Whoops ... race condition. Apparently fixed in upstream. All's well that
ends well. Lessons learned and all that.
On Thu, 2007-09-13 at 10:53 -0600, Richi Plana wrote:
> On Thu, 2007-09-13 at 16:44 +0000, Kevin Kofler wrote:
> > Nalin Dahyabhai <nalin <at> redhat.com> writes:
> > > Forgive me for wading in here, but upstream *has* to be where .pc files
> > > show up, and if they don't show up there, we absolutely shouldn't be
> > > adding them to binary packages. I believe this very strongly.
> > But there are actually cases where .pc files are being added in Fedora
> > packages, for reasons such as the upstream foo-config script not being
> > multilib-safe (so it gets replaced with multilibbed .pc files and a wrapper
> > foo-config script which just calls pkgconfig). There are also other reasons
> > for
> > adding .pc files in the distribution.
> > That said, I do think this point needs to be taken upstream.
> That's the point that Nalin is trying to make, :).
> Two things unanimously agreed on:
> 1) *.pc files are important because these contain the options that the
> providing package would know and the dependent package developer
> shouldn't guess.
> 2) A mechanism for providing this (whether it's a *.pc file or something
> else) should be provided by the package developer and that means
> upstream. They should answer the question "How do I develop against your
> Now, there are more immediate concerns like "What to do in the meantime
> because certain packages are waiting?" That's where Ralf and Chris
> should cooperate and Ralf begrudgingly (and quite sarcastically, I might
> add ... It's a good thing I can see the humor in it sometimes ;) ) has
> declared he will provide .. ahm ... OpenThreads-devel-dontuse and
> OpenSceneGraph-devel-dontuse for people to, err, use. That might sound
> silly, but that's the package maintainers last word, ;). Just be glad
> it's not worse, :-p.
> Don't forget the BuildRequires in your spec file.
> Richi Plana
fedora-devel-list mailing list