unable to read a stats file for reasons other than ENOENT, and having to reset
last_statrequest because it's later than current time in the collector.
Not clear if this will shed any light on the "pgstat wait timeout" business,
but it seems like a good idea in general.

In passing, do some message-style-police work on recently-added
pgstat_reset_shared_counters code.

properly; it's been making that comment uglier with each run.

This resulted in useless extra work during every call of parseRelOptions,
but no bad effects other than that.  Noted by Alvaro.

indexes, rather than printing them twice.  Per my gripe when the exclusion
constraint feature was committed.

corner cases that come up in certain timezones (apparently, only those with
lots and lots of distinct TZ transition rules, as far as I can gather from
a quick scan of their archives).  Per suggestion from Jeevan Chalke.

Back-patch to 8.4.  Possibly we need to push this into earlier releases
as well, but I'm hesitant to update them to the 64-bit tzcode without
more thought and testing.

assertion failure reported by Erik Rijkers, but this alone doesn't explain
the failure.

by joining to pg_constraint.conindid, instead of the former technique of
joining indirectly through pg_depend.  This is much more straightforward
and probably faster as well.  I had originally desisted from changing these
queries when conindid was added because I was worried about losing
performance, but if we join on conrelid as well as conindid then the index
on conrelid can be used when pg_constraint is large.

exclusion constraints.  Not sure how we managed to update the comment for
it in catalogs.sgml but miss this one.

Noted while reviewing a patch from KaiGai Kohei.

instead of an exclusive lock.

The change is almost for code cleanup. Since there seems to be no
performance benefits from it, backports should not be needed.

Fujii Masao

7.4.28.

The latter is considered unwarranted chumminess with the implementation,
and can lead to crashes with recent Perl versions.

Report and fix by Tim Bunce.  Back-patch to all versions containing the
questionable coding pattern.

Samoa, Chile; corrections to recent changes in Paraguay and Bangladesh.

ecpg now adds a unique counter to its varchar struct definitions to make these definitions unique, too. It used to use the linenumber but in the rare case of two definitions in one line this was not unique.

--single-transaction are both used and the failure happens in commit,
e.g. failed deferred trigger.  Also properly free BEGIN/COMMIT result
structures from --single-transaction.

Per report from Dominic Bevacqua

Tim Landscheidt

the GSSAPI libraries crash.

Noted by Zdenek Kotala

requires a hostname to function.

Noted by Zdenek Kotala

Bangladesh, Mexico, Paraguay.

when warning about column-level privileges.  This is more useful than before
and makes the apparent duplication complained of by Piyush Newe not so
duplicate.  Also fix lack of quote marks in a related message text.

Back-patch to 8.4, where column-level privileges were introduced.

Stephen Frost

file instead of ~/.psqlrc on startup.

unless (1) the @ isn't quoted and (2) the filename isn't empty.  This guards
against unexpectedly treating usernames or other strings in "flat files"
as inclusion requests, as seen in a recent trouble report from Ed L.
The empty-filename case would be guaranteed to misbehave anyway, because our
subsequent path-munging behavior results in trying to read the directory
containing the current input file.

I think this might finally explain the report at
http://archives.postgresql.org/pgsql-bugs/2004-05/msg00132.php
of a crash after printing "authentication file token too long, skipping",
since I was able to duplicate that message (though not a crash) on a
platform where stdio doesn't refuse to read directories.  We never got
far in investigating that problem, but now I'm suspicious that the trigger
condition was an @ in the flat password file.

Back-patch to all active branches since the problem can be demonstrated in all
branches except HEAD.  The test case, creating a user named "@", doesn't cause
a problem in HEAD since we got rid of the flat password file.  Nonetheless it
seems like a good idea to not consider quoted @ as a file inclusion spec,
so I changed HEAD too.

error sqlstate. This change makes it return a correct value..

with binary compatibility.

Backpatch to 8.4 where INOUT casts were introduced.

about the behavior.

Document that quotes in to_date, to_timestamp, to_number skip input
characters.

set ferror() but never set feof().  This is known to be the case for recent
glibc when trying to read a directory as a file, and might be true for other
platforms/cases too.  Per report from Ed L.  (There is more that we ought to
do about his report, but this is one easily identifiable issue.)

privileges of foreign servers is "GRANT ... ON *FOREIGN* SERVER ...".

in versions >= 8.3).  The core code is more robust and efficient than what
was there before, and this also reduces risks involved in swapping different
libxml error handler settings.

Before 8.3, there is still some risk of problems if add-on modules such as
Perl invoke libxml without setting their own error handler.  Given the lack
of reports I'm not sure there's a risk in practice, so I didn't take the
step of actually duplicating the core code into older contrib/xml2 branches.
Instead I just tweaked the existing code to ensure it didn't leave a dangling
pointer to short-lived memory when throwing an error.

too, instead of duplicating the functionality (badly).

I renamed xml_init to pg_xml_init, because the former seemed just a bit too
generic to be safe as a global symbol.  I considered likewise renaming
xml_ereport to pg_xml_ereport, but felt that the reference to ereport probably
made it sufficiently PG-centric already.

into proper sections, per suggestion from Tom.

section, throw an error message saying explicitly that the label must go
before DECLARE.  Per investigation of a recent pgsql-novice question,
this code did not work as intended in any modern PG version, maybe not ever.
Allowing such a thing would only create ambiguity anyway, so it seems better
to remove it than fix it.