sloppiness (insufficient parenthesization, etc).
It still fails regress test for me, however.

see if it's been updated.  Skip the whole mess if we didn't see any
old pidfile.

NEXTXID isn't really necessary, per previous discussion in pghackers,
but I mulishy insisted we should put it in anyway.  Mea culpa.

space in front of it.  Improve comments a little.

I'm betting the QNX4 port does not either ...

work in BeOS port.

which caused segfault in a data-only retore with BLOBs

slow backend, if we update unconditionally as the code did before).

the new log file with zeroes, only while renaming it into place.  This should
prevent problems with 'stuck spinlock' errors under heavy load.

Generated Changelog since last Beta ...

upgrade version to beta6

Sorry Eiji.

O_SYNC, or O_DSYNC (as available on a given platform).  Add GUC parameter
to control sync method.
Also, add defense to XLogWrite to prevent it from going nuts if passed
a target write position that's past the end of the buffers so far filled
by XLogInsert.

because we need page LSNs stored in the main database to be less than
the current XLOG position.  Hence, generate the new XLOG segment at last
old segment number plus one.

	Provide an extenisible scheme of encoding conversion.
	As the first step, SJIS and BIG5 are supported.
	From now on multibyte people would be happy to use
	this psqlodbc driver.

Eiji Tokuya e-tokuya@mail.sankyo-unyu.co.jp

Fix a bug where cs.execute('select %d + %d', (1, 2)) would get interpreted
as cs.executemany('select %d + %d', (1, 2))

to a join or subselect alias ... cf. Oliver Elphick's complaint 13-Mar.

is still a memory leak, but a little less bad than it was.

I'm not sure that it's really necessary to save insert events when there
are only after update or delete triggers, but certainly it's wrong for
COPY to behave differently from an INSERT query.

does have -o.

bug reports.  ;-)