RE: [sigc] libsigc++-2.0 TODO

[martin-ml hippogriff de]

Am 29.10.2003 um 08:33 Uhr haben Sie geschrieben:
> > I think it is time to seriously consider the 2.0 release.
> 
> Please not yet, I see no _need_ to go stable before gtkmm 2.4.0. Is
anybody
> trying to release a stable library or mature application that would
depend
> on libsigc++ 2.0? An API must get more usage before it's frozen.

Well, that's okay with me. However, in spite of the voluminous header
files the API has quite a clear structure and should be settled by now.
The ABI on the other hand is not a problem since there are only a few
functions built into the library and I don't expect these to change any
more.

> However it is good to keep threatening API stability so that people
keep
> checking the API.

Exactly :)

> > From my perspective, it would be well possible to release
> > libsigc++-2.0 in a few days time; let's say next Tuesday.
> 
> Please no. It's far too early and unnecessarily risky.

Let's wait a little longer then ...
I think we can make a kind of "soft freeze", though, meaning that apart
from obvious fixes (hopefully none) the API won't change any more. This
would allow us to release a beta version, at least.

Regards,

  Martin