Re: Ideas about allowing Co-maintainer
"Marcelo E. Magallon" <mmagallo@debian.org> writes:
> On Thu, Aug 14, 2003 at 10:04:18AM +0200, Goswin von Brederlow wrote:
>
> > I have a crazy idea about when/how to allow co-maintainers for
> > packages and this probably becomes one big flamewar. But here we go:
>
> My idea is even crazier: what about if we stop behaving like children,
> and finally admit that if we are going to move fordward with this
> project we have to do something unheard of, like, working with each
> other instead of against each other?
>
> Some people don't like NMUs because sometimes they come at inconvinient
> times (like in the middle of a packaging rework) or sometimes they
> cause more work for the maintainer (the road to hell is paved with good
> intentions). The solution for that is easy: if you can afford it (and
> many people in this project can), keep a CVS with your packaging work
> on a machine that anyone can reach. Some time ago I moved my wmaker
> CVS from my box to klecker. If you are in the "Debian" group, you can
> make commits to this repository.[0] That way the maintainer work is
> not disrupted and you can coorperatively contribute to improving
> packages.
>
> > Some maintainer have packages with lots of bugs but are very
> > teritorial about the package. Getting them to allow NMUs or offering
> > them to become co-maintainer just ends in them screaming obsceneties
> > back (overstated for dramnatic purposes).
>
> That's because people have had bad experiences with NMUs. Sometimes
> the maintainer is to blame, sometimes the NMUer is to blame. But
> there's no point in pointing fingers. Think of libc6 for example. I
> don't follow the mailing list (neither Debian's nor upstream's), and I
> really have no clue how that package is managed. What if I spot a
> simple yet RC bug and decide to fix it by uploading with a delay of 10
> days? I mean:
>
> -- GOTO Masanori <gotom@debian.or.jp> Sat, 19 Jul 2003 00:37:11 +0900
> -- Jeff Bailey <jbailey@nisa.net> Tue, 15 Jul 2003 14:35:58 -0400
> -- GOTO Masanori <gotom@debian.or.jp> Sat, 19 Apr 2003 22:01:40 +0900
> -- GOTO Masanori <gotom@debian.or.jp> Tue, 18 Mar 2003 00:04:13 +0900
> -- GOTO Masanori <gotom@debian.or.jp> Wed, 26 Feb 2003 18:44:08 +0900
>
> That's a sure sign for a bunch of lazy morons, isn't it? Five uploads
> in the last six months? Come on! And just look the bugs for glibc!
> There are some _trivial_ things in there for godness' sake! I'm going
> to make my upload with a 10 day delay, if if the maintainers care about
> the package they have _10_ _whole_ days to react. What harm could that
> cause?
If you contact the glibc team you will get a response out of them
within minutes. If you offer to join and help in that won't be a big
problem. glibc isn't a good example for your case.
> > I propose that when a package has <X> open bug (or RC bugs) for a
> > reasonable long time without activity anyone can just declare
> > themself co-maintainer and do the work.
>
> And that will only buy you bad blood. You might be fixing things in
> the short term but in the long term it's just not going to work.
I consider co-maintainership to be better than hijacking and some
things just don't seem to be working without some bad blood.
> > That way 3 5 year old bugs would be as bad a 50 3 month old ones.
>
> (JFYI, a year has 12 months)
>
> > Let the flamewar begin.
>
> That's some attitude. You'll get really far that way,
>
> Marcelo
>
> [0] Let this be an open invitation to anyone who cares about the wmaker
> package to feel free and fix whatever needs fixing. At the moment
> I don't have any non-commited work at home, whatever you find in
> the CVS repo is what's current. A couple of releases are not yet
> in the repository because I had expected to finish the work in
> progress a bit sooner, but RL got in the way. As long as you tag
> and properly document your work, I'm game. Just drop a note to
> wmaker@packages.debian.org if you want to start something big, to
> avoid duplication of work and stepping on each other's toes.
That is realy good and I would love to have that or something similar
for all packages.
Maybe instead of co-maintainership the creation of an alioth project
with the same name could be suggested policy or written down as common
practice for badly maintainer packages where people want to join in.
Having it written down would everyone know what to expect, what to do
and prevent anyone from complaining too much when its done.
MfG
Goswin
Reply to: