]> granicus.if.org Git - postgresql/log
postgresql
16 years agoInclude -lgss in libpq link, if available. Bjorn Munch
Tom Lane [Wed, 5 Mar 2008 05:39:11 +0000 (05:39 +0000)]
Include -lgss in libpq link, if available.  Bjorn Munch

16 years agoFix PREPARE TRANSACTION to reject the case where the transaction has dropped a
Tom Lane [Tue, 4 Mar 2008 19:54:06 +0000 (19:54 +0000)]
Fix PREPARE TRANSACTION to reject the case where the transaction has dropped a
temporary table; we can't support that because there's no way to clean up the
source backend's internal state if the eventual COMMIT PREPARED is done by
another backend.  This was checked correctly in 8.1 but I broke it in 8.2 :-(.
Patch by Heikki Linnakangas, original trouble report by John Smith.

16 years agoUse windows DACL fix for pg_regress as well.
Magnus Hagander [Tue, 4 Mar 2008 15:38:31 +0000 (15:38 +0000)]
Use windows DACL fix for pg_regress as well.

Dave Page

16 years agoUpdate pg_dump item:
Bruce Momjian [Tue, 4 Mar 2008 14:38:38 +0000 (14:38 +0000)]
Update pg_dump item:

<    produce a single dump output file.
>    produce a single dump output file.  It also would require
>    several sessions to share the same snapshot.

16 years agoShow example of ts_headline() using a configuration name.
Bruce Momjian [Tue, 4 Mar 2008 03:17:18 +0000 (03:17 +0000)]
Show example of ts_headline() using a configuration name.

16 years agoRemove:
Bruce Momjian [Tue, 4 Mar 2008 02:48:22 +0000 (02:48 +0000)]
Remove:

<
<  o To better utilize resources, restore data, primary keys, and
<     indexes for a single table before restoring the next table
<
<    Hopefully this will allow the CPU-I/O load to be more uniform
<    for simultaneous restores.  The idea is to start data restores
<    for several objects, and once the first object is done, to move
<    on to its primary keys and indexes.  Over time, simultaneous
<    data loads and index builds will be running.

16 years agoAdd URL for:
Bruce Momjian [Tue, 4 Mar 2008 01:38:37 +0000 (01:38 +0000)]
Add URL for:

        o Allow pg_restore to utilize multiple CPUs and I/O channels by
          restoring multiple objects simultaneously

>    http://archives.postgresql.org/pgsql-hackers/2008-02/msg00963.php

16 years agoAdd ideas for concurrent pg_dump and pg_restore:
Bruce Momjian [Tue, 4 Mar 2008 01:33:32 +0000 (01:33 +0000)]
Add ideas for concurrent pg_dump and pg_restore:

< * pg_dump
> * pg_dump / pg_restore
>  o Allow pg_dump to utilize multiple CPUs and I/O channels by dumping
>    multiple objects simultaneously
>
>    The difficulty with this is getting multiple dump processes to
>    produce a single dump output file.
>    http://archives.postgresql.org/pgsql-hackers/2008-02/msg00205.php
>
>  o Allow pg_restore to utilize multiple CPUs and I/O channels by
>           restoring multiple objects simultaneously
>
>    This might require a pg_restore flag to indicate how many
>    simultaneous operations should be performed.  Only pg_dump's
>    -Fc format has the necessary dependency information.
>
>  o To better utilize resources, restore data, primary keys, and
>     indexes for a single table before restoring the next table
>
>    Hopefully this will allow the CPU-I/O load to be more uniform
>    for simultaneous restores.  The idea is to start data restores
>    for several objects, and once the first object is done, to move
>    on to its primary keys and indexes.  Over time, simultaneous
>    data loads and index builds will be running.
>
>  o To better utilize resources, allow pg_restore to check foreign
>    keys simultaneously, where possible
>  o Allow pg_restore to create all indexes of a table
>    concurrently, via a single heap scan
>
>    This requires a pg_dump -Fc file because that format contains
>           the required dependency information.
>    http://archives.postgresql.org/pgsql-general/2007-05/msg01274.php
>
>  o Allow pg_restore to load different parts of the COPY data
>    simultaneously
<   single heap scan, and have a restore of a pg_dump somehow use it
>   single heap scan, and have pg_restore use it
<   http://archives.postgresql.org/pgsql-general/2007-05/msg01274.php

16 years agoUpdate Japanese FAQ.
Bruce Momjian [Mon, 3 Mar 2008 22:10:51 +0000 (22:10 +0000)]
Update Japanese FAQ.

Jun Kuwamura

16 years agoAdd:
Bruce Momjian [Mon, 3 Mar 2008 21:26:09 +0000 (21:26 +0000)]
Add:

> * Allow INSERT ... DELETE ... RETURNING, namely allow the DELETE ...
>   RETURNING to supply values to the INSERT
>   http://archives.postgresql.org/pgsql-hackers/2008-02/thrd2.php#00979

16 years agoAdd another URL for:
Bruce Momjian [Mon, 3 Mar 2008 21:00:35 +0000 (21:00 +0000)]
Add another URL for:

        o Consider using a ring buffer for COPY FROM
<
< http://archives.postgresql.org/pgsql-hackers/2008-02/msg01080.php
>    http://archives.postgresql.org/pgsql-hackers/2008-02/msg01080.php

16 years agoDocument that REVOKE doesn't remove all permissions if PUBLIC has permissions.
Bruce Momjian [Mon, 3 Mar 2008 19:17:27 +0000 (19:17 +0000)]
Document that REVOKE doesn't remove all permissions if PUBLIC has permissions.

16 years agoAdd:
Bruce Momjian [Mon, 3 Mar 2008 19:03:44 +0000 (19:03 +0000)]
Add:

>
> * Allow the UUID type to accept non-standard formats
>
>   http://archives.postgresql.org/pgsql-hackers/2008-02/msg01214.php

16 years agoAdd:
Bruce Momjian [Mon, 3 Mar 2008 18:45:24 +0000 (18:45 +0000)]
Add:

> * Speed WAL recovery by allowing more than one page to be prefetched
>
>   This involves having a separate process that can be told which pages
>   the recovery process will need in the near future.
>   http://archives.postgresql.org/pgsql-hackers/2008-02/msg01279.php
>

16 years agoRemove "reliably" from chr(0) doc mention.
Bruce Momjian [Mon, 3 Mar 2008 18:09:02 +0000 (18:09 +0000)]
Remove "reliably" from chr(0) doc mention.

16 years agoDocument that chr(0) is not supported, and why.
Bruce Momjian [Mon, 3 Mar 2008 17:11:13 +0000 (17:11 +0000)]
Document that chr(0) is not supported, and why.

16 years agoFix markup.
Bruce Momjian [Mon, 3 Mar 2008 16:22:55 +0000 (16:22 +0000)]
Fix markup.

16 years agoAdd new FAQ item:
Bruce Momjian [Mon, 3 Mar 2008 15:55:26 +0000 (15:55 +0000)]
Add new FAQ item:

    <H3 id="item1.15">1.15) How do I unsubscribe from the
    PostgreSQL email lists?  How do I avoid receiving duplicate
    emails?</H3>

16 years agoAdd URL's for sequence discussions:
Bruce Momjian [Mon, 3 Mar 2008 15:06:55 +0000 (15:06 +0000)]
Add URL's for sequence discussions:

>
>   http://archives.postgresql.org/pgsql-hackers/2008-03/msg00008.php
>
<  o %Have ALTER TABLE RENAME rename SERIAL sequence names
>  o Have ALTER TABLE RENAME rename SERIAL sequence names
>
>    http://archives.postgresql.org/pgsql-hackers/2008-03/msg00008.php
>
>    http://archives.postgresql.org/pgsql-hackers/2008-03/msg00008.php

16 years agoFixed bug that caused arrays of varchar to be output with incomplete name.
Michael Meskes [Sun, 2 Mar 2008 10:54:11 +0000 (10:54 +0000)]
Fixed bug that caused arrays of varchar to be output with incomplete name.
In the process expanded one test case,

16 years agoVenezuela Time now means UTC-4:30, not UTC-4:00. Adjust our treatment
Tom Lane [Sun, 2 Mar 2008 00:10:22 +0000 (00:10 +0000)]
Venezuela Time now means UTC-4:30, not UTC-4:00.  Adjust our treatment
of "VET" accordingly.  Per bug #3997 from Aaron Mizrachi.

16 years agoFix another place that was assuming that a local variable declared as
Tom Lane [Sat, 1 Mar 2008 19:26:22 +0000 (19:26 +0000)]
Fix another place that was assuming that a local variable declared as
"struct varlena" would be at least word-aligned.  Per buildfarm results
from gypsy_moth.  I did a little bit of trawling for other instances of
this coding pattern, and didn't find any; but if we turn up any more
of them I think we'd better revert the "char [4]" patch and find another
way of making tuptoaster.c alignment-safe.

16 years agoFix unportable usages of tolower(). On signed-char machines, it is necessary
Tom Lane [Sat, 1 Mar 2008 03:26:35 +0000 (03:26 +0000)]
Fix unportable usages of tolower().  On signed-char machines, it is necessary
to explicitly cast the output back to char before comparing it to a char
value, else we get the wrong result for high-bit-set characters.  Found by
Rolf Jentsch.  Also, fix several places where <ctype.h> functions were being
called without casting the argument to unsigned char; this is likewise
unportable, but we keep making that mistake :-(.  These found by buildfarm
member salamander, which I will desperately miss if it ever goes belly-up.

16 years agoDisable the undocumented xmlvalidate() function, which was unintentionally
Tom Lane [Sat, 1 Mar 2008 02:46:49 +0000 (02:46 +0000)]
Disable the undocumented xmlvalidate() function, which was unintentionally
left in the code though it was not meant to be provided.  It represents a
security hole because unprivileged users could use it to look at (at least the
first line of) any file readable by the backend.  Fortunately, this is only
possible if the backend was built with XML support, so the damage is at least
mitigated; and 8.3 probably hasn't propagated into any security-critical uses
yet anyway.  Per report from Sergey Burladyan.

16 years agoDon't call AddUserToDacl on Cygwin
Andrew Dunstan [Fri, 29 Feb 2008 23:31:20 +0000 (23:31 +0000)]
Don't call AddUserToDacl on Cygwin

16 years agoRemove long-unused and broken TCL_ARRAYS.
Alvaro Herrera [Fri, 29 Feb 2008 20:58:33 +0000 (20:58 +0000)]
Remove long-unused and broken TCL_ARRAYS.

16 years agoReducing the assumed alignment of struct varlena means that the compiler
Tom Lane [Fri, 29 Feb 2008 17:47:41 +0000 (17:47 +0000)]
Reducing the assumed alignment of struct varlena means that the compiler
is also licensed to put a local variable declared that way at an unaligned
address.  Which will not work if the variable is then manipulated with
SET_VARSIZE or other macros that assume alignment.  So the previous patch
is not an unalloyed good, but on balance I think it's still a win, since
we have very few places that do that sort of thing.  Fix the one place in
tuptoaster.c that does it.  Per buildfarm results from gypsy_moth
(I'm a bit surprised that only one machine showed a failure).

16 years agoFix handling of restricted processes for Windows Vista (mainly),
Magnus Hagander [Fri, 29 Feb 2008 15:31:33 +0000 (15:31 +0000)]
Fix handling of restricted processes for Windows Vista (mainly),
by explicitly adding back the user to the DACL of the new process.
This fixes the failure case when executing as the Administrator
user, which had no permissions left at all after we dropped the
Administrators group.

Dave Page with some modifications from me

16 years agoUpdate wording:
Bruce Momjian [Fri, 29 Feb 2008 14:56:50 +0000 (14:56 +0000)]
Update wording:

< * Add Oracle-style packages  (Pavel)
> * Add features of Oracle-style packages  (Pavel)

16 years agoWrite the objfiles.txt rules in a way that is compatible with GNU make 3.78,
Peter Eisentraut [Fri, 29 Feb 2008 10:34:51 +0000 (10:34 +0000)]
Write the objfiles.txt rules in a way that is compatible with GNU make 3.78,
and simpler, too.

16 years agoFix several memory leaks when rescanning SRFs. Arrange for an SRF's
Neil Conway [Fri, 29 Feb 2008 02:49:39 +0000 (02:49 +0000)]
Fix several memory leaks when rescanning SRFs. Arrange for an SRF's
"multi_call_ctx" to be a distinct sub-context of the EState's per-query
context, and delete the multi_call_ctx as soon as the SRF finishes
execution. This avoids leaking SRF memory until the end of the current
query, which is particularly egregious when the SRF is scanned
multiple times. This change also fixes a leak of the fields of the
AttInMetadata struct in shutdown_MultiFuncCall().

Also fix a leak of the SRF result TupleDesc when rescanning a
FunctionScan node. The TupleDesc is allocated in the per-query context
for every call to ExecMakeTableFunctionResult(), so we should free it
after calling that function. Since the SRF might choose to return
a non-expendable TupleDesc, we only free the TupleDesc if it is
not being reference-counted.

Backpatch to 8.3 and 8.2 stable branches.

16 years agoSupport for building contrib/uuid-ossp with MSVC.
Magnus Hagander [Thu, 28 Feb 2008 12:17:59 +0000 (12:17 +0000)]
Support for building contrib/uuid-ossp with MSVC.

Original patch from Hiroshi Saito, modified by me.

16 years agoChange expand_subsys function so that it preserves the relative order of
Peter Eisentraut [Wed, 27 Feb 2008 20:31:01 +0000 (20:31 +0000)]
Change expand_subsys function so that it preserves the relative order of
the files passed as argument.  This is desirable so that the dtrace rule
in src/backend/Makefile works.

16 years agoIf RelationBuildDesc() fails to open a critical system index, PANIC with
Tom Lane [Wed, 27 Feb 2008 17:44:19 +0000 (17:44 +0000)]
If RelationBuildDesc() fails to open a critical system index, PANIC with
a relevant error message instead of just dumping core.  Odd that nobody
reported this before Darren Reed.

16 years agoAdd more clarification about SSH tunnels from Faheem Mitha.
Peter Eisentraut [Tue, 26 Feb 2008 18:01:26 +0000 (18:01 +0000)]
Add more clarification about SSH tunnels from Faheem Mitha.

16 years agoIn the SSH setup instructions, change
Peter Eisentraut [Tue, 26 Feb 2008 16:07:16 +0000 (16:07 +0000)]
In the SSH setup instructions, change

ssh -L 3333:foo.com:5432 joe@foo.com

I think this should be changed to

ssh -L 3333:localhost:5432 joe@foo.com

The reason is that this assumes the postgres server on foo.com allows
connections from foo.com, which is not allowed by the default
listen_addresses setting.  Add more detail explaining this.

pointed out by Faheem Mitha

Also change the example port number 3333 to 63333 so no one can complain
that we are stealing a reserved port number.

16 years agoAdd information about format modifiers that apply to numeric formats.
Peter Eisentraut [Tue, 26 Feb 2008 15:32:30 +0000 (15:32 +0000)]
Add information about format modifiers that apply to numeric formats.
These were previously only documented in the context of date/time formats.

16 years agoFixed dtrace build
Peter Eisentraut [Tue, 26 Feb 2008 14:42:27 +0000 (14:42 +0000)]
Fixed dtrace build

found by Magne Mæhre

16 years agoDon't build the win32 support files in the all target, only in distprep and
Peter Eisentraut [Tue, 26 Feb 2008 14:26:16 +0000 (14:26 +0000)]
Don't build the win32 support files in the all target, only in distprep and
when they are actually needed as prerequisites.

16 years agoFix uninstall target.
Peter Eisentraut [Tue, 26 Feb 2008 13:48:57 +0000 (13:48 +0000)]
Fix uninstall target.

16 years agoCreate two separate libpq.rc's: One that is built at build time, and one
Peter Eisentraut [Tue, 26 Feb 2008 13:31:40 +0000 (13:31 +0000)]
Create two separate libpq.rc's: One that is built at build time, and one
that is shipped in the distribution, named libpq-dist.rc.  This way the
build system doesn't get upset when a distributed file is forcibly
overwritten by during a normal build.

16 years agoReorganize some of the exports list generation code. It seems that this
Peter Eisentraut [Tue, 26 Feb 2008 10:45:24 +0000 (10:45 +0000)]
Reorganize some of the exports list generation code.  It seems that this
has been reinvented about four different times throughout history (aix,
cygwin, win32, darwin/linux) and a lot of the concepts are actually shared,
which the code now shows better.

16 years agoEscape # character in variable assignment
Peter Eisentraut [Tue, 26 Feb 2008 10:30:06 +0000 (10:30 +0000)]
Escape # character in variable assignment

16 years agoNeed more dependencies to get the build order right when objfiles.txt
Peter Eisentraut [Tue, 26 Feb 2008 08:23:31 +0000 (08:23 +0000)]
Need more dependencies to get the build order right when objfiles.txt
doesn't exist yet.

16 years agoWe don't need to rebuild objfiles.txt every time an object file changes.
Peter Eisentraut [Tue, 26 Feb 2008 07:20:38 +0000 (07:20 +0000)]
We don't need to rebuild objfiles.txt every time an object file changes.
So only rebuild when a makefile changes (which presumably defines the
file list somewhere), and only touch the file if an object changed. The
touch is necessary so the parent make knows something changed and
ultimately rebuilds postgres.

16 years agoRefactor the code that creates the shared library export files to appear
Peter Eisentraut [Tue, 26 Feb 2008 06:41:24 +0000 (06:41 +0000)]
Refactor the code that creates the shared library export files to appear
only once in Makefile.shlib and not in four copies.

16 years agoFix encode(...bytea..., 'escape') so that it converts all high-bit-set byte
Tom Lane [Tue, 26 Feb 2008 02:54:08 +0000 (02:54 +0000)]
Fix encode(...bytea..., 'escape') so that it converts all high-bit-set byte
values into \nnn octal escape sequences.  When the database encoding is
multibyte this is *necessary* to avoid generating invalidly encoded text.
Even in a single-byte encoding, the old behavior seems very hazardous ---
consider for example what happens if the text is transferred to another
database with a different encoding.  Decoding would then yield some other
bytea value than what was encoded, which is surely undesirable.  Per gripe
from Hernan Gonzalez.

Backpatch to 8.3, but not further.  This is a bit of a judgment call, but I
make it on these grounds: pre-8.3 we don't really have much encoding safety
anyway because of the convert() function family, and we would also have much
higher risk of breaking existing apps that may not be expecting this behavior.
8.3 is still new enough that we can probably get away with making this change
in the function's behavior.

16 years agoReject year zero during datetime input, except when it's a 2-digit year
Tom Lane [Mon, 25 Feb 2008 23:36:28 +0000 (23:36 +0000)]
Reject year zero during datetime input, except when it's a 2-digit year
(then it means 2000 AD).  Formerly we silently interpreted this as 1 BC,
which at best is unwarranted familiarity with the implementation.
It's barely possible that some app somewhere expects the old behavior,
though, so we won't back-patch this into existing release branches.

16 years agoFix datetime input to behave correctly for Feb 29 in years BC.
Tom Lane [Mon, 25 Feb 2008 23:21:01 +0000 (23:21 +0000)]
Fix datetime input to behave correctly for Feb 29 in years BC.

Formerly, DecodeDate attempted to verify the day-of-the-month exactly, but
it was under the misapprehension that it would know whether we were looking
at a BC year or not.  In reality this check can't be made until the calling
function (eg DecodeDateTime) has processed all the fields.  So, split the
BC adjustment and validity checks out into a new function ValidateDate that
is called only after processing all the fields.  In passing, this patch
makes DecodeTimeOnly work for BC inputs, which it never did before.

(The historical veracity of all this is nonexistent, of course, but if
we're going to say we support proleptic Gregorian calendar then we should
do it correctly.  In any case the unpatched code is broken because it could
emit dates that it would then reject on re-inputting.)

Per report from Bernd Helmle.  Back-patch as far as 8.0; in 7.x we were
not using our own calendar support and so this seems a bit too risky
to put into 7.4.

16 years agoLink postgres from all object files at once, to avoid the error-prone
Peter Eisentraut [Mon, 25 Feb 2008 17:55:42 +0000 (17:55 +0000)]
Link postgres from all object files at once, to avoid the error-prone
SUBSYS.o step and allow for better optimization by the linker.

Instead of partial linking into SUBSYS.o, the list of object files is
assembled in objfiles.txt files that are expanded when the final
linking is done.

Because we are not yet sure how long command lines different platforms
can handle, the old way of linking is still available, by defining the
make variable PARTIAL_LINKING (e.g., make all PARTIAL_LINKING=1).  If
we determine that this is necessary for some platforms, then we will
document this in a more prominent place.

16 years agoUse our own getopt() and getopt_long() on Solaris, because that platform's
Tom Lane [Sun, 24 Feb 2008 05:21:54 +0000 (05:21 +0000)]
Use our own getopt() and getopt_long() on Solaris, because that platform's
versions don't handle long options the way we want.  Per Zdenek Kotala.

16 years agoAvoid trying to print a NULL char pointer in --describe-config. On some
Tom Lane [Sat, 23 Feb 2008 19:23:33 +0000 (19:23 +0000)]
Avoid trying to print a NULL char pointer in --describe-config.  On some
platforms this works, but on some it crashes.  Zdenek Kotala

16 years agoChange the declaration of struct varlena so that the length word is
Tom Lane [Sat, 23 Feb 2008 19:11:45 +0000 (19:11 +0000)]
Change the declaration of struct varlena so that the length word is
represented as "char ...[4]" not "int32".  Since the length word is never
supposed to be accessed via this struct member anyway, this won't break
any existing code that is following the rules.  The advantage is that C
compilers will no longer assume that a pointer to struct varlena is
word-aligned, which prevents incorrect optimizations in TOAST-pointer
access and perhaps other places.  gcc doesn't seem to do this (at least
not at -O2), but the problem is demonstrable on some other compilers.

I changed struct inet as well, but didn't bother to touch a lot of other
struct definitions in which it wouldn't make any difference because there
were other fields forcing int alignment anyway.  Hopefully none of those
struct definitions are used for accessing unaligned Datums.

16 years agoRename miscadmin.h's PG_VERSIONSTR macro to PG_BACKEND_VERSIONSTR to
Tom Lane [Wed, 20 Feb 2008 22:46:24 +0000 (22:46 +0000)]
Rename miscadmin.h's PG_VERSIONSTR macro to PG_BACKEND_VERSIONSTR to
make it a bit clearer what it is, and get rid of duplicate definitions
in initdb and pg_ctl.

16 years agoMake pg_regress -V consistent with the corresponding code in other
Tom Lane [Wed, 20 Feb 2008 22:44:16 +0000 (22:44 +0000)]
Make pg_regress -V consistent with the corresponding code in other
programs: use puts with a compile-time-constant string.

16 years agoFix mistakes in pg_ctl's code for "start -w" that tries to cope with
Tom Lane [Wed, 20 Feb 2008 22:18:15 +0000 (22:18 +0000)]
Fix mistakes in pg_ctl's code for "start -w" that tries to cope with
non-default settings for the postmaster's port number.  The code to parse
command line options and postgresql.conf entries wasn't quite right about
whitespace or quotes, and it was coded in a not-very-readable way too.
Per bug #3969 from Itagaki Takahiro, though this is more extensive than his
proposed patch (which fixed only the whitespace problem).
This code has been broken since it was put in in 8.0, so patch all the way
back.

16 years agoPut a CHECK_FOR_INTERRUPTS call into the loops that try to find a unique new
Tom Lane [Wed, 20 Feb 2008 17:44:09 +0000 (17:44 +0000)]
Put a CHECK_FOR_INTERRUPTS call into the loops that try to find a unique new
OID or new relfilenode.  If the existing OIDs are sufficiently densely
populated, this could take a long time (perhaps even be an infinite loop),
so it seems wise to allow the system to respond to a cancel interrupt here.
Per a gripe from Jacky Leng.

Backpatch as far as 8.1.  Older versions just fail on OID collision,
instead of looping.

16 years agoImprove error messages emitted when VACUUM and ANALYZE skip a table.
Alvaro Herrera [Wed, 20 Feb 2008 14:31:35 +0000 (14:31 +0000)]
Improve error messages emitted when VACUUM and ANALYZE skip a table.

Per gripe from Clodoaldo Pinto Neto on
Message-ID: <a595de7a0801060326qbfc790ax2a60573043c2e2be@mail.gmail.com>

16 years agoChange error message to be able to differentiate the two cases. Per suggestion
Alvaro Herrera [Wed, 20 Feb 2008 14:01:45 +0000 (14:01 +0000)]
Change error message to be able to differentiate the two cases.  Per suggestion
from Jaime Casanova.

16 years agoBackport fixed AC_FUNC_FSEEKO
Peter Eisentraut [Tue, 19 Feb 2008 18:02:30 +0000 (18:02 +0000)]
Backport fixed AC_FUNC_FSEEKO

16 years agoUn-break msvc port yet again (it started pulling in the Darwin
Magnus Hagander [Tue, 19 Feb 2008 16:15:14 +0000 (16:15 +0000)]
Un-break msvc port yet again (it started pulling in the Darwin
port files, which obviously didn't compile)

16 years agoMore refactoring, so that the SUBSYS.o rules are now all in one place.
Peter Eisentraut [Tue, 19 Feb 2008 15:29:58 +0000 (15:29 +0000)]
More refactoring, so that the SUBSYS.o rules are now all in one place.

16 years agoThis subdirectory has been unused, dead, and broken for 10 years.
Peter Eisentraut [Tue, 19 Feb 2008 13:08:56 +0000 (13:08 +0000)]
This subdirectory has been unused, dead, and broken for 10 years.

16 years agoFix function prototype to silence compiler warnings.
Magnus Hagander [Tue, 19 Feb 2008 12:06:35 +0000 (12:06 +0000)]
Fix function prototype to silence compiler warnings.

16 years agoUnbreak MSVC build after recent addition of HTMLDIR.
Magnus Hagander [Tue, 19 Feb 2008 12:00:03 +0000 (12:00 +0000)]
Unbreak MSVC build after recent addition of HTMLDIR.

16 years agoRemove another target I forgot during the refactoring
Peter Eisentraut [Tue, 19 Feb 2008 11:49:12 +0000 (11:49 +0000)]
Remove another target I forgot during the refactoring

16 years agoRefactor backend makefiles to remove lots of duplicate code
Peter Eisentraut [Tue, 19 Feb 2008 10:30:09 +0000 (10:30 +0000)]
Refactor backend makefiles to remove lots of duplicate code

16 years agoPut back code modularization of fseeko() configure checks.
Bruce Momjian [Tue, 19 Feb 2008 01:05:28 +0000 (01:05 +0000)]
Put back code modularization of fseeko() configure checks.

16 years agoPut fseeko check back in old location, in hopes of silencing build warnings.
Bruce Momjian [Tue, 19 Feb 2008 00:46:43 +0000 (00:46 +0000)]
Put fseeko check back in old location, in hopes of silencing build warnings.

16 years agoRemove unnecessary opening of other relation in RI_FKey_keyequal_upd_pk
Tom Lane [Mon, 18 Feb 2008 23:00:32 +0000 (23:00 +0000)]
Remove unnecessary opening of other relation in RI_FKey_keyequal_upd_pk
and RI_FKey_keyequal_upd_fk, as well as no-longer-needed calls of
ri_BuildQueryKeyFull.  Aside from saving a few cycles, this avoids needless
deadlock risks when an update is not changing the columns that participate
in an RI constraint.  Per a gripe from Alexey Nalbat.

Back-patch to 8.3.  Earlier releases did have a need to open the other
relation due to the way in which they retrieved information about the RI
constraint, so this problem unfortunately can't easily be improved pre-8.3.

Tom Lane and Stephan Szabo

16 years agoautoconf 2.61's AC_FUNC_FSEEKO reports success/failure differently, so
Bruce Momjian [Mon, 18 Feb 2008 21:46:22 +0000 (21:46 +0000)]
autoconf 2.61's AC_FUNC_FSEEKO reports success/failure differently, so
reorganize code for NetBSD/BSDi port/fseeko.c usage, and make code more
modular.

16 years agoObserve errors in makefile
Peter Eisentraut [Mon, 18 Feb 2008 16:04:32 +0000 (16:04 +0000)]
Observe errors in makefile

16 years agoAdded --htmldir option to pg_config, equivalent to the new configure option.
Peter Eisentraut [Mon, 18 Feb 2008 14:51:48 +0000 (14:51 +0000)]
Added --htmldir option to pg_config, equivalent to the new configure option.

