Re: [HACKERS] Features for next release

Thomas Lockhart wrote:

>
> Hmm. I see a problem coming up for myself :(
>
> I'm working on JOIN syntax and (hopefully) OUTER JOIN capabilities for
> the next release. That will take a lot of time afaik. We also have
> lots of great new features (almost certainly) coming up, such as WAL
> and RI. Lots of stuff, and taken together it should lead to a v7.0
> release.

What I actually see when starting/stopping the postmaster
lets me think that Vadim alread has the WAL target in sight.

And for RI I can say that we'll have at least FOREIGN KEY for
MATCH FULL without PENDANT in the next release. The key
features are in place and there's plenty of time left to do
all the stuff around.

>
> However! I've been waiting for v7.0 to do internal remapping of the
> datetime types (and possibly other types; I'm not remembering right
> now) so that the 4-byte TIMESTAMP becomes, internally, what is
> currently DATETIME and so that the current DATETIME becomes a synonym
> for TIMESTAMP. If we are releasing v7.0 rather than v6.6, then I
> should work on that now to ensure it is completed. If we might release
> as v6.6, I should continue to work on OUTER JOIN, since the type
> changes shouldn't go in yet.
>
> Any opinions/comments on the right path to take? With code changes and
> docs changes I would think that the datetime stuff is two or a few
> weeks work with my off-work time, which I suppose wouldn't impace a
> release schedule too much...

Break TIMESTAMP right now. IMHO it would be THE reason (a
functional incompatibility to v6.5) we where looking for to
be 100% sure the next release must be v7.0.

I clearly vote for v7.0 - even if I personally hoped we would
have a little chance to produce the ultimate devils
PostgreSQL-v6.6.6 someday.

Apropos devil: tomorrow Braunschweig Lions vs. Hamburg Blue
Devils here in Hamburg - German bowl! Go Devils go!

Jan

--

#======================================================================#
# It's easier to get forgiveness for being wrong than for being right. #
# Let's break this rule - forgive me. #
#========================================= wieck(at)debis(dot)com (Jan Wieck) #