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

Re: Proposal to help Debian release on time



On Fri, Feb 22, 2002 at 07:40:17PM -0300, Gustavo Noronha Silva wrote:
> On Fri, 22 Feb 2002 19:27:03 +0100
> Jeroen Dekkers <jeroen@dekkers.cx> wrote:
> 
> > On Fri, Feb 22, 2002 at 01:44:14PM -0300, Gustavo Noronha Silva wrote:
> > > the problem is not on Debian's release framework nor in our
> > > Release Manager, the problem is on Debian Developers not fixing
> > > RC bugs as has been said *many times* already
> > 
> > I agree, but having a good infrastructure can help the developers who
> > have to fix the bugs.
> what do you think is still missing? I think the BTS and the pages put 
> together by Debian QA people are very helpful... there are probably
> more useful things to come, but I'm sure they won't be some kind
> of magic to crush RC bugs without looking at the code and fixing it

I actually didn't want to discuss about it before woody is released,
but now you're asking me what to think, I feel I've to answer your
question. But it's your fault we are discussing about this issue
*again*! :)

The current infrastructure is indeed very helpful, it considered very
good even. But that doesn't mean that we can't or shouldn't improve
it.

The first thing to do is identifying where the problems are. Somebody
juse told me to on an IRC channel that Debian should decrease the
number of packages. 20 Ftp-servers aren't needed. However, as long as
it's causing no problems it doesn't matter either. If there are
RC-bugs in ftp servers will just be removed and debian ships with 15
ftp-servers.

The problem this time were that base was buggy and we were waiting for
the lawyers regarding the braindamages US crypto laws. I don't think
you can do anything about the second, and it's a thing which only
happens once. I think that you can do something about the first
problem and there is a chance that we will have the same problem with
woody+1.

I already talked about an idea in an previous mail in this
thread. Giving the base packages multiple maintainers, i.e. having 2
or 3 people in the Maintainer: field. The PTS already helps for
this. I think it does make sense (and probably also was the meaning of
the PTS and somebody is already working on this :-) to use it for
other things now sent to the maintainer.

We could also implement the @packages.debian.org address as a
mailinglist about the package. Now the general discussion about a
packages happens on the BTS I think. This can also help fixing, for
example if the bugreports are automatically send to that mailinglist
and people can easily reply with "this is the same as #12345, merge
it" or "this fix is trivial, the fix is attached". Having those mails
archived somewhere might help a developer adopting a package. Also
more NMUs might happen more as developers feel that they know in what
direction the developer wants the packages to go.

This is just one idea of mine which might solve the problem that base
is in a bad state. It might also help cooperation between the
increasing number of debian developers.

This also needs to be implemented, but I think that will be done
better after woody is released. I personally don't have much time in
my normal interest is in the Debian GNU/Hurd port (but most packages
are just the same so I think it's important for us too), so I've to
say that I'm not volunteering.

I also know about a lot of problems related to the different ports
(especially the Hurd port and the new *BSD ports, but also the
increasing number hardware Debian GNU/Linux runs on) and have a lot of
ideas how to fix those.

Those problems are in a lot of different areas of Debian, i.e. the
packaging system and the infrastructure for the developers. I'm
thinking about writing a document about all those problems, possible
fixes and examples showing the things. There are already a few Debian
developers who want to help me.

This is certainly a woody+1 issue (as the Hurd doesn't release with
woody and the *BSD ports aren't created in the debian archive yet) so
I'm not going to waste your time you could spend on making woody
better and release earlier. But it would be nice if the changes could
at least partially made for woody+1, especially because there is a big
chance the Hurd will release then.

Jeroen Dekkers
-- 
Jabber supporter - http://www.jabber.org Jabber ID: jdekkers@jabber.org
Debian GNU supporter - http://www.debian.org http://www.gnu.org
IRC: jeroen@openprojects

Attachment: pgpWLiAjesm8z.pgp
Description: PGP signature


Reply to: