treated as a new sql statement
 Modified Files:
  jdbc/org/postgresql/Driver.java.in
  jdbc/org/postgresql/jdbc1/AbstractJdbc1Statement.java

21 years agoRepair 7.3 breakage in timestamp-to-date conversion for dates before 2000.
Tom Lane [Thu, 24 Jul 2003 00:21:26 +0000 (00:21 +0000)]
Repair 7.3 breakage in timestamp-to-date conversion for dates before 2000.

21 years agoDon't refer to AF_UNIX in code not protected with HAVE_UNIX_SOCKETS.
Tom Lane [Thu, 24 Jul 2003 00:02:53 +0000 (00:02 +0000)]
Don't refer to AF_UNIX in code not protected with HAVE_UNIX_SOCKETS.

21 years agoHave a go at fixing various outstanding portability issues in code that
Tom Lane [Wed, 23 Jul 2003 23:30:41 +0000 (23:30 +0000)]
Have a go at fixing various outstanding portability issues in code that
was modified for IPv6.  Use a robust definition of struct sockaddr_storage,
do a proper configure test to see if ss_len exists, don't assume that
getnameinfo() will handle AF_UNIX sockets, don't trust getaddrinfo to
return the protocol we ask for, etc.  This incorporates several outstanding
patches from Kurt Roeckx, but I'm to blame for anything that doesn't
work ...

21 years agoNew WAL fix release text is:
Bruce Momjian [Wed, 23 Jul 2003 22:58:38 +0000 (22:58 +0000)]
New WAL fix release text is:

    Prevent rare possibility of server startup failure (Tom)

21 years agoMention increasing sort_mem during a restore.
Bruce Momjian [Wed, 23 Jul 2003 20:30:35 +0000 (20:30 +0000)]
Mention increasing sort_mem during a restore.

21 years agoAdd:
Bruce Momjian [Wed, 23 Jul 2003 18:16:39 +0000 (18:16 +0000)]
Add:

> * Add GUC variable to prevent waiting on locks

21 years agoRename --with-threads to --enable-thread-safeness, per Peter.
Bruce Momjian [Wed, 23 Jul 2003 17:27:28 +0000 (17:27 +0000)]
Rename --with-threads to --enable-thread-safeness, per Peter.

21 years agoUpdate German FAQ.
Bruce Momjian [Wed, 23 Jul 2003 17:08:04 +0000 (17:08 +0000)]
Update German FAQ.

21 years agoUpdate FAQ.
Bruce Momjian [Wed, 23 Jul 2003 16:55:47 +0000 (16:55 +0000)]
Update FAQ.

Ian Barwick

21 years agoJust a spot of copy-editing.
Tom Lane [Wed, 23 Jul 2003 15:05:42 +0000 (15:05 +0000)]
Just a spot of copy-editing.

21 years agoCzech translation updates from Karel Zak
Peter Eisentraut [Wed, 23 Jul 2003 09:36:43 +0000 (09:36 +0000)]
Czech translation updates from Karel Zak

21 years agoUpdate German translations.
Peter Eisentraut [Wed, 23 Jul 2003 08:49:30 +0000 (08:49 +0000)]
Update German translations.

21 years agoApply message style guide to frontend programs.
Peter Eisentraut [Wed, 23 Jul 2003 08:47:41 +0000 (08:47 +0000)]
Apply message style guide to frontend programs.

21 years agoUpdate to 7.3.4 as newest release.
Bruce Momjian [Wed, 23 Jul 2003 04:11:21 +0000 (04:11 +0000)]
Update to 7.3.4 as newest release.

21 years agoStamp 7.3.4.
Bruce Momjian [Wed, 23 Jul 2003 04:08:44 +0000 (04:08 +0000)]
Stamp 7.3.4.

21 years agoError message editing in backend/bootstrap, /lib, /nodes, /port.
Tom Lane [Tue, 22 Jul 2003 23:30:39 +0000 (23:30 +0000)]
Error message editing in backend/bootstrap, /lib, /nodes, /port.

21 years agoError message editing for foreign-key triggers.
Tom Lane [Tue, 22 Jul 2003 22:14:57 +0000 (22:14 +0000)]
Error message editing for foreign-key triggers.

21 years agoA few parentheses shy of a load here ...
Tom Lane [Tue, 22 Jul 2003 21:19:22 +0000 (21:19 +0000)]
A few parentheses shy of a load here ...

21 years agoAdd GUC parameter to control rendezvous name.
Bruce Momjian [Tue, 22 Jul 2003 20:29:13 +0000 (20:29 +0000)]
Add GUC parameter to control rendezvous name.

21 years agowups, took out one memset too many ...
Tom Lane [Tue, 22 Jul 2003 19:13:19 +0000 (19:13 +0000)]
wups, took out one memset too many ...

21 years agoError message editing in backend/libpq, backend/postmaster, backend/tcop.
Tom Lane [Tue, 22 Jul 2003 19:00:12 +0000 (19:00 +0000)]
Error message editing in backend/libpq, backend/postmaster, backend/tcop.

Along the way, fix some logic problems in pgstat_initstats, notably the
bogus assumption that malloc returns zeroed memory.

21 years agoFix threading configure tests, from Peter.
Bruce Momjian [Tue, 22 Jul 2003 16:39:55 +0000 (16:39 +0000)]
Fix threading configure tests, from Peter.

21 years agoRemove link of tarball
Teodor Sigaev [Tue, 22 Jul 2003 14:06:36 +0000 (14:06 +0000)]
Remove link of tarball

21 years agoMark tsearch v1 as deprecated
Teodor Sigaev [Tue, 22 Jul 2003 13:21:48 +0000 (13:21 +0000)]
Mark tsearch v1 as deprecated

21 years agoFix to prevent SQL injection attacks for code calling setObject(int,Object,int)
Barry Lind [Tue, 22 Jul 2003 05:17:09 +0000 (05:17 +0000)]
Fix to prevent SQL injection attacks for code calling setObject(int,Object,int)
where Object is a user supplied String and the type is a numeric type
(i.e. INTEGER,LONG,etc).
Also applied a patch from Kim Ho that fixes compile problems under jdk1.2

 Modified Files:
  jdbc/org/postgresql/Driver.java.in
  jdbc/org/postgresql/jdbc1/AbstractJdbc1Statement.java

21 years agoAdd:
Bruce Momjian [Tue, 22 Jul 2003 01:06:51 +0000 (01:06 +0000)]
Add:

>  o Allow PL/pgSQL to name columns by ordinal position, e.g. rec.(3)

21 years agoAdd:
Bruce Momjian [Tue, 22 Jul 2003 00:49:22 +0000 (00:49 +0000)]
Add:

> * Prevent COMMENT ON DATABASE from using a database name

21 years agoDon't report sequendces in pg_tables.
Bruce Momjian [Tue, 22 Jul 2003 00:02:55 +0000 (00:02 +0000)]
Don't report sequendces in pg_tables.

21 years agoBack out comment on GucContexts.
Bruce Momjian [Mon, 21 Jul 2003 21:02:12 +0000 (21:02 +0000)]
Back out comment on  GucContexts.

21 years agoApplied patch from dmitry@openratings.com to fix parsing of array values
Barry Lind [Mon, 21 Jul 2003 20:48:31 +0000 (20:48 +0000)]
Applied patch from dmitry@openratings.com to fix parsing of array values

 Modified Files:
  jdbc/org/postgresql/Driver.java.in
  jdbc/org/postgresql/jdbc2/Array.java

21 years agoError message editing in backend/access.
Tom Lane [Mon, 21 Jul 2003 20:29:40 +0000 (20:29 +0000)]
Error message editing in backend/access.

21 years agoFix regression test broken by recent view change.
Tom Lane [Mon, 21 Jul 2003 20:24:31 +0000 (20:24 +0000)]
Fix regression test broken by recent view change.

21 years agoAdd:
Bruce Momjian [Mon, 21 Jul 2003 19:15:48 +0000 (19:15 +0000)]
Add:

>  o Add PL/PHP (Joe, Jan)

21 years agoAdd:
Bruce Momjian [Mon, 21 Jul 2003 19:04:54 +0000 (19:04 +0000)]
Add:

> * Have AFTER triggers execute after the appropriate SQL statement in a
>   function, not at the end of the function

21 years agoError message editing in backend/executor.
Tom Lane [Mon, 21 Jul 2003 17:05:12 +0000 (17:05 +0000)]
Error message editing in backend/executor.

21 years agoFix relkind test for pg_tables.
Bruce Momjian [Mon, 21 Jul 2003 15:20:27 +0000 (15:20 +0000)]
Fix relkind test for pg_tables.

Mike Quinn

21 years agoFix output to psql:tsearch2.sql:13: NOTICE: ... "pg_ts_dict_pkey"
Teodor Sigaev [Mon, 21 Jul 2003 15:15:19 +0000 (15:15 +0000)]
Fix output to psql:tsearch2.sql:13: NOTICE:  ... "pg_ts_dict_pkey"

21 years agotsearch2 module
Teodor Sigaev [Mon, 21 Jul 2003 10:27:44 +0000 (10:27 +0000)]
tsearch2 module

21 years ago I updated Japanese translation of PostgreSQL FAQ
Bruce Momjian [Mon, 21 Jul 2003 04:18:41 +0000 (04:18 +0000)]
  I updated Japanese translation of PostgreSQL FAQ
according to your version on May/30.
(HTML and TEXT files are gzipped.)

  Jun Kuwamura

21 years agoError message editing in backend/catalog.
Tom Lane [Mon, 21 Jul 2003 01:59:11 +0000 (01:59 +0000)]
Error message editing in backend/catalog.

21 years agoAnother round of error message editing, covering backend/commands/.
Tom Lane [Sun, 20 Jul 2003 21:56:35 +0000 (21:56 +0000)]
Another round of error message editing, covering backend/commands/.

21 years agoUpdate URL's.
Bruce Momjian [Sun, 20 Jul 2003 21:52:46 +0000 (21:52 +0000)]
Update URL's.

