Hi Julien,
On Tue, May 07, 2013 at 11:15:39PM +0200, Julien Puydt wrote:
> >It is a bit sad that nobody has given a hint to palp when I was calling
> >for help for the metapackage creation[2].
>
> Well, there is already a palp in debian, outdated ; maintained (but
> mostly abandoned) by Tim Abbott. I was supposed to wait for wheezy's
> release before taking over and pushing (well RFS-ing) the latest
> upstream.
It seems my message was not really clear. *Because* palp is just in
Wheezy it should have been mentioned in the metapackages in Wheezy to
advertise it to Wheezy users. I was trying to collect information about
packages that should be mentioned when I wrote [2]. This information
does not necessarily need to come from the maintainer and it is not
related to any specific version number.
> It is available in debian science's git.debian.org corner, as I
> mentioned in my RFS.
Ups, sorry - I simply missed this piece of information.
> >[1] https://lists.debian.org/debian-devel-announce/2013/04/msg00010.html
> >[2] http://lists.debian.org/debian-science/2013/04/msg00013.html
>
> Ah, yes, I saw those pass... but didn't feel concerned since they
> were about wheezy, and I feel the sagemath work is still
> experimental/unstable more than stable/testing.
As said above: I palp seems to be relevant for mathematicans and is in
Wheezy so it should be mentioned in the metapackages. This seems to me
pretty orthogonal to the fact that it is a sagemath precondition because
other users could simply use it without using sagemath, right?
> After all we don't have all pieces yet, and when we'll have all of
> them, I'm pretty sure we'll have to modify them to make everything
> work.
That's no problem.
So I see that palp is in a Blends VCS and I might have a first look at
the packaging wearing some sponsors hat:
debian/control:
I'd regard the description as quite weak. Any package management
tools are not capable to browse the internet and just relying to
two URLs to let the potential user understand what the program is
actually doing is not acceptable. Perhaps the focus on the pure
sagemath precondition is to strong here - each package itself should
feature a proper description to enable users (and also tools like
xapian indexer) find out easily what it is doing.
debian/copyright:
E-mail Address to Upstream-Contact would be nice if known.
debian/upstream:
For any scientific package I would expect some relevant publication.
Please confirm that you have checked that there is no such article
around or provide a proper debian/upstream file[3] (you can also use
the example for scientific references only here [4]).
debian/changelog:
It has been turned out practical to use the target distribution
"UNRELEADED" (instead of unstable) as long as the package is not
yet uploaded.
So far my remarks to the package. Your mail does not really convince me
that you are ad familiar with the Blends concept as I mentioned in my
SoB effort[5] that the first condition is fullfilled. I guess condition
3. will be fullfilled in about one hour after the cron job that
generates the tasks pages was running and I would really like you to
check out this page[6] to learn what I'm talking about.
Kind regards
Andreas.
[3] https://wiki.debian.org/UpstreamMetadata
[4] http://anonscm.debian.org/viewvc/debian-med/trunk/package_template/upstream?revision=11012&view=markup
[5] https://wiki.debian.org/DebianPureBlends/SoB
[6] http://blends.alioth.debian.org/science/tasks/mathematics#palp
--
http://fam-tille.de