configure.

Regression tests and documentation have both been updated.

SQL2003 requires that both ceiling() and ceil() be present, so I have
documented both spellings. SQL2003 doesn't mention pow() as far as I
can see, so I decided to replace pow() with power() in the documentation:
there is little reason to encourage the continued usage of a function
that isn't compliant with the standard, given a standard-compliant
alternative.

RELEASE NOTES: should state that pow() is considered deprecated
(although I don't see the need to ever remove it.)

Allow additional thread flags to be added via port templates.

Change thread flag names to PTHREAD_CFLAGS and PTHREAD_LIBS to match new
configure script.

contains OIDs. Also, minor documentation improvements to the
psql reference page.

This saves a small amount of per-backend memory for LP64 machines.

Update log_statement to more clearly state it doesn't filter based on
the statement type of the prepared statement.

report socket errors.

Magnus Hagander

1. Fixed bug where sequences were being mirrored incorrectly if they
exceeded 127
2. Fixed a bug in the perl script with mirroring sequences(John
Burtenshaw sent an email to patches describing the bug in March but I
htink he forgot to attach his patch)
3. The dates/times in the transaction files will always use 2 digits.

Steven Singer

completion tag of the actual statement executed.  This allows the
correct update count to be returned for UPDATE/INSERT/DELETE
statements.

Kris Jurka

errors. This is the second submission, which integrates Tom comments about
localisation and exit code. I also added some comments about one sql
command which is not ignored.

Fabien COELHO

builtin aggregate functions.

the next are handled by ReleaseAndReadBuffer rather than separate
ReleaseBuffer and ReadBuffer calls.  This cuts the number of acquisitions
of the BufMgrLock by a factor of 2 (possibly more, if an indexscan happens
to pull successive rows from the same heap page).  Unfortunately this
doesn't seem enough to get us out of the recently discussed context-switch
storm problem, but it's surely worth doing anyway.

of whether we have successfully read data into a buffer; this makes the
error behavior a bit more transparent (IMHO anyway), and also makes it
work correctly for local buffers which don't use Start/TerminateBufferIO.
Collapse three separate functions for writing a shared buffer into one.
This overlaps a bit with cleanups that Neil proposed awhile back, but
seems not to have committed yet.

	FORCE QUOTE to force quotes
	FORCE NOT NULL to quote null input values

< * Add NO WAIT option to various SQL commands
> * -Add NO WAIT LOCKs

    process directly.  Some parameters can only be set at server start;
    any changes to their entries in the configuration file will be ignored
    until the server is restarted.

> * Allow AFTER triggers on system tables

> * -Exit postmaster if postgresql.conf can not be opened

> >> have to accept a full table scan when locating records.
> >
> > It indexes them, but "is null" is not an indexable operator, so you
> > can't directly solve the above with a 3-column index.  What you can do
> > instead is use a partial index, for instance
> >
> > create index i on CUSTOMER.WCCustOrderStatusLog (WCOrderStatusID)
> > where Acknowledged is null and Processing is null;
>
> That's a very nifty trick and exactly the sort of answer I was after!

Add CREATE INDEX doc mention of using partial indexes for IS NULL
indexing;  idea from Tom.

Remove all of the libpgtcl and pgtclsh files, including all references to
them within the various makefiles

with_tcl is still required for the src/pl/tcl language

reference DEALLOCATE in any way. It points to EXECUTE, but not to
DEALLOCATE. Suggested fix:

... This also means that a single  prepared statement cannot be used by
multiple simultaneous database clients; however, each client can create
their own prepared statement  to use. The prepared statement can be
manually cleaned up using the DEALLOCATE command.

James Robinson

of VACUUM cases so that VACUUM requests don't affect the ARC state at all,
avoid corner case where BufferSync would uselessly rewrite a buffer that
no longer contains the page that was to be flushed.  Make some minor
other cleanups in and around the bufmgr as well, such as moving PinBuffer
and UnpinBuffer into bufmgr.c where they really belong.