OSDN Git Service

pg-rex/syncrep.git
17 years agoAdd:
Bruce Momjian [Sat, 3 Feb 2007 22:32:49 +0000 (22:32 +0000)]
Add:

>    o Allow recovery.conf to allow the same syntax as
>      postgresql.conf, including quoting
>
>      http://archives.postgresql.org/pgsql-hackers/2006-12/msg00497.php

17 years agoUpdate SQL conformance information about XML features.
Peter Eisentraut [Sat, 3 Feb 2007 17:59:36 +0000 (17:59 +0000)]
Update SQL conformance information about XML features.

17 years agoImplement XMLSERIALIZE for real. Analogously, make the xml to text cast
Peter Eisentraut [Sat, 3 Feb 2007 14:06:56 +0000 (14:06 +0000)]
Implement XMLSERIALIZE for real.  Analogously, make the xml to text cast
observe the xmloption.

Reorganize the representation of the XML option in the parse tree and the
API to make it easier to manage and understand.

Add regression tests for parsing back XML expressions.

17 years agoFix configure detection code when --with-ldap and --enable-thread-safety
Bruce Momjian [Sat, 3 Feb 2007 02:43:38 +0000 (02:43 +0000)]
Fix configure detection code when --with-ldap and --enable-thread-safety
are both used.

Albe Laurenz

17 years agoAdd patches thread for:
Bruce Momjian [Sat, 3 Feb 2007 01:59:27 +0000 (01:59 +0000)]
Add patches thread for:

* Reduce checkpoint performance degredation by forcing data to disk
  more evenly

>   http://archives.postgresql.org/pgsql-patches/2006-12/msg00104.php

17 years agoAdd URL for:
Bruce Momjian [Fri, 2 Feb 2007 23:05:36 +0000 (23:05 +0000)]
Add URL for:

* Allow sequential scans to take advantage of other concurrent
  sequential scans, also called "Synchronised Scanning"
>
>   http://archives.postgresql.org/pgsql-hackers/2006-12/msg00784.php

17 years agoAdd:
Bruce Momjian [Fri, 2 Feb 2007 22:55:08 +0000 (22:55 +0000)]
Add:

> * Reduce checkpoint performance degredation by forcing data to disk
>   more evenly
>
>   http://archives.postgresql.org/pgsql-hackers/2006-12/msg00337.php
>   http://archives.postgresql.org/pgsql-hackers/2007-01/msg00079.php

17 years agoAdd URL for:
Bruce Momjian [Fri, 2 Feb 2007 21:34:39 +0000 (21:34 +0000)]
Add URL for:

                o Fix RENAME to work on variables other than OLD/NEW
>    http://archives.postgresql.org/pgsql-hackers/2007-01/msg01587.php

17 years agoThis patch changes the installscript for vcbuild to actually parse the
Neil Conway [Fri, 2 Feb 2007 16:25:34 +0000 (16:25 +0000)]
This patch changes the installscript for vcbuild to actually parse the
generated solution files for what to install, instead of blindly copying
everything as it previously did. With the previous quick-n-dirty
version, it would copy old DLLs if you reconfigured in a way that didn't
include subprojects like a PL for example.

Magnus Hagander.

17 years agoReword suggestion that libpq.dll be installed in WINNT\SYSTEM32 under
Neil Conway [Fri, 2 Feb 2007 16:10:08 +0000 (16:10 +0000)]
Reword suggestion that libpq.dll be installed in WINNT\SYSTEM32 under
Windows. Per Magnus Hagander, this is not recommended.

17 years agoApplied Magnus Hagander's patch to take away some compiler warnings.
Michael Meskes [Fri, 2 Feb 2007 09:31:10 +0000 (09:31 +0000)]
Applied Magnus Hagander's patch to take away some compiler warnings.

17 years agoCleaned up va_list handling. Hopefully this now works on all archs.
Michael Meskes [Fri, 2 Feb 2007 08:58:23 +0000 (08:58 +0000)]
Cleaned up va_list handling. Hopefully this now works on all archs.

17 years agoAdd:
Bruce Momjian [Fri, 2 Feb 2007 05:42:56 +0000 (05:42 +0000)]
Add:

>  o Allow column display reordering by recording a display,
>    storage, and permanent id for every column?
>
>    http://archives.postgresql.org/pgsql-hackers/2006-12/msg00782.php
>

17 years agoUpdate release notes for security-related releases in all active branches.
Tom Lane [Fri, 2 Feb 2007 00:10:23 +0000 (00:10 +0000)]
Update release notes for security-related releases in all active branches.

Security: CVE-2007-0555, CVE-2007-0556

17 years agoRepair failure to check that a table is still compatible with a previously
Tom Lane [Fri, 2 Feb 2007 00:07:03 +0000 (00:07 +0000)]
Repair failure to check that a table is still compatible with a previously
made query plan.  Use of ALTER COLUMN TYPE creates a hazard for cached
query plans: they could contain Vars that claim a column has a different
type than it now has.  Fix this by checking during plan startup that Vars
at relation scan level match the current relation tuple descriptor.  Since
at that point we already have at least AccessShareLock, we can be sure the
column type will not change underneath us later in the query.  However,
since a backend's locks do not conflict against itself, there is still a
hole for an attacker to exploit: he could try to execute ALTER COLUMN TYPE
while a query is in progress in the current backend.  Seal that hole by
rejecting ALTER TABLE whenever the target relation is already open in
the current backend.

This is a significant security hole: not only can one trivially crash the
backend, but with appropriate misuse of pass-by-reference datatypes it is
possible to read out arbitrary locations in the server process's memory,
which could allow retrieving database content the user should not be able
to see.  Our thanks to Jeff Trout for the initial report.

Security: CVE-2007-0556

17 years agoRepair insufficiently careful type checking for SQL-language functions:
Tom Lane [Fri, 2 Feb 2007 00:02:55 +0000 (00:02 +0000)]
Repair insufficiently careful type checking for SQL-language functions:
we should check that the function code returns the claimed result datatype
every time we parse the function for execution.  Formerly, for simple
scalar result types we assumed the creation-time check was sufficient, but
this fails if the function selects from a table that's been redefined since
then, and even more obviously fails if check_function_bodies had been OFF.

This is a significant security hole: not only can one trivially crash the
backend, but with appropriate misuse of pass-by-reference datatypes it is
possible to read out arbitrary locations in the server process's memory,
which could allow retrieving database content the user should not be able
to see.  Our thanks to Jeff Trout for the initial report.

Security: CVE-2007-0555

17 years agoFix broken markup.
Tom Lane [Thu, 1 Feb 2007 22:06:14 +0000 (22:06 +0000)]
Fix broken markup.

17 years agoAdd "May/Can/Might" section to error message style guidlines, and
Bruce Momjian [Thu, 1 Feb 2007 21:28:34 +0000 (21:28 +0000)]
Add "May/Can/Might" section to error message style guidlines, and
"can't" -> "cannot" section.

17 years agoFix two more regression tests whose expected outputs were not updated
Neil Conway [Thu, 1 Feb 2007 21:05:29 +0000 (21:05 +0000)]
Fix two more regression tests whose expected outputs were not updated
for the recent may/might cleanup.

17 years agoMention file system replication as a high availability solution in the
Bruce Momjian [Thu, 1 Feb 2007 21:02:48 +0000 (21:02 +0000)]
Mention file system replication as a high availability solution in the
shared hardware section, and mention DRBD as a popular solution.

17 years agoTypo fix.
Tom Lane [Thu, 1 Feb 2007 20:28:08 +0000 (20:28 +0000)]
Typo fix.

17 years agoUpdate some of the "expected" regression test results for Bruce's
Neil Conway [Thu, 1 Feb 2007 20:11:18 +0000 (20:11 +0000)]
Update some of the "expected" regression test results for Bruce's
recent may/might cleanup, in the hopes that this will unbreak the
buildfarm. Per report from Stefan Kaltenbrunner.

17 years agoFix plpgsql so that when a local variable has no initial-value expression,
Tom Lane [Thu, 1 Feb 2007 19:22:07 +0000 (19:22 +0000)]
Fix plpgsql so that when a local variable has no initial-value expression,
an error will be thrown correctly if the variable is of a NOT NULL domain.
Report and almost-correct fix from Sergiy Vyshnevetskiy (bug #2948).

17 years agoAdd URL for:
Bruce Momjian [Thu, 1 Feb 2007 19:13:56 +0000 (19:13 +0000)]
Add URL for:

        o Add long file support for binary pg_dump output
>
>    http://archives.postgresql.org/pgsql-hackers/2006-12/msg00551.php

17 years agoWording cleanup for error messages. Also change can't -> cannot.
Bruce Momjian [Thu, 1 Feb 2007 19:10:30 +0000 (19:10 +0000)]
Wording cleanup for error messages.  Also change can't -> cannot.

Standard English uses "may", "can", and "might" in different ways:

        may - permission, "You may borrow my rake."

        can - ability, "I can lift that log."

        might - possibility, "It might rain today."

Unfortunately, in conversational English, their use is often mixed, as
in, "You may use this variable to do X", when in fact, "can" is a better
choice.  Similarly, "It may crash" is better stated, "It might crash".

17 years agoFix broken markup.
Neil Conway [Thu, 1 Feb 2007 07:00:32 +0000 (07:00 +0000)]
Fix broken markup.

17 years agoAdd 8.2.0 "Incomatibilities" documentation that pg_dump's -n and -t
Bruce Momjian [Thu, 1 Feb 2007 05:12:24 +0000 (05:12 +0000)]
Add 8.2.0 "Incomatibilities" documentation that pg_dump's -n and -t
behavior has changed.

17 years agoChinese FAQ update
Bruce Momjian [Thu, 1 Feb 2007 04:55:53 +0000 (04:55 +0000)]
Chinese FAQ update

Daojing.Zhou

17 years agoUpdate URL for function permission checks
Bruce Momjian [Thu, 1 Feb 2007 04:51:31 +0000 (04:51 +0000)]
Update URL for function permission checks

<   http://archives.postgresql.org/pgsql-hackers/2006-12/msg00564.php
>   http://archives.postgresql.org/pgsql-hackers/2006-12/msg00568.php
>

17 years agoAdd:
Bruce Momjian [Thu, 1 Feb 2007 04:50:58 +0000 (04:50 +0000)]
Add:

>
> * Tighten function permission checks
>
>   http://archives.postgresql.org/pgsql-hackers/2006-12/msg00564.php
>

17 years agoAdd:
Bruce Momjian [Thu, 1 Feb 2007 04:49:43 +0000 (04:49 +0000)]
Add:

>
> * Tighten trigger permission checks
>
>   http://archives.postgresql.org/pgsql-hackers/2006-12/msg00564.php
>

17 years agoThis patch adds documentation for the long-version parameters --username
Neil Conway [Thu, 1 Feb 2007 04:39:33 +0000 (04:39 +0000)]
This patch adds documentation for the long-version parameters --username
and --password for pg_dump, pg_dumpall and pg_restore, per complaint by
Michael Schmidt. Patch from Magnus Hagander.

17 years agoAdd:
Bruce Momjian [Thu, 1 Feb 2007 04:35:52 +0000 (04:35 +0000)]
Add:
>
> * Fix problem when multiple subtransactions of the same outer transaction
>   hold different types of locks, and one subtransaction aborts
>
>   http://archives.postgresql.org/pgsql-hackers/2006-11/msg01011.php
>   http://archives.postgresql.org/pgsql-hackers/2006-12/msg00001.php

17 years agoFix a few typos in comments in GiN.
Neil Conway [Thu, 1 Feb 2007 04:16:08 +0000 (04:16 +0000)]
Fix a few typos in comments in GiN.

17 years agoUpdate CREATE SEQUENCE documentation to show the same sequence being
Bruce Momjian [Thu, 1 Feb 2007 00:34:03 +0000 (00:34 +0000)]
Update CREATE SEQUENCE documentation to show the same sequence being
created and increments.  The old docs created the sequence, then showed
a nextval() of 114.

17 years agoAdd URL for:
Bruce Momjian [Thu, 1 Feb 2007 00:32:45 +0000 (00:32 +0000)]
Add URL for:

      o Fix RENAME to work on variables other than OLD/NEW
>       http://archives.postgresql.org/pgsql-hackers/2007-01/msg01615.php

17 years agoConsistenly use colons before '<programlisting>' blocks, where
Bruce Momjian [Thu, 1 Feb 2007 00:28:19 +0000 (00:28 +0000)]
Consistenly use colons before '<programlisting>' blocks, where
appropriate.

17 years agoUpdate reference documentation on may/can/might:
Bruce Momjian [Wed, 31 Jan 2007 23:26:05 +0000 (23:26 +0000)]
Update reference documentation on may/can/might:

Standard English uses "may", "can", and "might" in different ways:

        may - permission, "You may borrow my rake."

        can - ability, "I can lift that log."

        might - possibility, "It might rain today."

Unfortunately, in conversational English, their use is often mixed, as
in, "You may use this variable to do X", when in fact, "can" is a better
choice.  Similarly, "It may crash" is better stated, "It might crash".

17 years agoUpdate comment.
Bruce Momjian [Wed, 31 Jan 2007 22:17:03 +0000 (22:17 +0000)]
Update comment.

17 years agoRevert error message change for may/can/might --- needs discussion.
Bruce Momjian [Wed, 31 Jan 2007 21:03:37 +0000 (21:03 +0000)]
Revert error message change for may/can/might --- needs discussion.

17 years agoUpdate documentation on may/can/might:
Bruce Momjian [Wed, 31 Jan 2007 20:56:20 +0000 (20:56 +0000)]
Update documentation on may/can/might:

Standard English uses "may", "can", and "might" in different ways:

        may - permission, "You may borrow my rake."

        can - ability, "I can lift that log."

        might - possibility, "It might rain today."

Unfortunately, in conversational English, their use is often mixed, as
in, "You may use this variable to do X", when in fact, "can" is a better
choice.  Similarly, "It may crash" is better stated, "It might crash".

Also update two error messages mentioned in the documenation to match.

17 years agoAdd missing Makefile dependencies for the "bigtest" and "bigcheck"
Neil Conway [Wed, 31 Jan 2007 19:38:50 +0000 (19:38 +0000)]
Add missing Makefile dependencies for the "bigtest" and "bigcheck"
targets, and refactor the tablespace setup code into a separate
Makefile target.

17 years agoRewrite uuid input and output routines to avoid dependency on the
Neil Conway [Wed, 31 Jan 2007 19:33:54 +0000 (19:33 +0000)]
Rewrite uuid input and output routines to avoid dependency on the
nonportable "hh" sprintf(3) length modifier. Instead, do the parsing
and output by hand. The code to do this isn't ideal, but this is
an interim measure anyway: the uuid type should probably use the
in-memory struct layout specified by RFC 4122. For now, this patch
should hopefully rectify the buildfarm failures for the uuid test.

Along the way, re-add pg_cast entries for uuid <-> varchar, which
I mistakenly removed earlier, and bump the catversion.

17 years agoFix initdb to not generate misleading error messages when postgres.bki
Tom Lane [Wed, 31 Jan 2007 18:52:49 +0000 (18:52 +0000)]
Fix initdb to not generate misleading error messages when postgres.bki
or other share-directory files are inaccessible for some reason other
than not existing.  Inspired by trouble report from Simon Kinsella.

17 years agoRevert gincostestimate changes.
Teodor Sigaev [Wed, 31 Jan 2007 16:54:51 +0000 (16:54 +0000)]
Revert gincostestimate changes.

17 years agoUpdate FreeBSD DocBook installation notices
Teodor Sigaev [Wed, 31 Jan 2007 15:23:28 +0000 (15:23 +0000)]
Update FreeBSD DocBook installation notices

17 years agoAllow GIN's extractQuery method to signal that nothing can satisfy the query.
Teodor Sigaev [Wed, 31 Jan 2007 15:09:45 +0000 (15:09 +0000)]
Allow GIN's extractQuery method to signal that nothing can satisfy the query.
In this case extractQuery should returns -1 as nentries. This changes
prototype of extractQuery method to use int32* instead of uint32* for
nentries argument.
Based on that gincostestimate may see two corner cases: nothing will be found
or seqscan should be used.

Per proposal at http://archives.postgresql.org/pgsql-hackers/2007-01/msg01581.php

PS tsearch_core patch should be sightly modified to support changes, but I'm
waiting a verdict about reviewing of tsearch_core patch.

17 years agoAdd URL:
Bruce Momjian [Wed, 31 Jan 2007 04:23:47 +0000 (04:23 +0000)]
Add URL:

      o Fix RENAME to work on variables other than OLD/NEW
>
>    http://archives.postgresql.org/pgsql-hackers/2002-03/msg00591.php
>

17 years agoDocument need for periodic REINDEX in VACUUM FULL cases.
Bruce Momjian [Wed, 31 Jan 2007 04:13:22 +0000 (04:13 +0000)]
Document need for periodic REINDEX in VACUUM FULL cases.

17 years agoAdd missing colon.
Bruce Momjian [Wed, 31 Jan 2007 04:12:01 +0000 (04:12 +0000)]
Add missing colon.

17 years agoAdd:
Bruce Momjian [Wed, 31 Jan 2007 03:17:49 +0000 (03:17 +0000)]
Add:

>
> * Add REINDEX CONCURRENTLY, like CREATE INDEX CONCURRENTLY
>
>   This is difficult because you must upgrade to an exclusive table lock
>   to replace the existing index file.  CREATE INDEX CONCURRENTLY does not
>   have this complication.  This would allow index compaction without
>   downtime.

17 years agoUpdate for VACUUM FULL:
Bruce Momjian [Wed, 31 Jan 2007 03:07:36 +0000 (03:07 +0000)]
Update for VACUUM FULL:

<   reindex rather than update the index.
>   reindex rather than update the index.  Also, index updates can
>   bloat the index.

17 years agoAdd:
Bruce Momjian [Tue, 30 Jan 2007 22:55:12 +0000 (22:55 +0000)]
Add:
>  o ARRAY[[1,2],[3,4]])[1] should return the same values as
>    ARRAY[[1,2],[3,4]])[1:1];
>

17 years agoAdd FAQ item about timezones and daylight savings time.
Bruce Momjian [Tue, 30 Jan 2007 22:37:15 +0000 (22:37 +0000)]
Add FAQ item about timezones and daylight savings time.

17 years agoUpdate documentation for backslashes to mention escape string syntax
Bruce Momjian [Tue, 30 Jan 2007 22:29:23 +0000 (22:29 +0000)]
Update documentation for backslashes to mention escape string syntax
more, and standard_conforming_strings less, because in the future non-E
strings will not treat backslashes specially.

Also use E'' strings where backslashes are used in examples. (The
existing examples would have drawn warnings.)

Backpatch to 8.2.X.

17 years agoRepair oversights in the mechanism used to store compiled plpgsql functions.
Tom Lane [Tue, 30 Jan 2007 22:05:13 +0000 (22:05 +0000)]
Repair oversights in the mechanism used to store compiled plpgsql functions.
The original coding failed (tried to access deallocated memory) if there were
two active call sites (fn_extra pointers) for the same function and the
function definition was updated.  Also, if an update of a recursive function
was detected upon nested entry to the function, the existing compiled version
was summarily deallocated, resulting in crash upon return to the outer
instance.  Problem observed while studying a bug report from Sergiy
Vyshnevetskiy.

Bug does not exist before 8.1 since older versions just leaked the memory of
obsoleted compiled functions, rather than trying to reclaim it.

17 years agoAdd SPI_push/SPI_pop calls so that datatype input and output functions called
Tom Lane [Tue, 30 Jan 2007 18:02:22 +0000 (18:02 +0000)]
Add SPI_push/SPI_pop calls so that datatype input and output functions called
by plpgsql can themselves use SPI --- possibly indirectly, as in the case
of domain_in() invoking plpgsql functions in a domain check constraint.
Per bug #2945 from Sergiy Vyshnevetskiy.

Somewhat arbitrarily, I've chosen to back-patch this as far as 8.0.  Given
the lack of prior complaints, it doesn't seem critical for 7.x.

17 years agoUpdate documentation for pg_get_serial_sequence() function.
Bruce Momjian [Tue, 30 Jan 2007 02:39:27 +0000 (02:39 +0000)]
Update documentation for pg_get_serial_sequence() function.

17 years agoAdd support for cross-type hashing in hash index searches and hash joins.
Tom Lane [Tue, 30 Jan 2007 01:33:36 +0000 (01:33 +0000)]
Add support for cross-type hashing in hash index searches and hash joins.
Hashing for aggregation purposes still needs work, so it's not time to
mark any cross-type operators as hashable for general use, but these cases
work if the operators are so marked by hand in the system catalogs.

17 years agoAdd comment noting that hashm_procid in a hash index's metapage isn't
Tom Lane [Mon, 29 Jan 2007 23:22:59 +0000 (23:22 +0000)]
Add comment noting that hashm_procid in a hash index's metapage isn't
actually used for anything.

17 years agoUpdate wording for minor release upgrades, per Alvaro.
Bruce Momjian [Mon, 29 Jan 2007 21:49:17 +0000 (21:49 +0000)]
Update wording for minor release upgrades, per Alvaro.

