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

Re: Fwd: RFS: arrayfire/3.0~beta-1 [ITP] -- High performance library for parallel computing





2015-05-26 14:35 GMT+01:00 Andreas Tille <andreas@an3as.eu>:
Hi Ghislain,

On Tue, May 26, 2015 at 11:34:18AM +0100, Ghislain Vaillant wrote:
> 2015-05-26 10:07 GMT+01:00 Andreas Tille <andreas@an3as.eu>:
>
> > If your master branch
> > does not contain a debian/ dir it can not parse the metadata.  I'm
> > not comfortable with the tag based packaging and whether there is
> > some safe way to fetch the debian/ dir from a raw Git repository.
>
> Assuming more and more people, like myself, adopt DEP-14 for their
> packaging repository
> layout, then it would make sense to update this code so that it fetches the
> metadata not only from
> the master branch but from a potential debian/sid or debian/experimental
> branch ? I don't know
> enough about Git scripting to comment on whether this could be easily done.

I guess this will be doable if the number of branches is restricted to
these two as a fallback if there is no debian/ dir in master.

However, what I really like to see is a discussion in Debian Science to
update our policy document.  It is far from beeing up to date and if new
workflows gain traktion amongst developers this should be mentioned
there.  I see this a way more urgent issue than whether a hand full of
metadata of our packages are not properly read and displayed on the
tasks pages.


I could not agree more. The policy badly needs an update IMO.
 
> This situation would happen regardless of whether pristine-tar is used and
> DEP-14 covers both
> pristine-tar and upstream-tag based workflows. I foresee the same situation
> happening for at least
> two other packages I am currently working on.

Which is fine for me - but as I said we should discuss this in Debian
Science and I wonder whether you would like to start a discussion of the
policy (which I personally do not want since I'll be not regularly
online until mid of June (which also means for you that SoB will be
delayed (heavily?) for the next four weeks.

I am happy to start this discussion and guide us through to a revision of the Debian
science policy which is more inline with current packaging practices.

Shall I start it now though or wait until your return mid-June ? I did not understand that
part.

Cheers,
Ghis 

Reply to: