Re: New openal-soft packages in mentors.d.n
Andreas Barth <firstname.lastname@example.org> writes:
> * Reinhard Tartler (email@example.com) [080608 21:51]:
>> Executive summary: the new package is basically an soname bump, all
>> reverse depends need to be rebuilt. This has been tested, only one
>> package is problematic: boson. We expect that this issue can be handled
>> befor the lenny release.
> Can we do it with an compatible package for the transition? (Which
> means: have both versions available for some time in testing and
> unstable, and transition the packages step by step)
Just discussed on irc:
22:03:26 < siretart> aba: in principle, that should be doable. however, I'm not sure if that makes sense. both packages provide an
'libopenal-dev' package with headers, and I don't think you want to adjust packages to switch from 'libopenal-dev' to
some 'libopenal-soft-dev' header package, do you?
22:03:56 < aba> siretart: just kill the -dev-package from the old package
22:04:04 < aba> so packages will use the new one after being rebuild
22:04:17 < siretart> aba: ah, right. that should be no problem
22:05:56 < aba> siretart: so I would propose: you first upload a version of the new lib into unstable without the -dev-package. After that
migrated to testing, you upload a new version of both libs to unstable where the -dev-package changes source package
22:07:04 < aba> if you do it that way, it isn't a transition we can worry about
22:07:18 < aba> (of course, we can still schedule binNMUs)
22:07:49 < siretart> aba: I could do that, however I see no much point in a library package without any users. I'd therefore prefer to switch
the -dev package before openal-soft reaches testing.
22:07:59 < siretart> however if you say that this is the way to go, no problem
22:08:25 < aba> siretart: because it makes sure that no other package can be blocked if somehow the package FTBFS or whatever else happens.
22:08:35 < aba> it just is "move any potential blocker out of the way"
22:09:16 < siretart> okay, I see. well, first it needs to get out of NEW anyways...
22:10:14 < aba> it currently waits for experimental?
22:10:54 < siretart> yes. because I didn't want to start a new transition without talking to you
22:11:04 < siretart> I can reupload targeting to unstable, if that helps
22:11:05 < aba> yep, good.
22:11:13 < aba> yes, but please without the dev-package
22:11:50 < siretart> okay, will do that
Reinhard Tartler, KeyID 945348A4