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

Re: Kudos to High Energy Physics packagers (Re: r3479 - /projects/science/trunk/debian-science/tasks/highenergy-physics)


On Sat, Jul 14, 2012 at 12:48:48AM +0800, Lifeng Sun wrote:
> On 14:21 Fri 07/13/12 Jul     , Andreas Tille wrote:
> > I'd like to say kudos to Lifeng Sun (and probably others who helped out
> > to get a lot of high energy physics applications packaged for Debian!
> > Really great work!!!
> I greatly appreciate Sylvestre for sponsoring my packages.

Yes, Debian Science would be not half that comprehensive without Sylvestre!
> Unfortunately, Herwig++ is also missed and failed to enter Wheezy,
> although I filed RFS more than one month ago.

That's a bit unfortunate.  I did not realised this - perhaps it would
have been a good idea to ping a bit harder.  While my main focus is on
Debian Med I'd heavily step in in case of urgent issues like this.

I have no idea whether this is somehow related to the really wide scope
of Debian Science.  Since some time I have the gut feeling that we
should split some more focussed Blends from the large Debian Science
umbrella which would create some more focussed teams and it might
enhance the communication in these a bit.  We discussed this in the
Debian Science workshop in Grenoble and perhaps Frédéric (CC) might say
a bit about what we did prepared at the Cheese and Wine party ;-).
(Frédéric, I just talked with Yann (CC) from ETRF and I guess we really
should step foreward with this plan.)
> The geant4 packaging stuffs are outdated for the latest upstream
> version due to the large changes in the build system, some nontrivial
> work is required.

I recently changed the debian/changelog file in Git to inject the WNPP
bug and changed the target distribution to UNRELEASED.  I realised that
the control file does not explicitely mention the Debian Science team
as maintainer.  Perhaps this should be discussed - but probably in a
separate thread.
> > Moreover I created a highenergy-physics-dev task because all the
> > packaged libraries have according -dev packages.  For the task
> > highenergy-physics it would be better in principle to add some packages
> > that might contain some according tools / applications because the
> > library itself is nothing a user can actually execute.
> > 
> > Could you revise the task according to this?
> done.

Many thanks for doing so!  This was exactly the needed input that was
needed and I wished more maintainers would work on this as well to make
their packages more popular.  I did some syntactical changes because
each Depends / Suggests needs to be in a separate paragraph but you
can also group those in a komma separated list.

Because I'm not perfectly sure whether I made my point clear in my mail
above I would like to rephrase the library issue again:  You injected
libcojets2-gfortran in the highenergy-physics task.  It is

$ apt-cache rdepends libcojets2-gfortran
Reverse Depends:

Considering this I wonder whether some user wants to install the library
libcojets2-gfortran on its own or only if he wants to use it for
development - the library package might be simply installed as a
dependency of the development package.

An other example is libcojets2-gfortran:

$ apt-cache rdepends libcojets2-gfortran
Reverse Depends:
tillea@sputnik:~$ apt-cache rdepends libgraflib1-gfortran
Reverse Depends:

The user most probably want to use the application point of view it is
dzedit what is in the focus is not the library in the first place.
These libraries are just dependencies of dzedit and will be installed
anyway if you install dzedit (perhaps via the dependencies of the
metapackage science-highenergy-physics).  So I do not really question
your choice to inject the library packages into the tasks but I just
want to make sure that the intention of the tasks became really clear.

> I also added givaro and fflas-ffpack packages to mathematics{,-dev}
> tasks.


Finally I would remind people about the way to add citation / references
in debian/upstream files.  I could perfectly imagine that a lot of these
packages are covered by according publications and we could do the
upstream authors a favour to mention them here.

Kind regards



Reply to: