Getting collaberative maintainace up (Re: Upcoming NMU)
On Sun, Feb 01, 2004 at 06:34:44PM +0000, Colin Watson wrote:
> > What is needed is then only a policy on who and when and with the
> > consent of who a release can be done. I suggest something like:
> > - after 2 weeks of the latest not-uploaded patch is in subversion, there
> > is branced for a release
> > - testing starts on it, at least two 'changes reviewed & testing OK'
> > reports from maintainers needed, and for example 2 days. Only real
> > bugfixes are accepted, and no veto's for any reason
>
> I think all that's overkill, really; much simpler to find people on IRC
> or mail the mailing list and say you want to do a release whenever it
> seems worthwhile, and then do whatever testing feels appropriate. I
Agreed, I had the XFree 4.3.0 upload in mind while typing it, but indeed
lintian is (relatively) so small and also this isn't about one big new
release, much more incremental development.
Consider that suggestion as withdrawn then :)
> think everybody working on lintian has enough common sense for that to
> work fine.
I do believe and trust that is the case, yes.
> > > Note also that this list is not limited just to Lintian -- with CVS/SVN
> > > commit logs and/or diffs being sent to it, we can improve both tools at the
> > > same time.
> >
> > You are referring to Linda I assume? It'd be best if both could be in
> > the same repository then, as you can then commit one fix/improvemt to
> > linda and lintian in one commit.
>
> The problem with that is that you'd be starting out from very different
> bases and working in different languages. For my part, I know that my
> Python abilities are very weak. Separate commits are probably more
> reasonable.
Advantage is indeed minor currently... but:
<possibly dangerous (for me) rant>
It always stroke me as odd that there exists both linda and lintian.
Isn't it possible to have lintian include arbitrary executables as
tests, so that python-only tests can be simply plugged into lintian, and
then drop linda? AFAICS, the interface could be very simple: just echo
the I/W/E: lines, and keep existing lintian structure.
That would prevent a lot of double work... And easier for the user, as
only one tool remains. And existing infrastructure like the global
lintian check on the whole archive will for free feature the linda
checks too then.
</rant>
--Jeroen
--
Jeroen van Wolffelaar
Jeroen@wolffelaar.nl (also for Jabber & MSN; ICQ: 33944357)
http://Jeroen.A-Eskwadraat.nl
Reply to: