dnf replacing yum and dnf-yum

[Radek Holy]

----- Original Message -----
> From: "Jan Zelený" <jzeleny at redhat.com>
> To: devel at lists.fedoraproject.org
> Cc: "Ralf Corsepius" <rc040203 at freenet.de>
> Sent: Friday, April 10, 2015 3:53:39 PM
> Subject: Re: dnf replacing yum and dnf-yum
> 
> On 10. 4. 2015 at 15:29:02, Ralf Corsepius wrote:
> > On 04/08/2015 08:41 AM, Jan Zelený wrote:
> > > On 7. 4. 2015 at 17:53:42, Ralf Corsepius wrote:
> > >> On 04/07/2015 05:07 PM, Kevin Fenzi wrote:
> > >>> On Tue, 7 Apr 2015 08:38:57 -0500
> > >>> 
> > >>> Bruno Wolff III <bruno at wolff.to> wrote:
> > >>>> On Tue, Apr 07, 2015 at 10:22:25 -0300,
> > >>>> 
> > >>>>     Paulo César Pereira de Andrade
> > >>>> 
> > >>>> <paulo.cesar.pereira.de.andrade at gmail.com> wrote:
> > >>>>>    I had also switched back to yum in rawhide due to --skip-broken,
> > >>>>> 
> > >>>>> and
> > >>>>> in a few updates not even needing it (I would first see what is
> > >>>>> broken, and if not something "vital", use --skip-broken), while dnf
> > >>>>> would just fail with cryptic messages. I can keep up if kde or gnome
> > >>>>> is broken, or some other stuff that does not prevent boot and a
> > >>>>> functional system.
> > >>>> 
> > >>>> dnf really does need --skip-broken like support if it is to replace
> > >>>> yum. yum can be a lot faster than the needed work around to get dnf
> > >>>> to work equivalently. I am considering going back to yum in rawhide
> > >>>> rather than continuig to test dnf in rawhide because of this issue.
> > >>> 
> > >>> dnf's default behavior is like yum with --skip-broken already.
> > >> 
> > >> WHAT?
> > >> 
> > >> --skip-broken is a band-aid to work around packaging mistakes and bugs
> > >> and NOT be the default.
> > >> 
> > >> IMO, this kind of behavior is not helpful and therefore should be
> > >> reverted.
> > > 
> > > This behavior is actually helpful, as it doesn't bother users with a
> > > bunch
> > > of broken deps messages they usually don't fully understand (check out
> > > how many of these bugs were filed against yum over the years).
> > 
> > I vehemently disagree: Users having been seeing the symptoms of bugs.
> > Now you are lying and cheating, pretending their systems would be OK in
> > situations their systems are broken (and potentially vulnerable).
> > 
> > I can not see anything helpful in this behaviour and am not impressed.
> 
> Their systems are not broken (dnf does not install the broken packages), the
> repos are broken. I still maintain my opinion that users should not fix
> problems of the distribution chain unless they explicitly don't want to, nor
> they should see them unless they explicitly don't want to.

I believe Ralf meant that the systems are broken in the sense that they don't have the latest version (aka the bug-fixed version) of the given package installed.

> Displaying a list of packages that could not be installed because they have
> some problems might be ok but that's where I would draw the line.
> 
> > > Putting the opinion of myself and the dnf team aside, I'd like to point
> > > out
> > > that the information you want is still available - dnf check-update will
> > > show you all the updates, even those that have broken deps. Running this
> > > command right after dnf upgrade will list you those that could not be
> > > installed.
> > This is  similiarly stupid.
> > 
> > With the dnf behavioral change
> > - dnf needs to inform users about the broken packages by default
> > - dnf now needs an option which does the opposite to --skip-broken
> >    (--no-skip-broken).
> > 
> > I am very sure you'll see a similiar amount of mails related to broken
> > packages as before.
> 
> To be honest with you, I haven't seen a single one in dnf. If you find it, I
> will rest my case.
> 
> Thanks
> Jan

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