DNF as default package manager

[Jan Zelený]

On 21. 1. 2015 at 13:42:01, drago01 wrote:
> On Wed, Jan 21, 2015 at 1:31 PM, Peter Robinson <pbrobinson at gmail.com> 
wrote:
> > On Wed, Jan 21, 2015 at 11:44 AM, Vít Ondruch <vondruch at redhat.com> wrote:
> >> Dne 21.1.2015 v 12:34 Peter Robinson napsal(a):
> >>> Are we expected to cross referencing previous logs to see if there's
> >>> changes or if it's the same and provide you that information? We
> >>> already have too much to do so it's easier to stick with yum where we
> >>> know what the outcome is. Sorry, not going to do your work for you!
> >>> Peter
> >> 
> >> I'd expect that if we are speaking about DNF as default (and it was
> >> approved by FESCo), that releng do scratch mass rebuild of all these 18k
> >> packages built using DNF and give us list of failed packages. What are
> >> these failures is not your concern but package maintainers concern.
> > 
> > If they build with yum why is it a bug in the packaging?
> 
> Because the deps expressed in the package apparently really on a
> specific implementation detail of yum which
> is just wrong so:
> 
> 1) There is bug in dnf as you state
> 2) Assuming the package set that dnf resolves satsify the expressed
> deps it *is* a bug in the package.

Exactly. Over the years many packages have set their dependencies to utilize 
some specific parts of yum depsolving algorithm to get a specific result. In 
these cases it's the packages what needs to be fixed.

I'd like to stress that depsolving problem is NP complete. The fact that DNF 
offers a different solution doesn't mean the solution is wrong. It might be in 
some cases but it shouldn't be automatic assumption.

Thanks
Jan