[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Regarding unresponsive Debian maintainers

Kevin Mark <kmark+debian-devel@pipeline.com> writes:

> Just a thought. How about setting up an aging system for who can fix the
> bugs. Give the maintainer N time period to act on the bug and then if
> the maintainer can not fix it or will not fix it, other folks who have a
> patch should be able to apply to fix it. if the maintainer feels that
> the patch should not be applyed, there should some authority to hear the
> pros and cons of the issue and arbitrate the result--would that be the
> tech commity, app. manag., RM or ?

I think there are several different levels to this problem.

Bugs with patches that fix the problem but which aren't being applied are
at a very different level than bugs that are just open but that no one
knows how to fix (which are, themselves, at a very different level than
bugs which aren't reproducible by the maintainer).  More aggressive use of
the tags in the bug tracking system could help differentiate.

Note, though, that many of the patches I've seen submitted for bugs are
wrong or otherwise ill-advised.  Either they're bad hacks or they'd create
other problems for other uses of the same program.  Such patches probably
aren't the majority, but they're a definite minority, which means that
blind statistics based on the patch tag won't work.  (Not that anyone is
suggesting that -- it's just something to be constantly aware of.)

Also, I think it's important not to underestimate the usefulness of just
asking the package maintainer directly if one can help, or just going
through the bugs and trying to provide patches for them and seeing if one
then gets a response.  In many cases, packages have lots of bugs because
no one has time to fix them, not because the maintainer is doing anything
wrong.  If no one has time to fix the bugs, orphaning the package isn't
going to help anything; in fact, it will just make it worse by removing
the time the maintainer *was* able to spend on the package.

No procedural method is needed except in those cases where the maintainer
is an obstacle in the way of improving the package, and it's very
important not to leap to the conclusion that that is the case just because
the package has a lot of bugs.

> It seems that folks go MIA for legitimate reasons but the package
> suffers. it seems like allowing someone to come in to fix a package or
> takeover a packages has to deal with the ego of the maintainer.

I really think this is a small percentage of the problems.  The cases
where there are real personality conflicts get the most ink in the mailing
lists, since they're far more likely to turn into big public fights, but I
think they're a significant minority in the set of packages with lots of

Russ Allbery (rra@stanford.edu)             <http://www.eyrie.org/~eagle/>

Reply to: