Re: Plans regarding GTK+ 2.4 and G-Wrap TNG/2.0

[Andy Wingo]

On Mon, 2004-03-29 at 12:38, Andreas Rottmann wrote:
> Andreas Rottmann <address@hidden> writes:
> 
> > Hi!

Hey bro. I am pleased to announce I have a (hopefully) stable internet
connection! So, I'll be joining you in this endeavor :)

> > I've come pretty far in converting guile-gobject to G-Wrap TNG/2.0[0],
> > leading also to a bit simpler support code (gw-spec-utils and
> > defs-support).

Sweet. Those files were essentially the first things I wrote in Scheme,
so there's quite a bit of nastiness. Any changes are welcome.

WRT TNG: I don't care at all whether we use the old g-wrap or the new
one. I'm sure your code is better (the old record system sucked). My
priority is releasing something, as I'm sure you know. If you can
release, on your own, the TNG stuff, and maybe get permission for it to
be the next version (so that it can go into debian), why not release
with the new? But if that's going to take forever (it shouldn't?), then
release with the old stuff.

I'm going to update to 2.4 now. I see no reason for supporting 2.2 at
this point.

> 1) Cover the GNOME 2.6 API

Perhaps only nominally. Covering all of the API can be done over time,
as long as we're set for inclusion in the 2.8 platform bindings.

> 2a) Release this as one big pile of code (guile-gobject)
> 2b) Do 3) first

Release something before the split, so I can have a proper dependency
for my apps.

> 3) Break the stuff up into modules, like proposed in [0].
> 4) Join the Platform bindings for 2.8. We have to be real quick if we
>    want to achieve this, as the first tarballs are due in April for
>    the 2.8 platform bindings.

These are together.

> Also, I really should have G-Wrap 2.0 ready before doing 3), since I
> really would like to have a unified codebase before doing the split.

Agreed.

BTW, gst-guile stopped handling multiple API versions a couple of months
ago (when HEAD gst started to churn a lot).

Regards,
-- 
Andy Wingo <address@hidden>