On Sun, 2012-06-10 at 04:38 +0200, hasufell wrote:
> Bug #420433 lately introduced the discussion again if and when we should
> support older (deprecated) toolkit versions.
> 
> As for the named bug it may make sense to support it, cause the gtk3
> useflag would lead to different (reduced) functionality of that package.
> (but that shall not be the discussion here)
> 
> Generally I think gtk3 useflags should be avoided and only be a
> workaround during migration to gtk+:3. Optimally gtk+:3 should always be
> forced when available and not leading to major issues.
> On the other hand... if gtk+:3 implementation is broken I would suggest
> to simply force gtk+:2 without any gtk3 useflag. So we have ONE working
> toolkit version.
> 
> Introducing stuff like gtk3 useflag will let users think this is about
> choice, but it's actually not (gtk+:2 is not being developed any longer
> afais).
> 
> Would it make sense to add a tracker for packages currently using gtk3
> useflag, so this will not become a habit and only be a workaround?

The Gnome team's recommendation is to avoid gtk3 or gtk2 USE flags.

For libraries, if possible, try splitting gtk2 and gtk3 support into
different slots (see net-libs/webkit-gtk for an example; the gtk2-based
versions have -r2xx revision numbers and go in slot 2, while the
gtk3-based versions have -r3xx revision numbers and go in slot 3).
Unfortunately, for a few libraries, this splitting is difficult to do in
a sane and maintainable manner, so then a gtk3 USE flag could be the
least bad solution.

For applications, just pick one version of gtk. If a particular version
works better, use only that one (e.g. if building an application against
gtk3 would result in reduced functionality or introduce crashes, or if
upstream calls it experimental, you should probably stick with gtk2 for
now). If the results of building against gtk2 or gtk2 are mostly
equivalent, I suggest only building against gtk3, because gtk2 is
basically legacy code that doesn't get much attention from gtk upstream.

-Alexandre


Reply via email to