21 years agoDone:
Bruce Momjian [Sun, 20 Jul 2003 19:21:41 +0000 (19:21 +0000)]
Done:

* -Allow LIMIT/OFFSET to use expressions (Tom)

21 years agoAdd Opteron/Itanium comment.
Bruce Momjian [Sun, 20 Jul 2003 04:31:32 +0000 (04:31 +0000)]
Add Opteron/Itanium comment.

21 years agoAdd:
Bruce Momjian [Sun, 20 Jul 2003 03:12:13 +0000 (03:12 +0000)]
Add:

> * Delay resolution of array expression type so assignment coercion
>   can be performed on empty array expressions (Joe)

21 years agoError message editing: finish up undone task of reporting the problem
Tom Lane [Sat, 19 Jul 2003 21:37:37 +0000 (21:37 +0000)]
Error message editing: finish up undone task of reporting the problem
xid when we fail to access pg_clog.

21 years agoAdd ereport-related functions to GETTEXT_TRIGGERS list.
Tom Lane [Sat, 19 Jul 2003 20:32:12 +0000 (20:32 +0000)]
Add ereport-related functions to GETTEXT_TRIGGERS list.

21 years agoAnother round of error message editing, covering backend/parser/.
Tom Lane [Sat, 19 Jul 2003 20:20:53 +0000 (20:20 +0000)]
Another round of error message editing, covering backend/parser/.

21 years agoAdd tablespace email.
Bruce Momjian [Sat, 19 Jul 2003 14:09:24 +0000 (14:09 +0000)]
Add tablespace email.

21 years agoFirst bits of work on error message editing.
Tom Lane [Fri, 18 Jul 2003 23:20:33 +0000 (23:20 +0000)]
First bits of work on error message editing.

21 years agoAdd cleaner formatting to config file.
Bruce Momjian [Fri, 18 Jul 2003 19:16:03 +0000 (19:16 +0000)]
Add cleaner formatting to config file.

21 years agoAdded more compat stuff ot the parser.
Michael Meskes [Fri, 18 Jul 2003 14:32:56 +0000 (14:32 +0000)]
Added more compat stuff ot the parser.

21 years agoFix subpath and subltree. Allow to return '' value.
Teodor Sigaev [Fri, 18 Jul 2003 13:27:43 +0000 (13:27 +0000)]
Fix subpath and subltree. Allow to return '' value.
subpath(ltree,0,0) returns ''.

21 years agoStephen Robert Norris wrote:
Bruce Momjian [Fri, 18 Jul 2003 03:45:06 +0000 (03:45 +0000)]
Stephen Robert Norris wrote:
> Well, no. What it says is that certain values must be escaped (but
> doesn't say which ones). Then it says there are alternate escape
> sequences for some values, which it lists.
>
> It doesn't say "The following table contains the characters which must
> be escaped:", which would be much clearer (and actually useful).

Attached documentation patch updates the wording for bytea input
escaping, per complaint by Stephen Norris above.

Joe Conway

21 years agoIt seems some case-insensitive locales sort upper case before lower
Tom Lane [Fri, 18 Jul 2003 03:21:53 +0000 (03:21 +0000)]
It seems some case-insensitive locales sort upper case before lower
(BBBB before bbbb) and others the other way around.  Provide comparison
files that cater to both approaches.

21 years agoAdd an upper limit to IS_VALID_JULIAN() to defend against overflow in
Tom Lane [Thu, 17 Jul 2003 22:28:42 +0000 (22:28 +0000)]
Add an upper limit to IS_VALID_JULIAN() to defend against overflow in
date2j().  This ensures we give reasonable errors instead of bizarre
behavior for input dates far in the future.

21 years agoOh, for crying in a bucket ... relax Assert so that glibc's strxfrm
Tom Lane [Thu, 17 Jul 2003 22:20:14 +0000 (22:20 +0000)]
Oh, for crying in a bucket ... relax Assert so that glibc's strxfrm
does not dump core.

21 years agoWork around buggy strxfrm() present in some Solaris releases.
Tom Lane [Thu, 17 Jul 2003 20:52:36 +0000 (20:52 +0000)]
Work around buggy strxfrm() present in some Solaris releases.

21 years agoFor COMMENT ON DATABASE where database name is unknown or not the current
Tom Lane [Thu, 17 Jul 2003 20:13:57 +0000 (20:13 +0000)]
For COMMENT ON DATABASE where database name is unknown or not the current
database, emit a WARNING and do nothing, rather than raising ERROR.
Per recent discussion in which we concluded this is the best way to deal
with database dumps that are reloaded into a database of a new name.

21 years agoRepair boundary-case bug introduced by patch of two months ago that
Tom Lane [Thu, 17 Jul 2003 16:45:04 +0000 (16:45 +0000)]
Repair boundary-case bug introduced by patch of two months ago that
fixed incorrect initial setting of StartUpID.  The logic in XLogWrite()
expects that Write->curridx is advanced to the next page as soon as
LogwrtResult points to the end of the current page, but StartupXLOG()
failed to make that happen when the old WAL ended exactly on a page
boundary.  Per trouble report from Hannu Krosing.