[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Wed, 27 May 2009 00:09:54 +0200
Gnus/5.13 (Gnus v5.13) Emacs/23.0.92 (gnu/linux)
So, IMO we need to start pushing towards a release. How about getting a
1.9.1 out soon?
My feeling is that the current feature set in master is ready for a
release. We could just distcheck and that's that. (I tried distcheck
today at work, and it worked fine.)
But there are a few features that would be nice for 2.0, and some of
them would be nice for a 1.9.1.
Here are the features that I know of, in no particular order:
Looks really promising. All the code is there. Unknown implications,
needs lots more testing. Needs to see at least one and hopefully two
micro releases. (1.9.x)
R6RS library support:
No code yet. Julian? :-)
Code is there, needs to be merged in and reviewed
Code is there. I think it needs to go in ;-) But does have impacts
wrt interpretation performance. Will need a couple micro releases.
Code shaping up, needs review and R6RS-libs. Unclear how much will be
in shape for 2.0. Acceptable to ship with only the finished parts.
Basically at this point the features are coming together, and we need
testing and exposure to focus the hackpower on what's important.
My proposal would be a time-based release schedule. We can decide the
Guile 2.0 release date now, and three or four micro releases before
August in Europe is rather slow, and September people are catching up,
so I would propose October 15 for Guile 2.0. Then we can make four micro
releases until then, on the 15th of every month, pulling together
features as they are ready. We can have a one-week or 10-day freeze in
master preceding each release, for stabilization.
What think ye?