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

Re: bugs and patches in Debian

Hi Gabor,

On Tue, May 20, 2008 at 12:04 PM, Gabor Szabo <szabgab@gmail.com> wrote:

> a month or two ago in Oslo we discussed with Jeremiah and others
> that I would like to make it easy for everyone, including the CPAN module
> authors to see which of the CPAN modules are distributed by Debian.

Jeremiah told us a little about this, and I'm very happy to see this initiative!

> I hope that having such metric on CPANTS will encourage CPAN developers
> to make sure they pick up the patches from Debian and fix the bugs reported
> in Debian.


> So we figured out the best would be if there was a cron job on some
> debian server
> generating a csv file with fileds like
> Perl-Package-Name,CPAN Version,Number of patches,Number of bugs,Link
> to its home page in Debian

Nowadays we don't have all that information, but let's examine it:

- Perl-Package-name: we currently have a pointer to the homepage and
the watchfile which strongly suggest this (we almost always use
seach.cpan.org/dist/foo URLs). I played with the idea of having a more
specific metadata available.
- CPAN version: got that from watchfiles.
- Number of patches: the problem here is the many ways of doing this
in debian; but in the Perl Group we have more or less standarised on
using quilt, and the few remaining use dpatch/simple-patchsys which
are quite similar. I guess this can be added pretty easily for us, not
for Debian in general (there are many Perl packages maintained outside
the group).
- Number of bugs: we have it
- Home page in Debian: there's not exactly a home page, but we can
automatically generate links to the various QA pages easily

In general, our qa-report scripts can be extended to grab this data
for the perl group, but as I said before, it'd be a little difficult
to extend them beyond that. Ad-hoc reporting could be done for the
rest, but I don't think it'd be an easy task.

> I guess I am talking about Debian Unstable here but we can discuss this.

I don't think it has much sense to track stable.

> As Jeremiah probably has other things to do now and I'd like to make this move
> forward I would like to ask if anyone on the list could create such a csv file.

I'm all for this to happen, so count me on working on this. As soon as
I have some free time :)
The CSV generating part is trivial, the main work is on capturing the
now missing data.

Martín Ferrari

Reply to: