pidgin.next.minor: f00a8560: Use up the last padding for PurplePlugin...
mark at kingant.net
Tue May 6 04:40:14 EDT 2008
On Mon, 5 May 2008 13:27:51 -0400, Ethan Blanton wrote
> Gary Kramlich spake unto us the following wisdom:
> > The fact that it's been in the vv branch for a while now seems like
> > nothing more than a red herring to me. I mean vv is going to be a major
> > bump. Theres no reason we can't add everything we need, and then add
> > some obscene number of reserved slots.
> I agree that .vv is a red herring.
> > Also, the fact that we're even looking at this, to me, says that we
> > don't want to follow the rules we all agreed to about our versioning. I
> > mean, if we need to add another member and don't have room, is there any
> > reason feature X is that important that we need to work around a our own
> > rules, and that just seems silly to me.
> The thing is, this *isn't* breaking our rules -- our rules are about
> ABI compatability, not "prettiness". I think this is an ugly hack,
> and I don't like it from a principle standpoint, but I don't see any
> reason why it's illegitimate from the ABI side.
I think it's worth it. I'm very much enjoying the stability of our 2.x.y
releases, and I like that we're continuing to refine them. I think this
change is pretty awesome, and I feel like it'll be pretty easy to forget about
this underlying ugly hack and go on living my life like nothing ever happened.
We can make it all nice and purty for 3.0.0 (either by moving the struct size
to the front and using the macro for everything, or by adding a larger number
of prpl callbacks, or doing something fancy with GObjects).
More information about the Devel