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

Re: Watching mentors and nifticlib



On Sun, 11 Nov 2007, Michael Hanke wrote:

[ not sure whether the original post should have been posted to
debian-med instead of debian-devel. CC'ing it for now, but I guess we
should move there... ]

The posting was intended to Debian-Med list, because we try to enhance the
observation of packages that are interesting for this CDD.  The techniques
should be made available for other CDDs as well in case we consider the
stuff as useful - perhaps it is pure discussion.

I like the idea, but it would only help solving the problem, if there
are actually some DDs reading it and acting accordingly.

Sure. The sentinel we are building for Debian-Med relevant packages
is infact meant to be read by Debian-Med interested developers - that's
the reason we are developing it. ;-)
This is the explanation why I posted it on Debian-Med because we are
focussing on a subset of packages to pick out the relevant packages
from mentors.  A posting on debian-devel would have been apropriate
if I would have had a suggestion to make an enhancement for general
enhancing handling packages on mentors - which I just have no idea about.

Indeed, I have a very reliable sponsor. In fact he is more than a
sponsor and (at least to some degree) we group-maintain all of 'my'
packages.

Great.

In general I think that packages should not be uploaded without a formal
request. Some people (including me) use mentors.d.n. to provide
unfinished packages for sponsors and beta-tester. Having such a package
uploaded could introduce unecessary, because already known, bugs to
Debian unstable.

Yes, this is exactly the reason why I did not even downloaded your package
from mentors - because I expect you to have a sponsor if you don't ask
explicitely for sponsoring.  The problem is, that I have no chance to
*verify* that there is a sponsor working on this.  That's why my idea
was to list the sponsor in addition to the maintainer in the task
file to enable others to ping the sponsor if needed.

BTW, Michael, what do you think about pushing your packaging stuff
of nifticlib into our Debian-Med SVN?
We switched to Git some time ago and all my packages are available from
git.debian.org. The control files should list the corresponding
repository. If one package is not there (yet), it is because no
upload/update was necessary since I started moving. The next upload will
make the package repository available as well.

I admit that I have the strange feeling that the pure existence of several
different repository management systems makes the world a worse place than
the existance of different editors (emacs/vim/...) or desktop environments
(Gnome/KDE/...).  The later ones are pure personal decisions but different
repository systems that are intended to let different people work together
fail in doing so because different people tend to have different habits.
So even if I would adopt your preference of git over say svn and would move
my packages to git other DDs prefere something else and we will finally
not be able to have a common repository of the Debian-Med stuff for
instance.  I admit this is really disgusting.

Because I use Git + git-buildpackage and a repository layout different
from the debian-med SVN I don't want to push things to it. But given the
distributed nature of Git, it should be equally easy to get the package
sources and provide patches for it.

Well, the idea was to form a group maintaining Debian-Med relevant stuff
in one repositore.  This does not mean I would try to convince you to
drop git and switch to svn.  This is rather a cry for help to find a
way to easily enable every interested person easily to cope with changes
in packaging stuff od Debian-Med packages.  David has written a great
tool that lists SVN commits:
   http://debian-med.alioth.debian.org/
but all changes outside SVN slip through this.  Do you see any chance
to cover "everybody pet repository" which this kind of tool?

Kind regards

           Andreas.

--
http://fam-tille.de



Reply to: