Fedora Extras packaging beta software into production repos, why?

[Christopher Stone]

On 11/29/06, Paul Howarth <paul at city-fan.org> wrote:
> On Tue, 2006-11-28 at 21:56 +0000, David Woodhouse wrote:
> > On Sat, 2006-10-28 at 15:53 -0700, Christopher Stone wrote:
> > > I *think* the point he is trying to make is that some people don't
> > > like 3rd party repositories overriding Fedora Extras without good
> > > reason and some other people dont like Fedora Extras maintainers
> > > adding packages in Fedora Extras without first consulting with the 3rd
> > > party repositories.
> > >
> > > Which goes back to my original idea of having an official Fedora wiki
> > > page that lists 3rd party repositories.  Fedora Extras maintainers
> > > could then check this wiki page to find out which repositories might
> > > already have a package available for the package they want to submit.
> >
> > I don't really understand. Whenever I want a package to be available, I
> > make sure it gets into Core or Extras? Why would I look elsewhere,
> > except in the special case of something that goes in Livna?
> >
> > I'm no more interested in packages from outside Core/Extras/Livna than I
> > am in installing Debian packages with Alien. Why is it interesting?
>
> End users may already be using packages from third-party repos, and may
> have issues if new and incompatible packages of those applications
> subsequently appear in Extras. The Extras packager could save their
> potential users some grief if they took notice of what's already
> available elsewhere and tried to avoid breaking upgrade paths between
> their package and others' (and vice versa). Of course if the third-party
> packager has done something stupid with their package, it doesn't mean
> that the Extras packager should follow suit, but at least as a courtesy
> to users try to avoid breaking things unnecessarily.

It has to go both ways.  I have a package (sdlmame) on Dribble which
is now also available on Freshrpms.  Yet the Freshrpms packager did
not bother to check my package before making his.  Seems pointless to
try and blame packagers for something like this when every packager
from every repo. is doing the same thing...