Re: Maintaining packages properly
[ Moving to -project ]
context: someone proposed a scoring system like this:
> x (= 10?) Important bugs are RC critical
> y (= 25?) Normal bugs are RC critical
On Fri, 20 Mar 2009, Patrick Schoenfeld wrote:
> On Thu, Mar 19, 2009 at 02:41:46PM +0100, Vincent Fourmond wrote:
> > Hmmm... I find that this is a good idea, but, just as you say, there
> > are rather painful side effects if we apply blindly this rule.
> I think that the original idea is better. Some automatic approach sureley
> works better instead of requiring someone to act on some problem that exists,
> because someone obviously is unable to act properly.
> Additional the so-called "side effects" are probably wanted. For several
> 1) The maintainer(s) should really be urged to act on the bugs. Its obvious
> 2) If basic tools (!) have a great number of valid important bugs that
> would be *very* bad. Just because tools are used more often does not
> mean that a higher quantity of bugs is to be ignored.
> The maintenance of standard, important and essential packages is btw. a
> topic on its own. I find it for example highly undesirable to have a
> package which is basically installed on *every* Debian system to have
> about 180 outstanding bugs, with 5-10 (or sometimes more) beeing
> priority important or higher. Its also undesirable to never care about
> wishlist bugs (which basically is what a lot of maintainers do).
Creating such rules serves no purpose. Instead you should work to recruit
volunteers to resolve those problems when they come up.
Take the dpkg package, it currently has 396 bugs open and 13 marked
important among them (and 176 normal). We had more than 500 bugs 2 years
ago and it has been a _lot_ of work to get to this state.
We welcome volunteers to fix more bugs but right now only 2 persons are
actively working on the code. At the current rate, all bugs will be dealt
with in 6-8 years.
What would your rules change to the situation ?
There's a lot of work to do to help responsabilize maintainers concerning
their duties but those rules would not help. I would rather that we try
the "self-assessment" idea that I presented on -qa a few months ago.
Would you be interested to work on that ?
Contribuez à Debian et gagnez un cahier de l'admin Debian Lenny :