Am Donnerstag, den 01.06.2006, 15:41 +0100 schrieb Jonathan Underwood: > On 01/06/06, Konstantin Ryabitsev <icon@xxxxxxxxxxxxxxxxx> wrote: > > In any case, this isn't a contingency we should really be spending > > that much time over, short of potentially developing a system of ACLs > > that would restrict CVS commits only to the actual package owners. > > Would it help this discussion if the technicalities of developing such > a system were put on the table (apologies if this has been discussed > before and I missed it)? The biggest problem probably is: There are plans to switch away from CVS to something else after FC6 (no, that's all I know). So investing to much time in the current system probably is not worth the trouble. I would sleep already a lot better if at least the issue with "hit CTRL +C at the right moment and no commit mail with the changes will be send" would be fixed. But I don't know CVS enough and would be really glad if someone could look into that. I would even sleep really good if there would be a mechanism that checks md5sum's against upstream packages. But that's quite complicated to implement and might be to much overhead. > This discussion would also be useful in the > context of developing a mechanism for having a team of people > responsible for a package, rather than a single owner. We really need that. But that's stalled mostly because nobody in FESCo really works on driving it forward and the proposal from Patrice is still in my Todo-Inbox. :-(( > Do the problems > with the apprach alluded to by Konstantin have their roots in the > limitations of CVS permissions, or are there other issues? I don't know. It was started with the current scheme and I don't know the details why every packager has access everywhere. And it seems a lot of people don't want fine-graded ACLs. CU thl -- Thorsten Leemhuis <fedora@xxxxxxxxxxxxx> -- fedora-extras-list mailing list fedora-extras-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/fedora-extras-list
