Re: Maintaining tasks files
Hi Afif,
On Sun, Mar 27, 2016 at 04:09:12PM -0700, Afif Elghraoui wrote:
> > since I started working on the Debian Science tasks files which are in a
> > way worse shape than the Debian Med ones I fired up my script that seeks
> > for packages not yet inside the tasks files.[1] While I added several
> > uninteresting preconditions to the ignore list I also found some packages
> > missing in the tasks files. May be it makes sense if somebody else takes
> > another look on the tasks files. Specifically the bio task is quite
> > crowded. Several packages where inbetween
> >
> > X-Mark: Packages in Vcs - Information about these is queried from UDD as well
> > and
> > X-Mark: The information below needs to be checked whether it can be obtained from Vcs or needs to stay here
> >
> > comments but are now official packages. While there is syntactically no
> > need to move these around having them in sensible groups might make
> > reading the task file more easy.
>
> At least for the latter, I think it would be better if those were filed
> (and tracked in our tasks page) as RFPs instead of written directly to
> the tasks file. I might volunteer to do the filing for the ones I
> consider relevant.
I admit my personal observation is that RFPs are just filling up the BTS
with "random content" without a real chance to end up as a package. To
my experience it needs a volunteer to do the task in the first place who
really intends to do the work (==ITP). Just hoping that people who
usually have no problem to fill their spare time with sensible tasks
will do you a favour and work on a RFP is an illusion.
> > I admit I'm to lazy (and partly
> > incompetent to find sensible groups) but since its one part of our
> > infrastructure I'd be happy if these files would get some more love.
> >
> > Any volunteer?
> >
>
> I'd be happy to help with categorizing the packages, but I'm not sure it
> is worthfile to shuffle them around the bio file if it doesn't make any
> difference to the online view or the structure of the metapackage.
As I said there is no effect on the online view or the metapackage. I
would consider some kind of internal grouping as a preparation for a
possible future split. From my point of view med-bio became to large
and should be split into sensible smaller tasks. Some kind of internal
structure could help to do this. We also have these bio-ngs and
bio-phylogeny tasks that could make some start but these are poorly
maintained. We also have med-cloud which is rather a technical split
but is also not properly maintained since we added a lot o new packages
to bio but the other tasks were not changed accordingly.
> For example, I'm a bit confused by the presence of groupings in the
> tasks/bio file like
>
> X-{Begin,End}-Category: Phylogenetic analysis
>
> while there is also a separate task file named bio-phylogeny. Should we
> rather have bio depend on bio-phylogeny and just put all the phylogeny
> packages in the latter file? I think this kind of division would be
> helpful, as http://blends.debian.org/med/tasks/bio is an extremely long
> page. Are there any strong feelings about this?
I'm in favour of splitting. The internal categories are older than
bio-phylogeny. I was trying to push a bit more into the direction of
splitting by creating the separate task but the maintenance was not
properly done and the bio task fas just filled up (since this is the
brain dead simple but in the end not the most helpful solution).
Kind regards
Andreas.
--
http://fam-tille.de
Reply to: