Axel Thimm schrieb: > On Sun, Oct 15, 2006 at 12:24:51PM -0500, Jason L Tibbitts III wrote: >>>>>>> "PD" == Patrice Dumas <pertusus@xxxxxxx> writes: >> PD> After some thinking and looking at some packages, I came to the >> PD> conclusion that having upstream as primary maintainer in fedora >> PD> should be avoided if possible. >> I think you've made a distinction between "primary" and other >> maintainers that does not exist. All maintainers get access. Does it >> really matter who submits the package for review, or which addresses >> appear where in owners.list? Does the new package database even >> attempt to prioritize owners? > > Is this the upcoming model of co-maintainers? Well, that's not written down anywhere and I'm glad it gets discussed. Maybe a bit early, because the technical framework for co-maintainership is still far away.... Anyway: > I'd prefer the model > Patrice assumes, e.g. a primary one and secondary co-maintainers that > *should* coordinate their actions with the primary one. Otherwise > suddenly all contributors become co-maintainers of everything and > we'll get trouble keeping it all in non-chaotic state. I currently prefer a middle ground: There is one primary maintainer -- normally the one that got the packages imported. Both primarly and co-maintainers are free to commit small changes without telling the others (they maybe should wait 24 before building so the other maintainers can veto things). Medium sized or big changes need coordination between all maintainers -- e.g. also the primary maintainer should also ask his co-maintainers for permission (e.g. wait at least 72 hours before building after importing and/or send the others a mails announcing the changes). Well, those are only some rough ideas, but you'll get the idea. But it also depends on how the maintainers of a particular package want to handle it. If one primary maintainers wants full control over his package: okay, then let him if there aren't five other well know packagers that are more friendly. But I don't like such a behavior to much, and it should be the exception and strongly discouraged. > To get back to upstream vs fedora experts maintership: Assuming the > co-maintainership model would be indeed hierarchical, I agree with > Patrice, better to have someone knowing the details in fedora doing > the packaging (in consulting with upstream), than having the package > experts trying to teach upstream how to package. Well, I see no problem to give some upstream developer the primary maintainership of a package in Fedora *if* the upstream developer actually is really interested in Fedora (e.g. he uses it and is aware of the Packaging Guidelines). But that's often not the case. > I guess before considering the relationship of upstream and package > maintainers closer one would need to see what the real model of > co-maintainership will look like. If this has already been decided on, > is there some pointer to wiki/mail that explains it? Thanks! Some stuff is at: http://www.fedoraproject.org/wiki/Extras/Schedule/Comaintainership Cu thl -- fedora-extras-list mailing list fedora-extras-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/fedora-extras-list
