On Thu, 2006-06-22 at 12:55 +0200, Michael Schwendt wrote: > On Wed, 21 Jun 2006 14:59:19 -0400, Jarod Wilson wrote: > > On Wed, 2006-06-21 at 21:41 +0300, Ville Skyttä wrote: > > > On Wed, 2006-06-21 at 10:27 -0400, Jarod Wilson wrote: > > > > > > > I think the ideal way to handle this is to > > > > build only 64-bit on 64-bit, 32-bit on 32-bit, but make the 32-bit > > > > packages available in the 64-bit tree as well. At the moment, this isn't > > > > possible, > > > > > > It ain't pretty, but it's possible from the repo management POV, we > > > already do copy i386 Wine packages + some dependencies over to the > > > x86_64 repo as part of the signing/pushing process. See "copydict" at > > > http://cvs.fedora.redhat.com/viewcvs/extras-buildsys/utils/extras-sign-move.py?root=fedora&rev=.&view=markup > > > > I suppose that would do as a workaround until something more like the ne > > algorithm used to determine what should be multilib in core gets > > implemented for extras as well... > > Is that algorithm documented anywhere? The basics of it were discussed in the thread on fedora-maintainers from last month titled "Improving the way we select multilib packages for trees". Jeremy -- fedora-extras-list mailing list fedora-extras-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/fedora-extras-list
