On Thursday 07 July 2011 13:42:24 Michael Orlitzky did opine thusly: > > Holy shit, that attitude from Samuli sucks big balls big time. > > > > > > > > He's always come across to me as an OK dev, never seen him pull > > THAT stunt before. > > For what it's worth, I was expecting much worse. In his defense, the > commenters list a bunch of bugs in other packages as the reason why > they want to retain gtk2 support in the gnome-mplayer ebuild. > > Per comment 21, the Gnome team suggests that packages use the latest > version of gtk that works.
Yes, that's "suggests" they use "that latest that works", not "demands", "insists", "mandates" or "requires", and not "only the latest version that works". > The gnome-mplayer package is supported on the alpha, amd64, ppc, > ppc64, x86, and x86-fbsd arches. Adding a gtk2 USE flag means that > the testing load would be doubled; that the maintainer would have > to recompile the package six times on six different machines to > make sure that it runs with gtk2. > > Then, to go stable (in addition to now being tied to the stable > gtk2), the arch testers would have to re-test on all six of those > arches. > > So, the additional burden isn't so small as it's made out to be in > the comments. And what about gnome? Does that not impose a fantastic testing burden, alongside which gnome-mplayer is small in comparison? How about the devs relook at this and do it sanely. When the major consumer of gtk+ (gnome itself) has a stable gtk+-3 very in stable, then other packages follow suit, not before. -- alan dot mckinnon at gmail dot com
