Hi Ricardo, Ricardo Mones: >> Please get in touch with the upstream project and suggest they >> port this application to libgtk3-perl. I've personally ported >> a couple Perl GTK+ apps from 2.x to 3.x and it's rather >> straightforward. Upstream for the GTK+ 3 and GObject >> Introspection Perl bindings is responsive and happy to add >> missing bits to the bindings.
> Can we have pointers to these migrations patches/commits/repos? > I think having working examples can help upstream here. Sure, here are a few ones: - parcimonie: 9fe3c6c..a082860 in Vcs-Git - openpgp-applet: df9a2ba^..913921b in Vcs-Git - https://git-tails.immerda.ch/persistence-setup 3fd749e...85248a3 (not in Debian) >> I see that upstream is active so I expect this port should be totally >> doable in this timeframe. > Don't understand what do you want to mean here. Most of the software currently in Debian that depend on libgtk2-perl have upstreams that have been inactive for years, in which case there's little hope that they'll suddenly become active again and port the app to current versions of their dependencies. But clawsker seems to be actively developed upstream, which gives hope :) > Porting it to C would also be doable, and probably more > future-proof, since the C library is probably the last piece to > be removed. I don't understand: Perl bindings for current versions of GTK+ are, and will be supported, both upstream and in Debian. The current version of GTK+ is 3.x. Porting to C is almost entirely orthogonal to the question at hand: what matters most is not the programming language, it's the expected version of the dependencies. Granted, the C library for GTK+ 2.x can't be removed before all bindings for non-C programming are, but porting to C while sticking to GTK+ 2 sounds like a waste of effort, i.e. much more work and less future-proof than simply porting to GTK+ 3. > What's the expected removal date for the libgtk3-perl bindings? If you really mean libgtk3-perl: probably a couple Debian release cycles after GTK+ 4 becomes stable, so I would guess not before 2025. If you meant libgtk2-perl: I intend to bump #912860 to RC severity during DebConf19, then the auto-removal from testing should trigger and remove it and its reverse-deps from testing after a couple months. Then removing from sid can wait until close to the Bullseye freeze: I want to give reverse-deps authors and maintainers as much time as I can to do the porting, while allowing to keep the GTK+ 2 version in sid in the meantime. Cheers, -- intrigeri