17 years agoDone:
Bruce Momjian [Mon, 29 Jan 2007 21:19:30 +0000 (21:19 +0000)]
Done:

< * Add Globally/Universally Unique Identifier (GUID/UUID)
> * -Add Globally/Universally Unique Identifier (GUID/UUID)

17 years agoUse sizeof() for snprintf() buffer length.
Bruce Momjian [Mon, 29 Jan 2007 20:22:35 +0000 (20:22 +0000)]
Use sizeof() for snprintf() buffer length.

17 years agoUpdate process termination message to display signal number and name
Bruce Momjian [Mon, 29 Jan 2007 20:17:40 +0000 (20:17 +0000)]
Update process termination message to display signal number and name
from exec.c and postmaster.c.

17 years agoUpdate documentation on XML functions
Peter Eisentraut [Mon, 29 Jan 2007 13:24:30 +0000 (13:24 +0000)]
Update documentation on XML functions

17 years agoImprove hash join to discard input tuples immediately if they can't
Tom Lane [Sun, 28 Jan 2007 23:21:26 +0000 (23:21 +0000)]
Improve hash join to discard input tuples immediately if they can't
match because they contain a null join key (and the join operator is
known strict).  Improves performance significantly when the inner
relation contains a lot of nulls, as per bug #2930.

17 years agoRemove unnecessary checkpoint from PL regression tests. This was once
Tom Lane [Sun, 28 Jan 2007 21:17:32 +0000 (21:17 +0000)]
Remove unnecessary checkpoint from PL regression tests.  This was once
handy to prevent core dump files from disappearing, but it's useless now
because (a) we don't drop core in individual DB subdirectories anymore,
and (b) CREATE DATABASE forces an internal checkpoint anyway.

17 years agoRename the uuid_t type to pg_uuid_t, to avoid a conflict with any
Neil Conway [Sun, 28 Jan 2007 20:25:38 +0000 (20:25 +0000)]
Rename the uuid_t type to pg_uuid_t, to avoid a conflict with any
definitions of uuid_t that may be provided by the system headers. This
should hopefully fix the Win32 build problems reported by Magnus.

17 years agoFix plpython MSVC build in non-debug mode.
Andrew Dunstan [Sun, 28 Jan 2007 19:36:46 +0000 (19:36 +0000)]
Fix plpython MSVC build in non-debug mode.

17 years agoRemove some unnecessary conversion work in build_regtype_array().
Tom Lane [Sun, 28 Jan 2007 19:05:35 +0000 (19:05 +0000)]
Remove some unnecessary conversion work in build_regtype_array().

17 years agoRepair oversight in creation of "append relations": we should set up
Tom Lane [Sun, 28 Jan 2007 18:50:40 +0000 (18:50 +0000)]
Repair oversight in creation of "append relations": we should set up
rel->tuples as well as rel->rows, since some estimation functions expect both
to be valid in every baserel.  Per report from Dave Dutcher.

17 years agoMake some small improvements in the accuracy of plpgsql's error location
Tom Lane [Sun, 28 Jan 2007 17:58:13 +0000 (17:58 +0000)]
Make some small improvements in the accuracy of plpgsql's error location
reports; inspired by the misleading CONTEXT lines shown in recent bug report
from Stefan Kaltenbrunner.  Also, allow statement-type names shown in these
messages to be translated.

17 years agoFix path problem in MSVC bison wrapper. per Joachim Wieland.
Andrew Dunstan [Sun, 28 Jan 2007 16:29:37 +0000 (16:29 +0000)]
Fix path problem in MSVC bison wrapper. per Joachim Wieland.

17 years agoAdd a new builtin type, "uuid". This implements a UUID type, similar to
Neil Conway [Sun, 28 Jan 2007 16:16:54 +0000 (16:16 +0000)]
Add a new builtin type, "uuid". This implements a UUID type, similar to
that defined in RFC 4122. This patch includes the basic implementation,
plus regression tests. Documentation and perhaps some additional
functionality will come later. Catversion bumped.

Patch from Gevik Babakhani; review from Peter, Tom, and myself.

17 years agoFix up plpgsql's "simple expression" evaluation mechanism so that it behaves
Tom Lane [Sun, 28 Jan 2007 16:15:49 +0000 (16:15 +0000)]
Fix up plpgsql's "simple expression" evaluation mechanism so that it behaves
safely in the presence of subtransactions.  To ensure that any ExprContext
shutdown callbacks are called at the right times, we have to have a separate
EState for each level of subtransaction.  Per "TupleDesc reference leak" bug
report from Stefan Kaltenbrunner.

Although I'm convinced the code is wrong as far back as 8.0, it doesn't seem
that there are any ways for the problem to really manifest before 8.2: AFAICS,
8.0 and 8.1 only use the ExprContextCallback mechanism to handle set-returning
functions, which cannot usefully be executed in a "simple expression" anyway.
Hence, no backpatch before 8.2 --- the risk of unforeseen breakage seems
to outweigh the chance of fixing something.

17 years agoDrat, can't fit an additional argument into log_error. Is it worth an
Tom Lane [Sun, 28 Jan 2007 07:29:32 +0000 (07:29 +0000)]
Drat, can't fit an additional argument into log_error.  Is it worth an
sprintf pushup to be sure we can report something useful for out-of-range
exitstatus?

17 years agoClean up broken usage of HAVE_DECL_SYS_SIGLIST and inconsistent/poorly
Tom Lane [Sun, 28 Jan 2007 06:32:03 +0000 (06:32 +0000)]
Clean up broken usage of HAVE_DECL_SYS_SIGLIST and inconsistent/poorly
formatted error messages.

17 years agoUse autoconf build-in sys_siglist macro AC_DECL_SYS_SIGLIST, rather than
Bruce Momjian [Sun, 28 Jan 2007 03:50:34 +0000 (03:50 +0000)]
Use autoconf build-in sys_siglist macro AC_DECL_SYS_SIGLIST, rather than
create our own.

17 years agoAdd a delay at the start of the stats test, to let any prior stats
Tom Lane [Sun, 28 Jan 2007 03:02:31 +0000 (03:02 +0000)]
Add a delay at the start of the stats test, to let any prior stats
activity quiesce.  Possibly this will fix the large increase in
non-reproducible stats test failures we've noted since turning on
stats_row_level by default.

17 years agoDept of second thoughts: the IQ of estimate_array_length() needs to be
Tom Lane [Sun, 28 Jan 2007 02:53:34 +0000 (02:53 +0000)]
Dept of second thoughts: the IQ of estimate_array_length() needs to be
kept on par with that of scalararraysel(), else estimates that should
track might not.  Hence teach it about binary-compatible cases, too.

17 years agoAdd signal.h for sys_siglist reference.
Bruce Momjian [Sun, 28 Jan 2007 02:33:09 +0000 (02:33 +0000)]
Add signal.h for sys_siglist reference.

17 years agoFix scalararraysel() to cope with binary-compatible cases, such as text[]
Tom Lane [Sun, 28 Jan 2007 01:37:38 +0000 (01:37 +0000)]
Fix scalararraysel() to cope with binary-compatible cases, such as text[]
versus varchar[].  This oversight probably explains Ryan Holmes' recent
complaint --- he was getting a generic selectivity estimate instead of
anything intelligent.

17 years agoUse sys_siglist[] to print out signal names for signal exits, rather
Bruce Momjian [Sun, 28 Jan 2007 01:12:05 +0000 (01:12 +0000)]
Use sys_siglist[] to print out signal names for signal exits, rather
than just numbers.

17 years agoModify SGML makefile to allow 'gmake draft' to build draft html ('draft'
Bruce Momjian [Sat, 27 Jan 2007 22:44:32 +0000 (22:44 +0000)]
Modify SGML makefile to allow 'gmake draft' to build draft html ('draft'
is now a target, no longer a modifier).

17 years agoCorrect an old logic error in btree page splitting: when considering a split
Tom Lane [Sat, 27 Jan 2007 20:53:30 +0000 (20:53 +0000)]
Correct an old logic error in btree page splitting: when considering a split
exactly at the point where we need to insert a new item, the calculation used
the wrong size for the "high key" of the new left page.  This could lead to
choosing an unworkable split, resulting in "PANIC: failed to add item to the
left sibling" (or "right sibling") failure.  Although this bug has been there
a long time, it's very difficult to trigger a failure before 8.2, since there
was generally a lot of free space on both sides of a chosen split.  In 8.2,
where the user-selected fill factor determines how much free space the code
tries to leave, an unworkable split is much more likely.  Report by Joe
Conway, diagnosis and fix by Heikki Linnakangas.

17 years agoFixed expected files, so they are in sync with tests again.
Michael Meskes [Sat, 27 Jan 2007 18:33:22 +0000 (18:33 +0000)]
Fixed expected files, so they are in sync with tests again.

17 years agoremove unnecessary and now inaccurate cast which I should have removed with other...
Andrew Dunstan [Sat, 27 Jan 2007 16:46:21 +0000 (16:46 +0000)]
remove unnecessary and now inaccurate cast which I should have removed with other old code.

17 years agoReactivate libxml memory management via palloc, now that I think I've
Peter Eisentraut [Sat, 27 Jan 2007 14:50:51 +0000 (14:50 +0000)]
Reactivate libxml memory management via palloc, now that I think I've
classified the conditions under which this is safe to do (see source
code comment).

17 years agoAdd trailing zero byte in Unicode codepoint conversion.
Peter Eisentraut [Sat, 27 Jan 2007 11:48:31 +0000 (11:48 +0000)]
Add trailing zero byte in Unicode codepoint conversion.

17 years agoAdd:
Bruce Momjian [Sat, 27 Jan 2007 03:25:49 +0000 (03:25 +0000)]
Add:

> * Enforce typmod for function inputs, function results and parameters for
>   spi_prepare'd statements called from PLs
>
>   http://archives.postgresql.org/pgsql-hackers/2007-01/msg01403.php

17 years agoAdd:
Bruce Momjian [Sat, 27 Jan 2007 02:29:32 +0000 (02:29 +0000)]
Add:

> * Consider having the background writer update the transaction status
>   hint bits before writing out the page

17 years agoAdd:
Bruce Momjian [Sat, 27 Jan 2007 02:28:16 +0000 (02:28 +0000)]
Add:

>
> * Consider increasing NUM_CLOG_BUFFERS

17 years agoAllow args to spi_prepare to be standard type aliaes as well as those known in pg_typ...
Andrew Dunstan [Sat, 27 Jan 2007 01:55:57 +0000 (01:55 +0000)]
Allow args to spi_prepare to be standard type aliaes as well as those known in pg_type. Fixes bug #2917. Add some regression tests for these cases.

17 years agoUpdate installation wording for an upgrade to state that dump/restore
Bruce Momjian [Sat, 27 Jan 2007 01:27:36 +0000 (01:27 +0000)]
Update installation wording for an upgrade to state that dump/restore
should not be done, per Peter.

17 years agoControl openjade draft-mode by variable DRAFT, rather than whether the
Bruce Momjian [Fri, 26 Jan 2007 23:51:39 +0000 (23:51 +0000)]
Control openjade draft-mode by variable DRAFT, rather than whether the
version tag is 'devel'.

17 years agoUpdate wording of installation when upgrading, to more clearly
Bruce Momjian [Fri, 26 Jan 2007 22:52:50 +0000 (22:52 +0000)]
Update wording of installation when upgrading, to more clearly
distinguish major vs minor release upgrades.

17 years agoModify draft SGML instructions to use DRAFT=Y because recursion was
Bruce Momjian [Fri, 26 Jan 2007 22:23:50 +0000 (22:23 +0000)]
Modify draft SGML instructions to use DRAFT=Y because recursion was
causing html to be called twice --- no way to exit the makefile after
the recursion returns.

17 years agoOn Windows, use pgwin32_waitforsinglesocket() instead of select() to wait for
Tom Lane [Fri, 26 Jan 2007 20:06:52 +0000 (20:06 +0000)]
On Windows, use pgwin32_waitforsinglesocket() instead of select() to wait for
input in the stats collector.  Our select() emulation is apparently buggy
for UDP sockets :-(.  This should resolve problems with stats collection
(and hence autovacuum) failing under more than minimal load.  Diagnosis
and patch by Magnus Hagander.

Patch probably needs to be back-ported to 8.1 and 8.0, but first let's
see if it makes the buildfarm happy...

17 years agoSquelch some VC++ compiler warnings. Mark float literals with the "f"
Neil Conway [Fri, 26 Jan 2007 17:45:42 +0000 (17:45 +0000)]
Squelch some VC++ compiler warnings. Mark float literals with the "f"
suffix, to distinguish them from doubles. Make some function declarations
and definitions use the "const" qualifier for arguments consistently.
Ignore warning 4102 ("unreferenced label"), because such warnings
are always emitted by bison-generated code. Patch from Magnus Hagander.

17 years agoUpdate BSD/OS platform for 8.2.
Bruce Momjian [Thu, 25 Jan 2007 23:34:28 +0000 (23:34 +0000)]
Update BSD/OS platform for 8.2.