Re: GNOME 2.11/2.12 targeting GTK+ 2.8 (ie cairo based)

[Luis Villa <luis villa gmail com>]

On 6/8/05, Luis Villa <luis villa gmail com> wrote:
> On 6/8/05, Mark McLoughlin <markmc redhat com> wrote:
> > Hey,
> >         I guess there's quite a few benefits/risks to be weighed up here:
> >
> >   - The benefit of having cool new rendering stuff in GNOME 2.12
> >
> >   - The benefit of being able to use all the other new APIs in GTK+
> >     2.8 for GNOME 2.12
> >
> >   - The benefit of getting all this stuff tested early (i.e. before
> >     GTK+ 2.8 is released, rather than after)
> 
> I'm all in favor of *testing* gtk 2.8 as much as possible now (so I
> think it should be left in jhbuild), but I'm very nervous about
> shipping with it at this point- we all know HEAD does not get as much
> testing as it should, and gtk 2.8 seems like some very big changes
> with potentially huge stability implications. Basically, unless Ubuntu
> breezy starts shipping it, or Fedora makes a very active push to get
> people to use it in Rawhide, I'm very nervous about shipping anything
> we'd call a .0 linking against gtk 2.8.

Oh, and after the last time we did this, the release team swore mighty
oaths to never depend on a released-close-to-gnome-schedule GTK again,
since it jeopardizes our release schedule for something that is less
tested than the rest of the stack and which in many cases isn't widely
used because developers haven't had time to integrate it. I suppose we
could reconsider that, but we did it the last time for the same
reasons Mark listed, more or less. As a result we had to delay our
release and (speaking with my QA hat on) after .0 we still had to
track down several issues in gtk that were caused by rushing out a
component low-in-the-stack with insufficient real-world testing.

So, yeah, I'm pretty strongly against this, though I'm open to persuasion.

Luis