Christopher Browne wrote:
> 2. Jan Wieck's work on SLONY-1
>
> Many are keen to see the code, but it's not out yet. And it is
> not self-evident that it will necessarily be ready to release in
> conjunction with 7.5 (Not to say it _can't_ happen, but just that
> we'll see the code when we see it...)
>
> It is planned to be implemented in C, so it would presumably be
> more suitable for inclusion in the main code than eRserv. But it
> stands as "hope," and not certainty.
I've just made one major step backward on that. Discovered that my first
thread model was deadlock prone, so I better throw that away instead of
building a lot of code on top of it.
What I currently do is documenting the Slony-I ERD and the new thread
model I have in mind. This will be a document in work, but I plan to
have something readable by the end of this week, latest mid next week. I
will create a mailing list for Slony-I on gborg so we can start
discussing the implementation details.
About the inclusion of a replication solution into the core distributon.
The much I personally would be proud to see this one added, as a CORE
member I do not see any of the replication "things" fit. All of them,
and neither Slony-I nor Slony-II will be any different here, have pros
and cons, none is the "one size that fits all" magic solution. To select
one of the replication projects that high above all the others that it
will be added to the core distribution, it should be really outstanding
and general purpose. I think that Slony in the end will be outstanding,
but only for what it was designed for.
Jan
--
#======================================================================#
# It's easier to get forgiveness for being wrong than for being right. #
# Let's break this rule - forgive me. #
#================================================== JanWieck@Yahoo.com #