Broken deps in FC6 Updates

[Daniel P. Berrange]

On Sun, Feb 04, 2007 at 09:52:56AM -0500, Jesse Keating wrote:
> On Saturday 03 February 2007 18:08, Michael Schwendt wrote:
> > BuildArch: noarch
> > ExcludeArch: ppc
> >
> >
> > [In comparison, the Extras push script examines the src.rpm for the
> > ExcludeArch tag and doesn't push such noarch packages to the excluded
> > target repos. That is something that has been said is done for Core,
> > too. IIRC, either jkatz or sopwith has said that.]
> 
> I'd really like to start discussion again on stop calling these 
> things 'noarch' when they aren't 'noarch'.  If your package doesn't work on 
> other arches, it can't be noarch.  It either needs to no-op on other arches, 
> or not be noarch.  I'm really tired of having to play games with digging at a 
> sourcerpm to figure out what arches a package is suitable for, since this 
> information isn't carried forward in the resultant rpm.  We either need to 
> fix rpm so that this information is carried forward, or just stop calling 
> these things noarch.

Well in the case of 'python-virtinst' the package itself works just fine
on ppc, and being pure python noarch is the right thing. We only had to 
add in the ExcludeArch: ppc  stuff because one of the packages it depends
on (libvirt) is not available on ppc. I don't see that we should switch
the package to be arch dependnant in this case. The build tools should be
able to figure out that a dependant package is not available on ppc and
automatically skip this python-virtinst, whether its noarch or not.

Dan.
-- 
|=- Red Hat, Engineering, Emerging Technologies, Boston.  +1 978 392 2496 -=|
|=-           Perl modules: http://search.cpan.org/~danberr/              -=|
|=-               Projects: http://freshmeat.net/~danielpb/               -=|
|=-  GnuPG: 7D3B9505   F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505  -=|