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

Re: Keeping a record of programming language in upstream metadata



Hi,

On Thu, May 24, 2012 at 04:30:29PM +0800, Paul Wise wrote:
> I think the implemented-in::* debtags should suffice and already
> exist. I guess it is hard for people to remember to debtag their
> packages, hopefully the PTS links to the debtags site will help. If
> you want that to be determined automatically, probably reviving DACA
> is the way to go since ohcount determines which languages are used by
> a source package:
> 
> http://qa.debian.org/daca/ohcount/

Thanks for these in principle interesting hints but both are not
helpfull for my specific use case because they work only for packages
which are just inside Debian.  My plan was to have the information for
packages which are in Vcs but not yet in Debian (so called prospective
packages in Blends terminology).  There are neither DebTags nor the
ohcount statistics for those packages available but there is some work
in VCS and there is a WNPP bug.  The later has the language information
(but I do not intend to parse the BTS text - at least not for the
moment) but I do not see a reasonable way to move the language
information into packaging information in VCS.

Perhaps the following example makes my point more clear.  If you look at

   http://debian-med.debian.net/tasks/bio#pkgvcs-debs

all data of this section is drained from information in Blends VCS by
only giving a package name as key (in the usual debian/control syntax
  "Depends: pkgname"  )
I would like to have access to the last missing information bit which
might be of interest and debian/upstream could be a reasonable way to
do this.

Kind regards

      Andreas.

-- 
http://fam-tille.de


Reply to: