OSDN Git Service

pg-rex/syncrep.git
18 years agoFix release item ordering.
Bruce Momjian [Sun, 12 Feb 2006 18:13:10 +0000 (18:13 +0000)]
Fix release item ordering.

18 years agoUpdate back branch release notes.
Bruce Momjian [Sun, 12 Feb 2006 17:57:37 +0000 (17:57 +0000)]
Update back branch release notes.

18 years agoFix release markup.
Bruce Momjian [Sun, 12 Feb 2006 17:27:33 +0000 (17:27 +0000)]
Fix release markup.

18 years agoUpdate release checlist.
Bruce Momjian [Sun, 12 Feb 2006 17:23:31 +0000 (17:23 +0000)]
Update release checlist.

18 years agoUpdate for 8.1.3.
Bruce Momjian [Sun, 12 Feb 2006 17:15:57 +0000 (17:15 +0000)]
Update for 8.1.3.

18 years agoUpdate release notes for 8.1.3.
Bruce Momjian [Sun, 12 Feb 2006 17:10:47 +0000 (17:10 +0000)]
Update release notes for 8.1.3.

18 years agoImprove tab completion macro for default keywords.
Bruce Momjian [Sun, 12 Feb 2006 15:25:04 +0000 (15:25 +0000)]
Improve tab completion macro for default keywords.

18 years agoWhen performing a parallel build (make -j N) with ./configure
Bruce Momjian [Sun, 12 Feb 2006 07:29:36 +0000 (07:29 +0000)]
When performing a parallel build (make -j N) with ./configure
--enable-depend it often tries to create the .deps directory twice and
bails out when it already exists due to a race condition of if doesn't
exist, then create.  This patch prevents mkdir from returning an error.

Kris Jurka

18 years agoList psql tab completion for various default keywords.
Bruce Momjian [Sun, 12 Feb 2006 07:21:40 +0000 (07:21 +0000)]
List psql tab completion for various default keywords.

Joachim Wieland

18 years agoDone:
Bruce Momjian [Sun, 12 Feb 2006 06:51:14 +0000 (06:51 +0000)]
Done:

> * -Add COMMENT ON for all cluster global objects (roles, databases

18 years agoApply code-reviewed version of for-scalar-list patch: mostly, fixing
Tom Lane [Sun, 12 Feb 2006 06:37:05 +0000 (06:37 +0000)]
Apply code-reviewed version of for-scalar-list patch: mostly, fixing
it to report reasonable errors in error cases.

18 years agoAdd MSVC support for utility commands and pg_dump.
Bruce Momjian [Sun, 12 Feb 2006 06:11:51 +0000 (06:11 +0000)]
Add MSVC support for utility commands and pg_dump.

Hiroshi Saito

18 years agoAllow PL/pgSQL FOR statement to return values to scalars as well as
Bruce Momjian [Sun, 12 Feb 2006 06:03:38 +0000 (06:03 +0000)]
Allow PL/pgSQL FOR statement to return values to scalars as well as
records and row types.

Pavel Stehule

18 years agoUse 0x01 for newlines in saved history. It was determined to be not
Bruce Momjian [Sun, 12 Feb 2006 05:24:38 +0000 (05:24 +0000)]
Use 0x01 for newlines in saved history.  It was determined to be not
used by multi-byte sequences, but futher invesetigation might prove this
to be false.

18 years agoClean up plpgsql grammar to make sure that check_assignable() is applied
Tom Lane [Sun, 12 Feb 2006 04:59:32 +0000 (04:59 +0000)]
Clean up plpgsql grammar to make sure that check_assignable() is applied
consistently.  This is mostly cosmetic right at the moment because
check_assignable() does nothing for ROW or RECORD datums, but that might
not always be so.  This also syncs several different places that read
INTO target lists.  They're just enough different that it seems
impractical to factor them into a single routine, but they surely
should be the same as much as possible.

18 years agoDone:
Bruce Momjian [Sun, 12 Feb 2006 04:44:17 +0000 (04:44 +0000)]
Done:

> * -Allow to_char() to print localized month names

18 years agoAllow to_char() to print localized month and day names.
Bruce Momjian [Sun, 12 Feb 2006 04:44:15 +0000 (04:44 +0000)]
Allow to_char() to print localized month and day names.

Euler Taveira de Oliveira

18 years agoAs part of previous commit, add pg_restore option:
Bruce Momjian [Sun, 12 Feb 2006 04:07:25 +0000 (04:07 +0000)]
As part of previous commit, add pg_restore option:

        -1 or --single-transaction

Simon Riggs

18 years agoAdd psql option:
Bruce Momjian [Sun, 12 Feb 2006 04:04:32 +0000 (04:04 +0000)]
Add psql option:

-1 or --single-transaction

Simon Riggs

18 years agoDone:
Bruce Momjian [Sun, 12 Feb 2006 03:56:13 +0000 (03:56 +0000)]
Done:

> * -Add system view to show free space map contents

18 years agoAdd contrib/pg_freespacemap to display free space map information.
Bruce Momjian [Sun, 12 Feb 2006 03:55:53 +0000 (03:55 +0000)]
Add contrib/pg_freespacemap to display free space map information.

Mark Kirkwood

18 years agoFix more fallout from line-wrap patch, to wit, arbitrarily changing
Tom Lane [Sun, 12 Feb 2006 03:30:21 +0000 (03:30 +0000)]
Fix more fallout from line-wrap patch, to wit, arbitrarily changing
the API of PQdsplen without bothering to fix its callers.  Although
ReportSyntaxErrorPosition could probably do with more smarts about
handling control characters, for the moment I'll just get it back to
handling tabs consistently.

18 years agoI've created a new shared catalog table pg_shdescription to store
Bruce Momjian [Sun, 12 Feb 2006 03:22:21 +0000 (03:22 +0000)]
I've created a new shared catalog table pg_shdescription to store
comments on cluster global objects like databases, tablespaces, and
roles.

It touches a lot of places, but not much in the way of big changes.  The
only design decision I made was to duplicate the query and manipulation
functions rather than to try and have them handle both shared and local
comments.  I believe this is simpler for the code and not an issue for
callers because they know what type of object they are dealing with.
This has resulted in a shobj_description function analagous to
obj_description and backend functions [Create/Delete]SharedComments
mirroring the existing [Create/Delete]Comments functions.

pg_shdescription.h goes into src/include/catalog/

Kris Jurka

18 years agoUndo changes of trailing space in recently-committed expected files.
Tom Lane [Sun, 12 Feb 2006 03:10:04 +0000 (03:10 +0000)]
Undo changes of trailing space in recently-committed expected files.
This is mostly just over-compulsiveness on my part, but the exercise
did reveal one real bug: errors.out has a space difference now where
it should not.

18 years agoRestore previous psql behavior of not printing useless trailing spaces
Tom Lane [Sun, 12 Feb 2006 02:56:21 +0000 (02:56 +0000)]
Restore previous psql behavior of not printing useless trailing spaces
after the data in the last column on a line.

18 years agoPlease find enclosed a patch that lets you use \c to connect
Bruce Momjian [Sun, 12 Feb 2006 02:54:30 +0000 (02:54 +0000)]
Please find enclosed a patch that lets you use \c to connect
(optionally) to a new host and port without exiting psql.  This
eliminates, IMHO, a surprise in that you can now connect to PostgreSQL
on a differnt machine from the one where you started your session. This
should help people who use psql as an administrative tool.

David Fetter

18 years agoActually there's a better way to do this, which is to count tuples
Tom Lane [Sun, 12 Feb 2006 00:18:17 +0000 (00:18 +0000)]
Actually there's a better way to do this, which is to count tuples
during the vacuumcleanup scan that we're going to do anyway.  Should
save a few cycles (one calculation per page, not per tuple) as well
as not having to depend on assumptions about heap and index being
in step.
I think this could probably be made to work for GIST too, but that
code looks messy enough that I'm disinclined to try right now.

18 years agoSkip ambulkdelete scan if there's nothing to delete and the index is not
Tom Lane [Sat, 11 Feb 2006 23:31:34 +0000 (23:31 +0000)]
Skip ambulkdelete scan if there's nothing to delete and the index is not
partial.  None of the existing AMs do anything useful except counting
tuples when there's nothing to delete, and we can get a tuple count
from the heap as long as it's not a partial index.  (hash actually can
skip anyway because it maintains a tuple count in the index metapage.)
GIST is not currently able to exploit this optimization because, due to
failure to index NULLs, GIST is always effectively partial.  Possibly
we should fix that sometime.
Simon Riggs w/ some review by Tom Lane.

18 years agoAllow ALTER TABLE ... ALTER CONSTRAINT ... RENAME
Bruce Momjian [Sat, 11 Feb 2006 22:17:19 +0000 (22:17 +0000)]
Allow ALTER TABLE ... ALTER CONSTRAINT ... RENAME

Joachim Wieland

18 years agoDone:
Bruce Momjian [Sat, 11 Feb 2006 22:17:17 +0000 (22:17 +0000)]
Done:
>  o -Allow ALTER TABLE ... ALTER CONSTRAINT ... RENAME

18 years agoDone:
Bruce Momjian [Sat, 11 Feb 2006 21:55:49 +0000 (21:55 +0000)]
Done:

<  o Improve psql's handling of multi-line statements
>  o -Improve psql's handling of multi-line statements

18 years agoo Improve psql's handling of multi-line statements
Bruce Momjian [Sat, 11 Feb 2006 21:55:35 +0000 (21:55 +0000)]
o Improve psql's handling of multi-line statements

Currently, while \e saves a single statement as one entry, interactive
statements are saved one line at a time.  Ideally all statements
would be saved like \e does.

Sergey E. Koposov

18 years agoFix incorrect addition, subtraction, and overflow checking in new
Tom Lane [Sat, 11 Feb 2006 20:39:59 +0000 (20:39 +0000)]
Fix incorrect addition, subtraction, and overflow checking in new
inet operators.

18 years agoRevert based on Tom's recommendation:
Bruce Momjian [Sat, 11 Feb 2006 17:14:09 +0000 (17:14 +0000)]
Revert based on Tom's recommendation:

> Allow VACUUM to complete faster by avoiding scanning the indexes when no
> rows were removed from the heap by the VACUUM.

18 years agoAllow VACUUM to complete faster by avoiding scanning the indexes when no
Bruce Momjian [Sat, 11 Feb 2006 16:59:09 +0000 (16:59 +0000)]
Allow VACUUM to complete faster by avoiding scanning the indexes when no
rows were removed from the heap by the VACUUM.

Simon Riggs

18 years agoUpdate catalog version for INET additions.
Bruce Momjian [Sat, 11 Feb 2006 16:28:56 +0000 (16:28 +0000)]
Update catalog version for INET additions.

18 years agoDone:
Bruce Momjian [Sat, 11 Feb 2006 03:33:36 +0000 (03:33 +0000)]
Done:

> * -Allow INET + INT8 to increment the host part of the address or

18 years agoAdd INET/CIDR operators: and, or, not, plus int8, minus int8, and inet
Bruce Momjian [Sat, 11 Feb 2006 03:32:41 +0000 (03:32 +0000)]
Add INET/CIDR operators: and, or, not, plus int8, minus int8, and inet
minus inet.

Stephen R. van den Berg

18 years agoClean up signedness warnings and 64-bit bugs in recent psql printing
Tom Lane [Fri, 10 Feb 2006 22:29:06 +0000 (22:29 +0000)]
Clean up signedness warnings and 64-bit bugs in recent psql printing
patch.  Martijn van Oosterhout and Tom Lane

18 years agoMake it possible to run initdb from an admin account on Windows,
Tom Lane [Fri, 10 Feb 2006 22:05:42 +0000 (22:05 +0000)]
Make it possible to run initdb from an admin account on Windows,
by giving up admin privileges (only works if newer than NT4).

Magnus

18 years agoEnable pg_ctl to give up admin privileges when starting the server under
Tom Lane [Fri, 10 Feb 2006 22:00:59 +0000 (22:00 +0000)]
Enable pg_ctl to give up admin privileges when starting the server under
Windows (if newer than NT4, else works same as before).

Magnus

18 years agoCheck that SID is enabled while checking for Windows admin privileges.
Tom Lane [Fri, 10 Feb 2006 21:52:21 +0000 (21:52 +0000)]
Check that SID is enabled while checking for Windows admin privileges.
Magnus

18 years agoChange search for default operator classes so that it examines all opclasses
Tom Lane [Fri, 10 Feb 2006 19:01:12 +0000 (19:01 +0000)]
Change search for default operator classes so that it examines all opclasses
regardless of the current schema search path.  Since CREATE OPERATOR CLASS
only allows one default opclass per datatype regardless of schemas, this
should have minimal impact, and it fixes problems with failure to find a
desired opclass while restoring dump files.  Per discussion at
http://archives.postgresql.org/pgsql-hackers/2006-02/msg00284.php.
Remove now-redundant-or-unused code in typcache.c and namespace.c,
and backpatch as far as 8.0.

18 years agoClean up some signedness warnings.
Tom Lane [Fri, 10 Feb 2006 15:57:58 +0000 (15:57 +0000)]
Clean up some signedness warnings.

18 years agoFix psql after newline patch for calloc(0).
Bruce Momjian [Fri, 10 Feb 2006 15:48:05 +0000 (15:48 +0000)]
Fix psql after newline patch for calloc(0).

Kris Jurka

18 years agoFix fallout from psql line-wrapping patch.
Tom Lane [Fri, 10 Feb 2006 15:47:44 +0000 (15:47 +0000)]
Fix fallout from psql line-wrapping patch.

18 years agoAllow "'" symbol in affixes ("'s" affix in english): it was diallowed during
Teodor Sigaev [Fri, 10 Feb 2006 12:56:14 +0000 (12:56 +0000)]
Allow "'" symbol in affixes ("'s" affix in english): it was diallowed during
multibyte support work.
Add line number to error output during affix file parsing.

18 years agoFix up remaining library checks.
Peter Eisentraut [Fri, 10 Feb 2006 11:35:47 +0000 (11:35 +0000)]
Fix up remaining library checks.

18 years agorenew output of regression test accordingly to
Teodor Sigaev [Fri, 10 Feb 2006 11:18:40 +0000 (11:18 +0000)]
renew output of regression test accordingly to
http://archives.postgresql.org/pgsql-committers/2006-02/msg00089.php

18 years agoDone:
Bruce Momjian [Fri, 10 Feb 2006 04:14:17 +0000 (04:14 +0000)]
Done:

<  o Allow multi-line column values to align in the proper columns
>  o -Allow multi-line column values to align in the proper columns

18 years agoAllow psql multi-line column values to align in the proper columns
Bruce Momjian [Fri, 10 Feb 2006 00:39:04 +0000 (00:39 +0000)]
Allow psql multi-line column values to align in the proper columns

  If the second output column value is 'a\nb', the 'b' should appear
  in the second display column, rather than the first column as it
  does now.

Change libpq's PQdsplen() to return more useful values.

> Note: this changes the PQdsplen function, it can now return zero or
> minus one which was not possible before. It doesn't appear anyone is
> actually using the functions other than psql but it is a change. The
> functions are not actually documentated anywhere so it's not like we're
> breaking a defined interface. The new semantics follow the Unicode
> standard.

BACKWARD COMPATIBLE CHANGE.

The only user-visible change I saw in the regression tests is that a
SELECT * on a table where all the columns have been dropped doesn't
return a blank line like before.  This seems like a step forward.

Martijn van Oosterhout

18 years agoThis fixes pg_dump so that when using the '-O' no owners option it does
Bruce Momjian [Thu, 9 Feb 2006 20:52:13 +0000 (20:52 +0000)]
This fixes pg_dump so that when using the '-O' no owners option it does
not print the owner name in the object comment.

eg:

--
-- Name: actor; Type: TABLE; Schema: public; Owner: chriskl; Tablespace:
--

Becomes:

--
-- Name: actor; Type: TABLE; Schema: public; Owner: -; Tablespace:
--

This makes it far easier to do 'user independent' dumps.  Especially for
distribution to third parties.

Christopher Kings-Lynne

18 years agoProvide the libpq error message when PQputline or PQendcopy fails.
Tom Lane [Thu, 9 Feb 2006 18:28:29 +0000 (18:28 +0000)]
Provide the libpq error message when PQputline or PQendcopy fails.

18 years ago1 Fix bug with very short word: prefix and suffix might be overlapped,
Teodor Sigaev [Thu, 9 Feb 2006 18:04:20 +0000 (18:04 +0000)]
1 Fix bug with very short word: prefix and suffix might be overlapped,
  sorry but fix can't be applyed to previous version: it's require
  refill tsvector...
2 Small optimize of load time for huge dictionaries
3 use palloc instead of malloc during load dict file

18 years agoUse "bitwise" rather than "binary for AND/OR descriptions in \d, to
Bruce Momjian [Thu, 9 Feb 2006 14:53:51 +0000 (14:53 +0000)]
Use "bitwise" rather than "binary for AND/OR descriptions in \d, to
match SGML documentation.

18 years agoC++ comments are verboten. Per gripe from Kris Jurka.
Tom Lane [Thu, 9 Feb 2006 07:22:22 +0000 (07:22 +0000)]
C++ comments are verboten.  Per gripe from Kris Jurka.

18 years agoReject out-of-range dates in date_in().
Tom Lane [Thu, 9 Feb 2006 03:39:17 +0000 (03:39 +0000)]
Reject out-of-range dates in date_in().

Kris Jurka

18 years agoFix broken non-YYDEBUG case.
Tom Lane [Thu, 9 Feb 2006 00:56:57 +0000 (00:56 +0000)]
Fix broken non-YYDEBUG case.

18 years ago Added just another test case.
Michael Meskes [Wed, 8 Feb 2006 09:10:05 +0000 (09:10 +0000)]
 Added just another test case.
 Fixed missing continuation line character.
 Do not translate $-quoting.
 Bit field notation belongs to a variable not a variable list.
 Output of line number only done by one function.

18 years agoRemove ancient hack to work around a peculiarity of libcurses on HPUX.
Tom Lane [Tue, 7 Feb 2006 17:36:13 +0000 (17:36 +0000)]
Remove ancient hack to work around a peculiarity of libcurses on HPUX.
Since we now use libtermcap in preference to libcurses, no need for hack
anymore.

18 years agoSource code cleanup.
Bruce Momjian [Tue, 7 Feb 2006 17:04:04 +0000 (17:04 +0000)]
Source code cleanup.

18 years agoI think that NUMERIC datatype has a problem in the performance that
Bruce Momjian [Tue, 7 Feb 2006 16:03:50 +0000 (16:03 +0000)]
I think that NUMERIC datatype has a problem in the performance that
the format on Tuple(Numeric) and the format to calculate(NumericVar)
are different. I understood that to reduce I/O. However, when many
comparisons or calculations of NUMERIC are executed, the conversion
of Numeric and NumericVar becomes a bottleneck.

It is profile result when "create index on NUMERIC column" is executed:

  %   cumulative   self              self     total
 time   seconds   seconds    calls   s/call   s/call  name
 17.61     10.27    10.27 34542006     0.00     0.00  cmp_numerics
 11.90     17.21     6.94 34542006     0.00     0.00  comparetup_index
  7.42     21.54     4.33 71102587     0.00     0.00  AllocSetAlloc
  7.02     25.64     4.09 69084012     0.00     0.00  set_var_from_num
  4.87     28.48     2.84 69084012     0.00     0.00  alloc_var
  4.79     31.27     2.79 142205745     0.00     0.00  AllocSetFreeIndex
  4.55     33.92     2.65 34542004     0.00     0.00  cmp_abs
  4.07     36.30     2.38 71101189     0.00     0.00  AllocSetFree
  3.83     38.53     2.23 69084012     0.00     0.00  free_var

The create index command executes many comparisons of Numeric values.
Functions other than comparetup_index spent a lot of cycles for
conversion from Numeric to NumericVar.

An attached patch enables the comparison of Numeric values without
executing conversion to NumericVar. The execution time of that SQL
becomes half.

o Test SQL (index_test table has 1,000,000 tuples)
 create index index_test_idx on index_test(num_col);

o Test results (executed the test five times)
(1)PentiumIII
 original: 39.789s  36.823s  36.737s  37.752s  37.019s
 patched : 18.560s  19.103s  18.830s  18.408s  18.853s
  4.07     36.30     2.38 71101189     0.00     0.00  AllocSetFree
  3.83     38.53     2.23 69084012     0.00     0.00  free_var

The create index command executes many comparisons of Numeric values.
Functions other than comparetup_index spent a lot of cycles for
conversion from Numeric to NumericVar.

An attached patch enables the comparison of Numeric values without
executing conversion to NumericVar. The execution time of that SQL
becomes half.

o Test SQL (index_test table has 1,000,000 tuples)
 create index index_test_idx on index_test(num_col);

o Test results (executed the test five times)
(1)PentiumIII
 original: 39.789s  36.823s  36.737s  37.752s  37.019s
 patched : 18.560s  19.103s  18.830s  18.408s  18.853s

(2)Pentium4
 original: 16.349s  14.997s  12.979s  13.169s  12.955s
 patched :  7.005s   6.594s   6.770s   6.740s   6.828s

(3)Itanium2
 original: 15.392s  15.447s  15.350s  15.370s  15.417s
 patched :  7.413s   7.330s   7.334s   7.339s   7.339s

(4)Ultra Sparc
 original: 64.435s  59.336s  59.332s  58.455s  59.781s
 patched : 28.630s  28.666s  28.983s  28.744s  28.595s

Atsushi Ogawa

18 years agoRemove question mark:
Bruce Momjian [Tue, 7 Feb 2006 14:49:17 +0000 (14:49 +0000)]
Remove question mark:

< * %Disallow changing default expression of a SERIAL column?
> * %Disallow changing default expression of a SERIAL column

18 years agoMake --version option print fixed program name.
Peter Eisentraut [Tue, 7 Feb 2006 11:36:36 +0000 (11:36 +0000)]
Make --version option print fixed program name.

18 years agoRemove some checks for libraries that no one can identify. We'll see how
Peter Eisentraut [Tue, 7 Feb 2006 11:22:41 +0000 (11:22 +0000)]
Remove some checks for libraries that no one can identify.  We'll see how
that works out...

18 years agoSplit up wal-logging items:
Bruce Momjian [Tue, 7 Feb 2006 02:08:08 +0000 (02:08 +0000)]
Split up wal-logging items:

< * Allow control over which tables are WAL-logged [walcontrol]
> * Allow WAL logging to be turned off for a table, but the table
>   might be dropped or truncated during crash recovery [walcontrol]
<   commit.  To do this, only a single writer can modify the table, and
<   writes must happen only on new pages.  Readers can continue accessing
<   the table.  This would affect COPY, and perhaps INSERT/UPDATE too.
<   Another option is to avoid transaction logging entirely and truncate
<   or drop the table on crash recovery.  These should be implemented
<   using ALTER TABLE, e.g. ALTER TABLE PERSISTENCE [ DROP | TRUNCATE |
<   STABLE | DEFAULT ].  Tables using non-default logging should not use
<   referential integrity with default-logging tables, and tables using
<   stable logging probably can not have indexes.  One complexity is
<   the handling of indexes on TOAST tables.
>   commit.  This should be implemented using ALTER TABLE, e.g. ALTER
>   TABLE PERSISTENCE [ DROP | TRUNCATE | DEFAULT ].  Tables using
>   non-default logging should not use referential integrity with
>   default-logging tables.  A table without dirty buffers during a
>   crash could perhaps avoid the drop/truncate.
>
> * Allow WAL logging to be turned off for a table, but the table would
>   avoid being truncated/dropped [walcontrol]
>
>   To do this, only a single writer can modify the table, and writes
>   must happen only on new pages so the new pages can be removed during
>   crash recovery.  Readers can continue accessing the table.  Such
>   tables probably cannot have indexes.  One complexity is the handling
>   of indexes on TOAST tables.

18 years agoFix HTML alignment in PQprint.
Bruce Momjian [Tue, 7 Feb 2006 00:26:15 +0000 (00:26 +0000)]
Fix HTML alignment in PQprint.

Christoph Zwerschke

18 years agoImprove the tests to see if ScalarArrayOpExpr is strict. Original coding
Tom Lane [Mon, 6 Feb 2006 22:21:12 +0000 (22:21 +0000)]
Improve the tests to see if ScalarArrayOpExpr is strict.  Original coding
would basically punt in all cases for 'foo <> ALL (array)', which resulted
in a performance regression for NOT IN compared to what we were doing in
8.1 and before.  Per report from Pavel Stehule.

18 years agoCheck number of affixes to prevent core dump with zero number of affixes
Teodor Sigaev [Mon, 6 Feb 2006 15:45:34 +0000 (15:45 +0000)]
Check number of affixes to prevent core dump with zero number of affixes

18 years agoFix PQprint HTML tag, "centre" -> "center".
Bruce Momjian [Mon, 6 Feb 2006 02:23:07 +0000 (02:23 +0000)]
Fix PQprint HTML tag, "centre" -> "center".

18 years agoFix pg_restore to properly discard COPY data when trying to continue
Tom Lane [Sun, 5 Feb 2006 20:58:47 +0000 (20:58 +0000)]
Fix pg_restore to properly discard COPY data when trying to continue
after an error in a COPY statement.  Formerly it thought the COPY data
was SQL commands, and got quite confused.

Stephen Frost

18 years agoFix typo in configuration docs.
Bruce Momjian [Sun, 5 Feb 2006 18:19:14 +0000 (18:19 +0000)]
Fix typo in configuration docs.

Devrim GUNDUZ

18 years agoImprove my initial, rather hacky implementation of joins to append
Tom Lane [Sun, 5 Feb 2006 02:59:17 +0000 (02:59 +0000)]
Improve my initial, rather hacky implementation of joins to append
relations: fix the executor so that we can have an Append plan on the
inside of a nestloop and still pass down outer index keys to index scans
within the Append, then generate such plans as if they were regular
inner indexscans.  This avoids the need to evaluate the outer relation
multiple times.

18 years agoUpdate PL/pgSQL trigger example to be clearer about how to "merge" data
Bruce Momjian [Sun, 5 Feb 2006 02:47:53 +0000 (02:47 +0000)]
Update PL/pgSQL trigger example to be clearer about how to "merge" data
into a table.

Jim C. Nasby

18 years agoFix constraint exclusion to work in inherited UPDATE/DELETE queries
Tom Lane [Sat, 4 Feb 2006 23:03:20 +0000 (23:03 +0000)]
Fix constraint exclusion to work in inherited UPDATE/DELETE queries
... in fact, it will be applied now in any query whatsoever.  I'm still
a bit concerned about the cycles that might be expended in failed proof
attempts, but given that CE is turned off by default, it's the user's
choice whether to expend those cycles or not.  (Possibly we should
change the simple bool constraint_exclusion parameter to something
more fine-grained?)

18 years agoFix broken markup.
Tom Lane [Sat, 4 Feb 2006 22:38:39 +0000 (22:38 +0000)]
Fix broken markup.

18 years agoAdded C bit fields to ecpg parser
Michael Meskes [Sat, 4 Feb 2006 20:54:44 +0000 (20:54 +0000)]
Added C bit fields to ecpg parser
Added some default rules to lexer
Added log output to prepare statement
Added some more stuff to a test case

18 years agoDROP IF EXISTS for ROLE/USER/GROUP
Andrew Dunstan [Sat, 4 Feb 2006 19:06:47 +0000 (19:06 +0000)]
DROP IF EXISTS for ROLE/USER/GROUP

18 years agoIssue a warning if a change-on-restart-only postgresql.conf value is
Peter Eisentraut [Sat, 4 Feb 2006 12:50:47 +0000 (12:50 +0000)]
Issue a warning if a change-on-restart-only postgresql.conf value is
modified  and the server config files are reloaded

18 years agoUpdate walcontrol item:
Bruce Momjian [Sat, 4 Feb 2006 03:23:21 +0000 (03:23 +0000)]
Update walcontrol item:

< * Allow control over which tables are WAL-logged
> * Allow control over which tables are WAL-logged [walcontrol]
1038c1038,1039
<   stable logging probably can not have indexes.  [walcontrol]
>   stable logging probably can not have indexes.  One complexity is
>   the handling of indexes on TOAST tables.

18 years agoAdd to TODO.detail/walcontrol.
Bruce Momjian [Sat, 4 Feb 2006 03:22:36 +0000 (03:22 +0000)]
Add to TODO.detail/walcontrol.

18 years agoIn ecpg, automatically double single quotes in $$ strings because
Bruce Momjian [Sat, 4 Feb 2006 02:32:38 +0000 (02:32 +0000)]
In ecpg, automatically double single quotes in $$ strings because
internally $$ strings are converted to single-quote strings.

In ecpg, output newlines in commands using standard C escapes, rather
than using literal newlines, which is not portable.

18 years agoUpdate Makefile for new thread_test location.
Bruce Momjian [Sat, 4 Feb 2006 01:04:20 +0000 (01:04 +0000)]
Update Makefile for new thread_test location.

18 years agoMove thread_test directory from /tools to /test so source-only tarballs
Bruce Momjian [Sat, 4 Feb 2006 01:00:02 +0000 (01:00 +0000)]
Move thread_test directory from /tools to /test so source-only tarballs
have the directory for the configure test.

18 years agoFirst attempt at removing some AC_CHECK_LIB(foo, main) calls.
Peter Eisentraut [Sat, 4 Feb 2006 00:42:54 +0000 (00:42 +0000)]
First attempt at removing some AC_CHECK_LIB(foo, main) calls.

18 years agoTeach planner to convert simple UNION ALL subqueries into append relations,
Tom Lane [Fri, 3 Feb 2006 21:08:49 +0000 (21:08 +0000)]
Teach planner to convert simple UNION ALL subqueries into append relations,
thereby sharing code with the inheritance case.  This puts the UNION-ALL-view
approach to partitioned tables on par with inheritance, so far as constraint
exclusion is concerned: it works either way.  (Still need to update the docs
to say so.)  The definition of "simple UNION ALL" is a little simpler than
I would like --- basically the union arms can only be SELECT * FROM foo
--- but it's good enough for partitioned-table cases.

18 years agoUpdate tested AIX memset platforms.
Bruce Momjian [Fri, 3 Feb 2006 13:56:20 +0000 (13:56 +0000)]
Update tested AIX memset platforms.

18 years agoAllow MEMSET_LOOP_LIMIT to be set on a per-platform basis, and turn off
Bruce Momjian [Fri, 3 Feb 2006 13:53:15 +0000 (13:53 +0000)]
Allow MEMSET_LOOP_LIMIT to be set on a per-platform basis, and turn off
MemSet on AIX by setting MEMSET_LOOP_LIMIT to zero.

Add optimization to skip MemSet tests in MEMSET_LOOP_LIMIT == 0 case and
just call memset() directly.

18 years agoUpdate random() usage so ranges are inclusive/exclusive as required.
Bruce Momjian [Fri, 3 Feb 2006 12:45:47 +0000 (12:45 +0000)]
Update random() usage so ranges are inclusive/exclusive as required.

18 years agoPrevent COPY from using newline or carriage return as delimiter or null.
Bruce Momjian [Fri, 3 Feb 2006 12:41:07 +0000 (12:41 +0000)]
Prevent COPY from using newline or carriage return as delimiter or null.
Disallow backslash as the delimiter in non-CVS mode.

David Fetter

18 years agoImprove tab whitespace in file.
Bruce Momjian [Fri, 3 Feb 2006 05:38:35 +0000 (05:38 +0000)]
Improve tab whitespace in file.

Add comment about $$ and '' SCONST strings.

18 years agoMore spacing cleanups.
Bruce Momjian [Thu, 2 Feb 2006 03:51:41 +0000 (03:51 +0000)]
More spacing cleanups.

18 years agoSource code alignment fixes for preproc.y.
Bruce Momjian [Wed, 1 Feb 2006 22:16:36 +0000 (22:16 +0000)]
Source code alignment fixes for preproc.y.

18 years agoMore alignment improvements.
Bruce Momjian [Wed, 1 Feb 2006 20:57:39 +0000 (20:57 +0000)]
More alignment improvements.

18 years agoMake pgc.l source code alignment consistent.
Bruce Momjian [Wed, 1 Feb 2006 20:56:44 +0000 (20:56 +0000)]
Make pgc.l source code alignment consistent.

18 years agoAdd:
Bruce Momjian [Wed, 1 Feb 2006 17:32:45 +0000 (17:32 +0000)]
Add:

> * Allow statistics collector information to be pulled from the collector
>   process directly, rather than requiring the collector to write a
>   filesystem file twice a second?

18 years agoAdd code comment about Linux stack randomization and shared memory.
Bruce Momjian [Wed, 1 Feb 2006 16:00:06 +0000 (16:00 +0000)]
Add code comment about Linux stack randomization and shared memory.

18 years agoFix const cast in get_progname().
Bruce Momjian [Wed, 1 Feb 2006 12:41:45 +0000 (12:41 +0000)]
Fix const cast in get_progname().

Backpatch.

18 years agoSet progname early in the postmaster/postgres binary, rather than doing
Bruce Momjian [Wed, 1 Feb 2006 00:31:59 +0000 (00:31 +0000)]
Set progname early in the postmaster/postgres binary, rather than doing
it later.  This fixes a problem where EXEC_BACKEND didn't have progname
set, causing a segfault if log_min_messages was set below debug2 and our
own snprintf.c was being used.

Also alway strdup() progname.

Backpatch to 8.1.X and 8.0.X.

18 years agoMove items:
Bruce Momjian [Wed, 1 Feb 2006 00:07:26 +0000 (00:07 +0000)]
Move items:

> * Add SQL99 WITH clause to SELECT
> * Add SQL99 WITH RECURSIVE to SELECT
< * Add SQL99 WITH clause to SELECT
< * Add SQL99 WITH RECURSIVE to SELECT