On Sat, 2009-05-09 at 04:27 +0200, Kevin Kofler wrote:
Adam Williamson wrote:
> (On a tangent - can anyone think of any reason why we shouldn't enforce
> a rule that no package could provide the actual name of another package?
> I can't envisage any scenario where (ab)using Provides in that manner
> would be useful, but I may be missing something).
It can be useful for compat packages. E.g. kdelibs3 has:
Provides: kdelibs = 6:%{version}-%{release}
(and yes, the actual kdelibs also has Epoch 6).
I don't quite see how that works...what does it achieve? Are there
things which can work happily with either the KDE 3 or KDE 4 versions of
kdelibs?
Even so, I think it'd be more correct to have a virtual provides that
both packages provide, but maybe I'm missing the point here.
--
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Fedora Talk: adamwill AT fedoraproject DOT org
http://www.happyassassin.net