16 years ago- Removed duplicate include of ecpgtype.h which meant I had to adapt all expected...
Michael Meskes [Sun, 17 Feb 2008 18:14:29 +0000 (18:14 +0000)]
- Removed duplicate include of ecpgtype.h which meant I had to adapt all expected results.
- Changed INFORMIX mode symbol definition yet again because the old way didn't work on NetBSD. Hopefully this one does.

16 years agoUpgrade to Autoconf 2.61:
Peter Eisentraut [Sun, 17 Feb 2008 16:36:43 +0000 (16:36 +0000)]
Upgrade to Autoconf 2.61:
- Change configure.in to use Autoconf 2.61 and update generated files.
- Update build system and documentation to support now directory variables
  offered by Autoconf 2.61.
- Replace usages of PGAC_CHECK_ALIGNOF by AC_CHECK_ALIGNOF, now available
  in Autoconf 2.61.
- Drop our patched version of AC_C_INLINE, as Autoconf now has the change.

16 years agoAdd back #include <time.h> in a couple of files that seem to need it
Tom Lane [Sun, 17 Feb 2008 04:21:05 +0000 (04:21 +0000)]
Add back #include <time.h> in a couple of files that seem to need it
on Linux.

16 years agoReplace time_t with pg_time_t (same values, but always int64) in on-disk
Tom Lane [Sun, 17 Feb 2008 02:09:32 +0000 (02:09 +0000)]
Replace time_t with pg_time_t (same values, but always int64) in on-disk
data structures and backend internal APIs.  This solves problems we've seen
recently with inconsistent layout of pg_control between machines that have
32-bit time_t and those that have already migrated to 64-bit time_t.  Also,
we can get out from under the problem that Windows' Unix-API emulation is not
consistent about the width of time_t.

There are a few remaining places where local time_t variables are used to hold
the current or recent result of time(NULL).  I didn't bother changing these
since they do not affect any cross-module APIs and surely all platforms will
have 64-bit time_t before overflow becomes an actual risk.  time_t should
be avoided for anything visible to extension modules, however.

16 years agoMove item as done:
Bruce Momjian [Sat, 16 Feb 2008 23:27:34 +0000 (23:27 +0000)]
Move item as done:

<
> * -Allow AS in "SELECT col AS label" to be optional in certain cases
>
< * Allow AS in "SELECT col AS label" to be optional (not wanted)

16 years agoUpdate docs to reflect the fact that we can now deal with DST rules
Tom Lane [Sat, 16 Feb 2008 21:51:04 +0000 (21:51 +0000)]
Update docs to reflect the fact that we can now deal with DST rules
outside the 32-bit-time_t range.  Also, refer to Olson's tz database
as the 'zoneinfo' database, a name that upstream sometimes uses, not
'zic database' which they never use.

16 years agoUpdate timezone code to track the upstream changes since 2003. In particular
Tom Lane [Sat, 16 Feb 2008 21:16:04 +0000 (21:16 +0000)]
Update timezone code to track the upstream changes since 2003.  In particular
this adds support for 64-bit tzdata files, which is needed to support DST
calculations beyond 2038.  Add a regression test case to give some minimal
confidence that that really works.

Heikki Linnakangas

16 years agoNo need to use checkpoint_warning to log checkpoints --- we added
Bruce Momjian [Sat, 16 Feb 2008 21:14:08 +0000 (21:14 +0000)]
No need to use checkpoint_warning to log checkpoints --- we added
log_checkpoints in 8.3.

16 years agoAdd suggestion about how to set checkpoint_warning to log all
Bruce Momjian [Sat, 16 Feb 2008 21:11:32 +0000 (21:11 +0000)]
Add suggestion about how to set checkpoint_warning to log all
checkpoints.

16 years agoRename a libpq NOT_USED SSL function to
Bruce Momjian [Sat, 16 Feb 2008 21:03:30 +0000 (21:03 +0000)]
Rename a libpq NOT_USED SSL function to
verify_peer_name_matches_certificate(), clarify some of the function's
variables and logic, and update a comment.  This should make SSL
improvements easier in the future.

16 years agoAllow AS to be omitted when specifying an output column name in SELECT
Tom Lane [Fri, 15 Feb 2008 22:17:06 +0000 (22:17 +0000)]
Allow AS to be omitted when specifying an output column name in SELECT
(or RETURNING), but only when the output name is not any SQL keyword.
This seems as close as we can get to the standard's syntax without a
great deal of thrashing.  Original patch by Hiroshi Saito, amended by me.

16 years agoRemove ancient restriction that LIMIT/OFFSET can't contain a sub-select.
Tom Lane [Fri, 15 Feb 2008 17:19:46 +0000 (17:19 +0000)]
Remove ancient restriction that LIMIT/OFFSET can't contain a sub-select.
This was probably protecting some implementation limitation when it was
put in, but as far as I can tell the planner and executor have no such
assumption anymore; the case seems to work fine.  Per a gripe from
Grzegorz Jaskiewicz.

16 years agoChanged the way symbols are defined in C in INFORMIX mode.
Michael Meskes [Fri, 15 Feb 2008 11:20:21 +0000 (11:20 +0000)]
Changed the way symbols are defined in C in INFORMIX mode.

16 years agoSync our regex code with upstream changes since last time we did this, which
Tom Lane [Thu, 14 Feb 2008 17:33:37 +0000 (17:33 +0000)]
Sync our regex code with upstream changes since last time we did this, which
was Tcl 8.4.8.  The main changes are to remove the never-fully-implemented
code for multi-character collating elements, and to const-ify some stuff a
bit more fully.  In combination with the recent security patch, this commit
brings us into line with Tcl 8.5.0.

Note that I didn't make any effort to duplicate a lot of cosmetic changes
that they made to bring their copy into line with their own style
guidelines, such as adding braces around single-line IF bodies.  Most of
those we either had done already (such as ANSI-fication of function headers)
or there is no point because pgindent would undo the change anyway.

16 years ago- EXECUTE can return NOT FOUND so it should be checked here too.
Michael Meskes [Thu, 14 Feb 2008 14:54:48 +0000 (14:54 +0000)]
- EXECUTE can return NOT FOUND so it should be checked here too.
- Changed regression test accordingly.

16 years agoSorry, accidently committed a patch I'm working on. Reverting it.
Michael Meskes [Thu, 14 Feb 2008 12:27:26 +0000 (12:27 +0000)]
Sorry, accidently committed a patch I'm working on. Reverting it.
The only correct change was:
- Added SQLSTATE macro closing bug #3961.

16 years ago*** empty log message ***
Michael Meskes [Thu, 14 Feb 2008 12:22:36 +0000 (12:22 +0000)]
*** empty log message ***

16 years agoCorrect XML markup typo, <book> -> </book>
Bruce Momjian [Wed, 13 Feb 2008 22:46:55 +0000 (22:46 +0000)]
Correct XML markup typo, <book> -> </book>

 XMLPARSE (DOCUMENT '<?xml
 version="1.0"?><book><title>Manual</title><chapter>...</chapter></book>')

Backpatch to 8.3.X.

16 years agoAdd section to Migration talking about typical areas that change between
Bruce Momjian [Wed, 13 Feb 2008 22:44:06 +0000 (22:44 +0000)]
Add section to Migration talking about typical areas that change between
major releases.

16 years agoUpdate minor version bumping policy.
Alvaro Herrera [Wed, 13 Feb 2008 21:09:24 +0000 (21:09 +0000)]
Update minor version bumping policy.

16 years agoAs sub-bullet decoration.
Bruce Momjian [Wed, 13 Feb 2008 18:30:21 +0000 (18:30 +0000)]
As sub-bullet decoration.

16 years agoUpdate wording for minor library bumping.
Bruce Momjian [Wed, 13 Feb 2008 18:29:08 +0000 (18:29 +0000)]
Update wording for minor library bumping.

16 years agoBump minor library version numbers for 8.4.
Bruce Momjian [Wed, 13 Feb 2008 18:14:46 +0000 (18:14 +0000)]
Bump minor library version numbers for 8.4.

16 years agoNo longer necessary:
Bruce Momjian [Wed, 13 Feb 2008 18:10:23 +0000 (18:10 +0000)]
No longer necessary:

        o update ecpg regression expected files for new library number

16 years agoStamp HEAD as 8.4devel.
Tom Lane [Wed, 13 Feb 2008 03:40:38 +0000 (03:40 +0000)]
Stamp HEAD as 8.4devel.

16 years agoFix SPI_cursor_open() and SPI_is_cursor_plan() to push the SPI stack before
Tom Lane [Tue, 12 Feb 2008 04:09:44 +0000 (04:09 +0000)]
Fix SPI_cursor_open() and SPI_is_cursor_plan() to push the SPI stack before
doing anything interesting, such as calling RevalidateCachedPlan().  The
necessity of this is demonstrated by an example from Willem Buitendyk:
during a replan, the planner might try to evaluate SPI-using functions,
and so we'd better be in a clean SPI context.

A small downside of this fix is that these two functions will now fail
outright if called when not inside a SPI-using procedure (ie, a
SPI_connect/SPI_finish pair).  The documentation never promised or suggested
that that would work, though; and they are normally used in concert with
other functions, mainly SPI_prepare, that always have failed in such a case.
So the odds of breaking something seem pretty low.

In passing, make SPI_is_cursor_plan's error handling convention clearer,
and fix documentation's erroneous claim that SPI_cursor_open would
return NULL on error.

Before 8.3 these functions could not invoke replanning, so there is probably
no need for back-patching.

16 years agoUpdate timezone mapping for Windows with new timezones added
Magnus Hagander [Mon, 11 Feb 2008 19:55:11 +0000 (19:55 +0000)]
Update timezone mapping for Windows with new timezones added
in windows servicepacks.
Fix timezone mapping for "Mexico 2"