Re: Debian Science taking over cimg-dev?
Am Samstag, den 10.10.2009, 17:44 +0200 schrieb Andreas Tille:
> On Sat, Oct 10, 2009 at 03:23:46PM +0200, Daniel Leidert wrote:
> >
> > No, we all now that you prefer hijacks into debian-science over NMUs,
>
> If you define a serious offer of group maintenance as hijack, then yes.
"Taking over" is not an "offer of group maintenance". These both are
excluding: Either forcing the maintainers to join debian-science (JFTR:
the opposite is obviously not ok for you) or letting the package down.
> > the common and correct behaviour (even to increase pressure on the
> > maintainer(s) of a package).
>
> The maintainer were not pressed by long standing bugs and you call the
> offer for group maintenance increased pressure?
You said: "give the maintainers some time to react", which in fact
points to some action to increase pressure on the maintainesr to care
about the package. q.e.d.
> > > I waitet
> > > for confirmation to take over the package, which IMHO should give the
> > > maintainers some time to react.
> >
> > We have NMU queues for this type of situation.
>
> No.
Yes.
> We have NMU queues for situations were maintainers are temporary
> absent but actually intend to continue maintenance. IMHO the situation
> was different in the case of cimg.
It was not. What you tried is not the common way. Even packages which
seem unmaintained for more then 2 or 3 years receive NMUs. And more
different to those: in this case you had someone who already cared about
the bugs: me. You are right with one point: It was different. It would
have required just a check and an upload of the NMU.
> > It would show a much
> > better light on your intentions, if you would use them instead to try to
> > grab more and more packages into your hands. I wonder if you really have
> > the necessary time to handle them.
>
> Reading this lets me wonder whether you are really understanding my
> intention. I'm not Debian Science in person. I try to work on a strong
> team.
There is no QA-team-like structure in debian-science to switch between
hot-spot-packages (to decrease their bug count, to prepare the latest
upstream version, to sort bugs and forward them, ...). So how should
pkg-scicomp or cimg benefit here? Can you disprove me?
[..]
> Cimg-dev is in the focus of Debian
> Med and if I would be able to commit some time to this package if it
> needs work I would do if I would not be forced to join another team.
That is simply not an argument: NMUing doesn't require any commit access
except to the Debian archive. This is, why Vcs-Svn should not contain
the 'svn+ssh' scheme.
> This is what would spoil my time scale.
>
> > > I have no idea what you want to tell me. You prepared an NMU in SVN
> >
> > JFTR: I don't have upload rights for cimg. So checking the build
> > (problem) and/or offering to sponsor the upload would have shown your
> > *serious" interest in this package. You did neither of both!
>
> Well, I checked the build and it builded fine - so I at least did one of
> both - perhaps this makes me half-serious?
Then: Why didn't you adapt the work and uploaded it? Why didn't you
offer to sponsor the upload? Why didn't you send a reminder in >4
months?
> I do not intend to work on a
> lot of NMUs if a package seems orphaned.
You had the chance to ask Sam or Christophe (both are not MIA to my
knowledge) to orphan it. You even had the chance to consolidate your
position with one or two NMUs. You did nothing.
[..]
> > > and accuses me to not have uploaded your work?
> >
> > You said: "that it hurts you so much". So again: Why then you didn't
> > make or upload an NMU (this is the thing I already suggested to you
> > "more than 4 months ago")? Why do you insist in hijacking a package,
> > where the much easier and common solution to fix long outstanding bugs
> > is to prepare/upload an NMU?
>
> That's easy: I'm not seeking for an easy solution which is rather a
> workaround but a real solution. Continous NMUs are no solution - at
> least I not for me.
Nobody was talking about "continous NMUs". At the time of writing we
were talking about the same 4 open bug reports which are there now. So
exactly one would have been enough. And during this time you could have
asked for O/ITA cimg too.
[..]
> > This would have even shown, that you care
> > about the package. Instead you didn't care about it nor touch
>
> I have no commit permissions in pkg-scicomp - so I was unable to touch
> it.
That is simply not an argument. I already told you somewhere above. NMUs
don't require commit access here. It's up to the maintainer to
incorporate your NMU into the subversion tree if he wants to do this.
You are further emphasizing team maintenance: NMUs are the mother and
the father of the team maintenance effort: packages are maintained by
the whole group of DDs (and by sponsoring uploads also by all non-DDs).
[..]
> I do not see the profit
> neither for cimg nor for the involved groups if we start a flame about
> right or wrong behaviour in this case.
I disagree. (a) You could have solved the whole issue several months ago
yourself. (b) Christopher answered you recently with "I claim that we
have done a great job at packaging and disseminating scientific
computing software.". You should ask yourself, why your mail implies the
opposite, so he has to say that? You did the same with several other
groups in the past, including debichem. I already said it in the past
and I repeat it: Your behaviour is rude and insulting to everybody, who
has worked to give scientific software in Debian the significance it
has.
[..]
> > I would really be interested in the answer.
>
> If your tone stays that agressive not any more but for this time yes.
That's interesting. Maybe some reasons you should think about:
- you re-opened a >4 months old discussion with me without CCing me
- you obviously construct arguments (against NMUs)
- you make claims about the quality of debian-science and other of your
groups I see clearly disproved
Debian has a very simple problem: The count of source packages increased
by around or even more than 20% in less than 1 year (from <25000 to
>30000). This creates issues in lacking manpower. And you don't solve
them, but you blame others to do a bad job. That's simply not ok [1].
You failed in creating a team, who can address this issue. That's my
point of view. If you can proove me wrong, don't hesitate to do so.
[1] And in the case of cimg I've uploaded the local fixes as promised. I
might have missde to care about the build issue - I cannot remember ...
too much happened with a lot of success for several other software
packages in Debian - but <ou did not care enough to upload the NMU or
care in any way about cimg. It simply appeared on your radar again
recently and now you again blame others.
Daniel
Reply to: