OSDN Git Service

pg-rex/syncrep.git
20 years agoDocument all the system views created by initdb (several of these were
Tom Lane [Fri, 17 Oct 2003 22:38:20 +0000 (22:38 +0000)]
Document all the system views created by initdb (several of these were
never documented anywhere, sigh).  Centralize the detailed documentation
of system views into catalogs.sgml, and provide cross-references.

20 years agoAdd a listing of our SQLSTATE error codes to the documentation.
Tom Lane [Fri, 17 Oct 2003 18:57:01 +0000 (18:57 +0000)]
Add a listing of our SQLSTATE error codes to the documentation.
Link to it from some appropriate places.

20 years agoFix busted markup.
Tom Lane [Fri, 17 Oct 2003 18:29:55 +0000 (18:29 +0000)]
Fix busted markup.

20 years agoFix elog tab-insertion code to insert tabs only where wanted.
Tom Lane [Fri, 17 Oct 2003 16:49:03 +0000 (16:49 +0000)]
Fix elog tab-insertion code to insert tabs only where wanted.

20 years agoRemove automatic '()' in other branch of UNION as well.
Peter Eisentraut [Fri, 17 Oct 2003 11:52:06 +0000 (11:52 +0000)]
Remove automatic '()' in other branch of UNION as well.

20 years agoAdjust display of actual runtimes in EXPLAIN output to use three fractional
Tom Lane [Fri, 17 Oct 2003 01:14:26 +0000 (01:14 +0000)]
Adjust display of actual runtimes in EXPLAIN output to use three fractional
digits, and label it 'ms' not 'msec', for consistency with psql's \timing
display.  Per recent discussions.

20 years agoUse pg_get_constraintdef() rather than pg_constraint.consrc; this is
Tom Lane [Fri, 17 Oct 2003 00:57:04 +0000 (00:57 +0000)]
Use pg_get_constraintdef() rather than pg_constraint.consrc; this is
a portion of a patch recently submitted by Christopher Kings-Lynne.
Applied by agreement that this is a bug fix.

20 years agoAdd note deprecating use of consrc.
Tom Lane [Fri, 17 Oct 2003 00:55:17 +0000 (00:55 +0000)]
Add note deprecating use of consrc.

20 years agoFix bugs in referential_constraints view.
Peter Eisentraut [Thu, 16 Oct 2003 23:46:17 +0000 (23:46 +0000)]
Fix bugs in referential_constraints view.

20 years agoThere should not be an automatic '()' after a function name when
Peter Eisentraut [Thu, 16 Oct 2003 23:45:29 +0000 (23:45 +0000)]
There should not be an automatic '()' after a function name when
tab-completing.

20 years agoSet attstattarget to -1 for all system catalog columns. All but a few
Tom Lane [Thu, 16 Oct 2003 22:52:22 +0000 (22:52 +0000)]
Set attstattarget to -1 for all system catalog columns.  All but a few
of the entries used to be zero, which I think I had deliberately done in
the name of saving cycles during ANALYZE, but it was really a rather
foolish decision.  Some of the more complex views in information_schema
were getting really bad plans for lack of statistics on the columns they
were joining over.

I'm not forcing an initdb for this, but I think there will be one soon
anyway to repair some bugs in the information_schema views.

20 years agoAvoid division by zero in estimate_num_groups() when table has no rows.
Tom Lane [Thu, 16 Oct 2003 21:37:54 +0000 (21:37 +0000)]
Avoid division by zero in estimate_num_groups() when table has no rows.

20 years agoEnsure that all places that are complaining about exhaustion of shared
Tom Lane [Thu, 16 Oct 2003 20:59:35 +0000 (20:59 +0000)]
Ensure that all places that are complaining about exhaustion of shared
memory say 'out of shared memory'; some were doing that and some just
said 'out of memory'.  Also add a HINT about increasing max_locks_per_transaction
where relevant, per suggestion from Sean Chittenden.  (The former change
does not break the strings freeze; the latter does, but I think it's
worth doing anyway.)

20 years agoCause tab completion to do something moderately reasonable with mixed-case
Tom Lane [Thu, 16 Oct 2003 20:03:09 +0000 (20:03 +0000)]
Cause tab completion to do something moderately reasonable with mixed-case
identifiers --- it will now complete these correctly with double quoting.
Fix a few other issues in passing.

20 years agoAdd mention of SIGHUP on the postmaster manual page.
Bruce Momjian [Thu, 16 Oct 2003 17:38:01 +0000 (17:38 +0000)]
Add mention of SIGHUP on the postmaster manual page.

20 years agoFix bad interaction between NOTIFY processing and V3 extended query
Tom Lane [Thu, 16 Oct 2003 16:50:41 +0000 (16:50 +0000)]
Fix bad interaction between NOTIFY processing and V3 extended query
protocol, per report from Igor Shevchenko.  NOTIFY thought it could
do its thing if transaction blockState is TBLOCK_DEFAULT, but in
reality it had better check the low-level transaction state is
TRANS_DEFAULT as well.  Formerly it was not possible to wait for the
client in a state where the first is true and the second is not ...
but now we can have such a state.  Minor cleanup in StartTransaction()
as well.

20 years agoMinor copy-editing.
Tom Lane [Thu, 16 Oct 2003 04:52:21 +0000 (04:52 +0000)]
Minor copy-editing.

20 years agoAdd:
Bruce Momjian [Thu, 16 Oct 2003 04:12:53 +0000 (04:12 +0000)]
Add:

> * Allow pg_dump to dump CREATE CONVERSION (Christopher)

20 years agoFix SGML markup.
Bruce Momjian [Thu, 16 Oct 2003 04:11:53 +0000 (04:11 +0000)]
Fix SGML markup.

Update HISTORY file.

20 years agoThis patch adds another item to the "migration to 7.4" list in the
Bruce Momjian [Thu, 16 Oct 2003 03:55:31 +0000 (03:55 +0000)]
This patch adds another item to the "migration to 7.4" list in the
HISTORY file, and attributes the shared-buffer-alignment change to
Manfred Spraul.

Neil Conway

20 years agoI have run pg_autovacuum on AIX, so that can, at least loosely, be
Bruce Momjian [Thu, 16 Oct 2003 03:47:28 +0000 (03:47 +0000)]
I have run pg_autovacuum on AIX, so that can, at least loosely, be
considered "tested."

And I have turned all of the 7.4 HISTORY material into DocBook for
release.sgml

Christopher Browne

20 years agoCause the default CFLAGS to be -O2 -fno-strict-aliasing when using gcc,
Tom Lane [Wed, 15 Oct 2003 22:23:56 +0000 (22:23 +0000)]
Cause the default CFLAGS to be -O2 -fno-strict-aliasing when using gcc,
per recent pghackers discussions.  Also ensure that explicitly setting
CFLAGS in the configure environment will be honored.

20 years agoTranslation updates
Peter Eisentraut [Wed, 15 Oct 2003 10:00:20 +0000 (10:00 +0000)]
Translation updates

20 years agoNew translation
Peter Eisentraut [Wed, 15 Oct 2003 09:49:07 +0000 (09:49 +0000)]
New translation

20 years agoTranslation update
Peter Eisentraut [Wed, 15 Oct 2003 09:45:34 +0000 (09:45 +0000)]
Translation update

20 years agoNew translation
Peter Eisentraut [Wed, 15 Oct 2003 09:39:04 +0000 (09:39 +0000)]
New translation

20 years agoUse PQescapeString to ensure that tab-completion queries are not messed
Tom Lane [Tue, 14 Oct 2003 22:47:12 +0000 (22:47 +0000)]
Use PQescapeString to ensure that tab-completion queries are not messed
up by quotes or backslashes in words that are being matched to database
names (per gripe from Ian Barwick, though I didn't use his patch).
Also fix possible memory leakage if _complete_with_query isn't run to
completion (not clear if that can happen or not, but be safe).

20 years agoRemove void* in MemSet until we understand the gcc 3.3.1 problem better.
Bruce Momjian [Tue, 14 Oct 2003 19:08:00 +0000 (19:08 +0000)]
Remove void* in MemSet until we understand the gcc 3.3.1 problem better.

20 years agoAdd compile -O flag only for non-debug mode, per Tom
Bruce Momjian [Tue, 14 Oct 2003 00:48:09 +0000 (00:48 +0000)]
Add compile -O flag only for non-debug mode, per Tom

20 years agopull_up_subqueries() should copy the subquery before starting to modify
Tom Lane [Mon, 13 Oct 2003 23:48:16 +0000 (23:48 +0000)]
pull_up_subqueries() should copy the subquery before starting to modify
it.  Not sure why I'd thought it would be a good idea to do differently
way back when, but Greg Stark exposed the folly of doing so ...

20 years agoBack out makeNode() patch to fix gcc 3.3.1 warning.
Bruce Momjian [Mon, 13 Oct 2003 22:47:15 +0000 (22:47 +0000)]
Back out makeNode() patch to fix gcc 3.3.1 warning.

20 years agoRemove registration message, after discussion concluded that it was of
Peter Eisentraut [Mon, 13 Oct 2003 21:39:41 +0000 (21:39 +0000)]
Remove registration message, after discussion concluded that it was of
little use.

20 years agoTranslation update
Peter Eisentraut [Mon, 13 Oct 2003 21:38:37 +0000 (21:38 +0000)]
Translation update

20 years agoDetermine max_connections first, then see how large shared_buffers can
Tom Lane [Mon, 13 Oct 2003 21:06:44 +0000 (21:06 +0000)]
Determine max_connections first, then see how large shared_buffers can
be made, to avoid corner cases where max_connections ends up unreasonably
small because shared_buffers is hogging too much shmem space.  Per pghackers
discussion about a week ago.  Also, fix the copy-newlines problem in a
more robust way, by using COPY FROM filename instead of COPY FROM STDIN;
per a suggestion from Peter.

20 years agoAdjust setRelhassubclassInRelation() to not perform actual heap_update
Tom Lane [Mon, 13 Oct 2003 20:02:52 +0000 (20:02 +0000)]
Adjust setRelhassubclassInRelation() to not perform actual heap_update
when the pg_class.relhassubclass value is already correct.  This should
avoid most cases of the 'tuple concurrently updated' problem that
Robert Creager recently complained about.  Also remove a bunch of dead
code in StoreCatalogInheritance() --- it was still computing the complete
list of direct and indirect inheritance ancestors, though that list has
not been needed since we got rid of the pg_ipl catalog.

20 years agoUse makeNode() to allocate structures that have to be cast to Node *,
Bruce Momjian [Sun, 12 Oct 2003 23:19:21 +0000 (23:19 +0000)]
Use makeNode() to allocate structures that have to be cast to Node *,
rather than allocating them on the stack.

Fixes complaint from gcc 3.3.1.

20 years agoNew translation
Peter Eisentraut [Sun, 12 Oct 2003 19:16:36 +0000 (19:16 +0000)]
New translation

20 years agoAdded missing "commit" in test case.
Michael Meskes [Sun, 12 Oct 2003 15:19:46 +0000 (15:19 +0000)]
Added missing "commit" in test case.

20 years agoAdd void * cast to MemSet to silence compiler, and add comment that we
Bruce Momjian [Sat, 11 Oct 2003 19:53:45 +0000 (19:53 +0000)]
Add void * cast to MemSet to silence compiler, and add comment that we
already check for alignment.

20 years agoBack out -fstrict-aliasing void* casting.
Bruce Momjian [Sat, 11 Oct 2003 18:04:26 +0000 (18:04 +0000)]
Back out -fstrict-aliasing void* casting.

20 years agoTranslation updates
Peter Eisentraut [Sat, 11 Oct 2003 17:11:02 +0000 (17:11 +0000)]
Translation updates

20 years agoThis patch will stop gcc from issuing warnings about type-punned objects
Bruce Momjian [Sat, 11 Oct 2003 16:30:55 +0000 (16:30 +0000)]
This patch will stop gcc from issuing warnings about type-punned objects
when -fstrict-aliasing is turned on, as it is in the latest gcc when you
use -O2

Andrew Dunstan

20 years agoUpdate max existing db size to 32tb.
Bruce Momjian [Fri, 10 Oct 2003 21:27:13 +0000 (21:27 +0000)]
Update max existing db size to 32tb.

20 years agoFix core dump on root starting pg_autovacuum.
Bruce Momjian [Fri, 10 Oct 2003 19:26:25 +0000 (19:26 +0000)]
Fix core dump on root starting pg_autovacuum.

Change conditionals to be more constent with our code

if (const == var) => if (var == const).

20 years agoTranslation updates
Peter Eisentraut [Fri, 10 Oct 2003 08:01:44 +0000 (08:01 +0000)]
Translation updates

20 years agoRename __arm__/__arm__ to __arm__/__arm, found by Neil Conway
Bruce Momjian [Fri, 10 Oct 2003 03:58:57 +0000 (03:58 +0000)]
Rename __arm__/__arm__ to __arm__/__arm, found by Neil Conway

20 years agoAdd NULL regression tests.
Bruce Momjian [Fri, 10 Oct 2003 03:46:51 +0000 (03:46 +0000)]
Add NULL regression tests.

Manfred Koizar

20 years agoChange Win32 rename/unlink timeout to 3 seconds.
Bruce Momjian [Fri, 10 Oct 2003 03:27:58 +0000 (03:27 +0000)]
Change Win32 rename/unlink timeout to 3 seconds.

20 years agoUpdate Russian FAQ.
Bruce Momjian [Fri, 10 Oct 2003 02:52:12 +0000 (02:52 +0000)]
Update Russian FAQ.

Viktor Vislobokov

20 years agoAdd:
Bruce Momjian [Fri, 10 Oct 2003 02:42:09 +0000 (02:42 +0000)]
Add:

> * Prevent libpq's PQfnumber() from lowercasing the column name
572d572
< * Vadim is Vadim B. Mikheev <vadim4o@email.com> of Sector Data

20 years ago> That's a fairly useless place to put it, though, since someone would
Bruce Momjian [Fri, 10 Oct 2003 02:08:42 +0000 (02:08 +0000)]
> That's a fairly useless place to put it, though, since someone would
> only think to look at sort_mem if they already had a clue.  It should
> be mentioned under bulk data load (in performance tips chapter)

Attached is a doc patch that does this. The way I've worded it may not
be the best, though.

Neil Conway

20 years agoAdd count(*) TODO detail.
Bruce Momjian [Fri, 10 Oct 2003 01:53:06 +0000 (01:53 +0000)]
Add count(*) TODO detail.

20 years agoAdd TODO.detail for count(*):
Bruce Momjian [Fri, 10 Oct 2003 01:52:15 +0000 (01:52 +0000)]
Add TODO.detail for count(*):

>   to allow fast COUNT(*) queries with no WHERE clause(?) [count]

20 years agoAllow pg_id to compile on Win32.
Bruce Momjian [Fri, 10 Oct 2003 01:34:51 +0000 (01:34 +0000)]
Allow pg_id to compile on Win32.

Andrew Dunstan

20 years agoAdd:
Bruce Momjian [Thu, 9 Oct 2003 23:41:22 +0000 (23:41 +0000)]
Add:

>  o Allow PL/pgSQL to handle %TYPE arrays, e.g. tab.col%TYPE[]

20 years agoRemove -pipe from compiles.
Bruce Momjian [Thu, 9 Oct 2003 22:55:46 +0000 (22:55 +0000)]
Remove -pipe from compiles.

20 years agoAdd:
Bruce Momjian [Thu, 9 Oct 2003 19:36:27 +0000 (19:36 +0000)]
Add:

> * Have VACUUM FULL use REINDEX rather than index vacuum

20 years agoAbout section 8.5.1.5. ("Special Values"):
Bruce Momjian [Thu, 9 Oct 2003 19:13:48 +0000 (19:13 +0000)]
About section 8.5.1.5. ("Special Values"):

I suggest adding LOCALTIMESTAMP and LOCALTIME to the first paragraph.
Maybe it should be phrased as:

  The following SQL-compatible functions can be used to obtain
  current datetime-related values: CURRENT_DATE, CURRENT_TIME,
  CURRENT_TIMESTAMP, LOCALTIME and LOCALTIMESTAMP.
  (See Section 9.8.4.)

Troels Arvin

20 years agoMake sure that -- comments extend to the end of the line. This fixes the
Peter Eisentraut [Thu, 9 Oct 2003 19:13:23 +0000 (19:13 +0000)]
Make sure that -- comments extend to the end of the line.  This fixes the
misscanning of this construct:

SELECT ''hello world''
-- SELECT ''goodbye world''
::text;

20 years agoSome updates
Peter Eisentraut [Thu, 9 Oct 2003 19:11:03 +0000 (19:11 +0000)]
Some updates

20 years agoAdd mention of logrotate for syslog, from Robert Treat
Bruce Momjian [Thu, 9 Oct 2003 19:05:09 +0000 (19:05 +0000)]
Add mention of logrotate for syslog, from Robert Treat

20 years agoUpdate Solaris FAQ to mention -fast.
Bruce Momjian [Thu, 9 Oct 2003 17:11:13 +0000 (17:11 +0000)]
Update Solaris FAQ to mention -fast.

20 years agoRemove -pipe on Linux, per reports from Peter.
Bruce Momjian [Thu, 9 Oct 2003 16:52:45 +0000 (16:52 +0000)]
Remove -pipe on Linux, per reports from Peter.

20 years agoAdd -pipe compile for Linux.
Bruce Momjian [Thu, 9 Oct 2003 14:47:59 +0000 (14:47 +0000)]
Add -pipe compile for Linux.

20 years agoMake template CFLAGS handling consistent.
Bruce Momjian [Thu, 9 Oct 2003 14:40:37 +0000 (14:40 +0000)]
Make template CFLAGS handling consistent.

20 years agoSomeone report me small bug in contrib/pg_dumplo today. It's problem
Bruce Momjian [Thu, 9 Oct 2003 13:38:05 +0000 (13:38 +0000)]
Someone report  me small  bug in contrib/pg_dumplo  today. It's problem
with a little dirty snprintf() usage which I used some years ago:

    snprintf(path, BUFSIZ, "%s/lo_dump.index", path);

Karel Zak

20 years agoRe-add -pipe compile flags for bsd's.
Bruce Momjian [Thu, 9 Oct 2003 04:34:02 +0000 (04:34 +0000)]
Re-add -pipe compile flags for bsd's.

20 years agoUpdate:
Bruce Momjian [Thu, 9 Oct 2003 04:29:40 +0000 (04:29 +0000)]
Update:

> * COMMENT ON [ CAST | CONVERSION | OPERATOR CLASS | LARGE OBJECT | LANGUAGE ]
>   (Christopher)

20 years agoThis centralizes the optimization defaults into configure.in, rather
Bruce Momjian [Thu, 9 Oct 2003 03:20:34 +0000 (03:20 +0000)]
This centralizes the optimization defaults into configure.in, rather
than having CFLAGS= in the template files.

It uses -O2 for gcc (generated by autoconf), and -O for non-gcc, unless
the template overrides it.

20 years agoAdd:
Bruce Momjian [Thu, 9 Oct 2003 03:17:24 +0000 (03:17 +0000)]
Add:

> * COMMENT ON [ CAST | CONVERSION | OPERATOR CLASS | LARGE OBJECT ] (Christopher)

20 years agoChange "query:" to "statement:".
Bruce Momjian [Thu, 9 Oct 2003 02:40:18 +0000 (02:40 +0000)]
Change "query:" to "statement:".

Have log_duration print when log_min_duration_statement prints.

20 years agoAdd quotes to CFLAG define.
Bruce Momjian [Thu, 9 Oct 2003 02:37:09 +0000 (02:37 +0000)]
Add quotes to CFLAG define.

20 years agoProtected access to variable m_preparedCount via synchronized
Jan Wieck [Thu, 9 Oct 2003 01:17:07 +0000 (01:17 +0000)]
Protected access to variable m_preparedCount via synchronized
function to prevent multiple threads using automatic cursors on
the same connection from stomping over each others cursor.

Jan

20 years agoAdd:
Bruce Momjian [Wed, 8 Oct 2003 22:29:58 +0000 (22:29 +0000)]
Add:
<
346a346
> * Remove CREATE CONSTRAINT TRIGGER

20 years agoAdd 2-phase info, add Win32 URL:
Bruce Momjian [Wed, 8 Oct 2003 21:44:06 +0000 (21:44 +0000)]
Add 2-phase info, add Win32 URL:

<  o Sample implementation in contrib/rserv
29c28,29
< * Create native Win32 port [win32]
> * Create native Win32 port, http://momjian.postgresql.org/main/writings/pgsql/win32.html
>
367c367,368
< * Two-phase commit to implement distributed transactions
> * Add two-phase commit to all distributed transactions with
>   offline/readonly server status or administrator notification for failure

20 years agoNew translations
Peter Eisentraut [Wed, 8 Oct 2003 20:35:39 +0000 (20:35 +0000)]
New translations

20 years agoAdd:
Bruce Momjian [Wed, 8 Oct 2003 19:33:53 +0000 (19:33 +0000)]
Add:

> * Use background process to write dirty shared buffers to disk

20 years agoNew translations
Peter Eisentraut [Wed, 8 Oct 2003 18:38:49 +0000 (18:38 +0000)]
New translations

20 years agoAdd quotes around -O setting for consistency.
Bruce Momjian [Wed, 8 Oct 2003 18:23:08 +0000 (18:23 +0000)]
Add quotes around -O setting for consistency.

20 years agoChange some awkward wording in the pg_autovacuum README file. I really
Bruce Momjian [Wed, 8 Oct 2003 18:18:04 +0000 (18:18 +0000)]
Change some awkward wording in the pg_autovacuum README file. I really
only read this because of Niel :-)

Robert Treat

20 years agoRemoved my debug printf() call from numeric.c
Michael Meskes [Wed, 8 Oct 2003 08:28:57 +0000 (08:28 +0000)]
Removed my debug printf() call from numeric.c

20 years agoUse calloc() to allocate empty structures.
Bruce Momjian [Wed, 8 Oct 2003 03:52:32 +0000 (03:52 +0000)]
Use calloc() to allocate empty structures.

Fix pg_restore tar log output bug where Special flag wasn't being
initialized; bug seen on XP.

20 years agoHave log_min_duration_statement = 0 always print duration/statement.
Bruce Momjian [Wed, 8 Oct 2003 03:49:38 +0000 (03:49 +0000)]
Have log_min_duration_statement = 0 always print duration/statement.

Change log line to be "duration:  ms  query:"

Indent multi-line queries with a tab in the server logs.

20 years agoFixed floating point exception in long=>numeric conversion.
Michael Meskes [Tue, 7 Oct 2003 18:36:46 +0000 (18:36 +0000)]
Fixed floating point exception in long=>numeric conversion.

20 years agoFix include used by entab.
Bruce Momjian [Tue, 7 Oct 2003 17:40:09 +0000 (17:40 +0000)]
Fix include used by entab.

20 years agoAdd missing include file.
Peter Eisentraut [Tue, 7 Oct 2003 08:07:56 +0000 (08:07 +0000)]
Add missing include file.

20 years agoFixed error handling in Informix compat str to date conversion.
Michael Meskes [Tue, 7 Oct 2003 05:46:56 +0000 (05:46 +0000)]
Fixed error handling in Informix compat str to date conversion.

20 years agoNew Italian translation
Peter Eisentraut [Mon, 6 Oct 2003 21:50:45 +0000 (21:50 +0000)]
New Italian translation

20 years agoFix binary_oper_exact() so that the heuristic 'an unknown literal on
Tom Lane [Mon, 6 Oct 2003 20:09:47 +0000 (20:09 +0000)]
Fix binary_oper_exact() so that the heuristic 'an unknown literal on
one side of a binary operator is probably supposed to be the same type
as the other operand' will be applied for domain types.  This worked
in 7.3 but was broken in 7.4 due to code rearrangements.  Mea culpa.

20 years agoTranslation updates
Peter Eisentraut [Mon, 6 Oct 2003 17:37:39 +0000 (17:37 +0000)]
Translation updates

20 years agoDuring ALTER TABLE ADD FOREIGN KEY, try to check the existing rows using
Tom Lane [Mon, 6 Oct 2003 16:38:28 +0000 (16:38 +0000)]
During ALTER TABLE ADD FOREIGN KEY, try to check the existing rows using
a single LEFT JOIN query instead of firing the check trigger for each
row individually.  Stephan Szabo, with some kibitzing from Tom Lane and
Jan Wieck.

20 years agoNew Slovene translations
Peter Eisentraut [Mon, 6 Oct 2003 16:31:16 +0000 (16:31 +0000)]
New Slovene translations

20 years ago- Fixed constant listing in execute using clause.
Michael Meskes [Mon, 6 Oct 2003 06:44:55 +0000 (06:44 +0000)]
- Fixed constant listing in execute using clause.
- Fixed typo in ecpg for Informix dec_t type.
- Fixed precision handling in Informix compat funxtions.

20 years agoTranslation updates
Peter Eisentraut [Mon, 6 Oct 2003 06:20:11 +0000 (06:20 +0000)]
Translation updates

20 years agoModify COPY FROM to match the null-value string against the column value
Tom Lane [Mon, 6 Oct 2003 02:38:53 +0000 (02:38 +0000)]
Modify COPY FROM to match the null-value string against the column value
before it is de-backslashed, not after.  This allows the null string \N
to be reliably distinguished from the data value \N (which must be
represented as \\N).  Per bug report from Manfred Koizar ... but it's
amazing this hasn't been reported before ...
Also, be consistent about encoding conversion for null string: the form
specified in the command is in the server encoding, but what is sent
to/from client must be in client encoding.  This never worked quite
right before either.

20 years agoFix order of operations within SendQuery() so that the time spent in
Tom Lane [Mon, 6 Oct 2003 01:11:12 +0000 (01:11 +0000)]
Fix order of operations within SendQuery() so that the time spent in
data transfer during COPY is included in the \timing display.  Also
avoid portability problems if tv_usec is unsigned on some platform.

20 years agoAdd a little more smarts to estimate_hash_bucketsize(): if there's no
Tom Lane [Sun, 5 Oct 2003 22:44:25 +0000 (22:44 +0000)]
Add a little more smarts to estimate_hash_bucketsize(): if there's no
statistics, but there is a unique index on the column, we can safely
assume it's well-distributed.

20 years agoMake psql \timing show three digits.
Bruce Momjian [Sun, 5 Oct 2003 22:36:00 +0000 (22:36 +0000)]
Make psql \timing show three digits.

20 years agoTranslation updates
Peter Eisentraut [Sun, 5 Oct 2003 21:52:37 +0000 (21:52 +0000)]
Translation updates

20 years agoWhen revoking privileges from the owner, don't revoke the grant options,
Peter Eisentraut [Sun, 5 Oct 2003 21:49:12 +0000 (21:49 +0000)]
When revoking privileges from the owner, don't revoke the grant options,
to avoid recursively revoking everything from everyone.