Bug#727642: [tg@mirbsd.de: Bug#727642: dose-distcheck: fails on input that works with edos-debcheck and is "correct"]

[Ralf Treinen <treinen@pps.univ-paris-diderot.fr>]

This concerns debian bug 727642.

On Fri, Oct 25, 2013 at 12:37:28PM +0200, Pietro Abate wrote:

> As you can see there is a conflict in m68k%3alibc6 with package
> m68k%3aunscd < 4 . This conflict comes from the Breaks conflict with
> nscd (<< 2.17) . Since unscd provides nscd (all versions), the two packages are
> indeed in conflict. I would say edos-debcheck is wrong.

I don't see why. As I understand it, a versionend break should apply only
to real packages and not to virtual packages, as it is also the case with
Dependencies. The Debian Policy says in Section 7.5:

  If a relationship field has a version number attached, only real packages
  will be considered to see whether the relationship is satisfied (or the
  prohibition violated, for a conflict or breakage). In other words, if a
  version number is specified, this is a request to ignore all Provides
  for that package name and consider only real packages. 

OTOH, the action of dose-distcheck seems to be specific to Multi-arch=same -
If I remove the multiarch from the libc6 package then no conflict is
detected. Is there a reason for doing that exception?

-Ralf.