Re: [Debian-med-packaging] Packaging by students
Hi,
On Fri, Jun 29, 2012 at 09:23:14PM +0200, Laszlo Kajan wrote:
> Hello Andreas, Olivier!
>
> Well the main policy differences seem to me (I have two simple packages accepted so far):
>
> * They do not support Debian Maintainer uploads, they do not use the DM-Upload-Allowed field, so getting a package uploaded involves a
> particular procedure (you set it to unstable from UNRELEASED, then one of the DD team members has a look and either uploads or comments and sets
> back to UNRELEASED).
I personally think that our quite liberal use of DM-Upload-Allowed is OK
so in this point I think it makes sense to derive from their policy.
> * They use git exclusively.
Probably they need to settle to some common standard. IMHO in a
heterogen group like Debian Med you can not force random people to
accept such a choice (and I'm afraid we might lose 2-3 supporters if we
would push for this). Time will show how this can be managed but
for the moment I see no reason to force people into Git when they
prefer SVN.
> * They are sensitive to lintian experimental, info and pedantic tags.
I personally use `lintian -I` and ask people usually to pay also
attention to those infos of lintian if the amount of work needed to fix
the issue has some sensible relation to the problem. This should be
considered with a grain of salt and we possibly are facing other
upstream software than the Perl team who might have some quality
means applied upfront at CPAN.
> Indeed, it only makes sense to follow both their and our (Med) policy, and be sensitive to lintian tags.
I think their rules are quite good but not always applicable in total
strictness - even if we should try hard to meet those standards.
> I do not want to have all my students added to the Perl team as well. If the one that gets the NHGRI::Blastall can work with git, I think I will
> ask to have him/her added to that team and then that pack can go there. It makes sense, because that is likely to be reused by others. The other
> Perl modules are just components of PredictProtein, so I would have those at Debian Med.
Just decide on a case by case basis by keeping in mind what might be in
the sense of getting things working in a reasonable time frame by
gaining for high quality.
Kind regards
Andreas.
--
http://fam-tille.de
Reply to: