DNF vs YUM, $pkg, $pkg-mpi, $pkg-openmpi having same provides

[Radek Holy]

----- Original Message -----
> From: "James Antill" <james at fedoraproject.org>
> To: "Development discussions related to Fedora" <devel at lists.fedoraproject.org>
> Sent: Tuesday, June 16, 2015 9:16:27 PM
> Subject: Re: DNF vs YUM, $pkg, $pkg-mpi, $pkg-openmpi having same provides
> 
> On Fri, 2015-06-12 at 15:09 +0200, Kalev Lember wrote:
> > On 06/12/2015 10:28 AM, Radek Holy wrote:
> > > If a package "Requires: foo" and both "bar" and "barbaz" "Provides:
> > > foo", they are handled as being equally suitable. DNF/libsolv is not
> > > going to prefer packages with shorter names.
> > 
> > Yes, very much agreed here. Please don't add the yum shortest name hack
> > to DNF. This and the rest of the depsolving quirks that choose one
> > provider over the other depending on what else happens to be installed,
> > have made my life quite hard over the years :)
> 
>  You would rather have it undocumented and/or random which of the
> providers it chooses ... and this will make your life easier? Ok.

Not random. And the documentation is not needed. All packages providing the given provide are handled as being equally suitable. And that's how it should work. The actual choice is an implementation detail and packages should not depend on it. They don't need to depend on it since they should work the same with all the packages that provide the same requirement/interface. Otherwise the virtual provides would have no sense.

> > With the yum system, yum would pretty much always just pick the shortest
> > name. It _happened_ to be PackageKit-yum, but that was totally an
> > accident. And it made it impossible to change the default without
> > renaming the packages - so in a new release that was supposed to use
> > hawkey, we'd have very little options besides renaming PackageKit-hawkey
> > to PackageKit-h for example to make sure it wins over -yum.
> 
>  That's not true, the obvious change would be to have versions on the
> provide and have the hawkey (default) one be higher. This comes before
> shortest name (which is the last thing yum uses):
> 
> http://yum.baseurl.org/wiki/CompareProviders

Yes, that's the hack which is used by mariadb and community-mysql. From my point of view, such provide is a plain lie. If both packages implement the same version of MySQL interface, the provide should have the same version.

Despite the fact that this "solution" is not very systematic since it requires all the packages to keep this inequality between the provides forever.

> ...although it would also just work if yum wasn't installed by default,
> given the yum backend would bring in a bunch more deps.
> 
>  Although, to be fair, this case is a bit more complicated as you'd want
> to migrate everyone on the -yum backend to the -hawkey backend ... so
> you can't just rely on new install defaults to accomplish that, random
> or documented.
> 
> > What I feel would be a good solution to the problem above would be to
> > have a way to specify the default. I believe this problem is already
> > solved in apt-get with a very nice syntax: the OR syntax:
> > 
> > Requires: PackageKit-hawkey | PackageKit-backend
> 
>  Look at the weak deps. and rich deps. work, which might help here (but
> again, not really in the case you chose).
> 
> --
> devel mailing list
> devel at lists.fedoraproject.org
> https://admin.fedoraproject.org/mailman/listinfo/devel
> Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct

-- 
Radek Holý
Associate Software Engineer
Software Management Team
Red Hat Czech