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

Re: About EMBOSS



On Fri, 1 Sep 2006, Charles Plessy wrote:

I still wish that we could ship EMBOSS in Etch; here are a few thoughts
about this.

I would call this very optimistic - but it would be realy great.

First of all, there is the package of Matt Hope. If properly fixed for
the xlibs transition, it builds fine on Etch. However, it is two major
versions lagging behind the current release. Do you think that it makes
sense to upload it anyway to NEW?

In general there might be certain reasons to upload not really up to
date software to NEW.  I do not think that "package does not contain
latest upstream" would be a reason not to upload.

The NEW queue is getting long again,
and I do not know if it means that some packages will not make it in
Etch because they will be stuck in the queue before the freeze.
Uploading emboss 2.8 shortly would solve the problem in advance.

I would make the decision dependant from the fact whether our users
(i.e. readers od Debian-Med list) think that this old version is
useful for their daily work.

In any case, I will try contact Matt Hope today and ask if it is OK if
we take over the packaging. As he seems to be MIA, I do not really
expect an answer, but we never know. One or two weeks later, I will send
a last call on -devel.

IMHO you can save this time.  I tried to contact Matt several times
about EMBOSS and other stuff and never got any answer.  So this part
of your job should be regarded as 'done'.

I will ask him to communicate them, but as I really would like
to create a friendly collaboration with his team, I will not _require_
the sources as the GPL would allow me to do so. In any case, the GPL
also allow them to withdraw the binary package from public download, so
there is nothing to win in direct confrontation.

Well, perhaps they are just lacking time and do not regard the
sources as such important for their clients that they make them
available.  GPL is not really violated if they send you the sources
at request (at least if I'm not misleaded).  So we should not
really go the confrontation path.

Lastly, I have started some preliminary work on dependancies of EMBOSS.
I have obtained from the REBASE copyright holders the permisson to
distribute REBASE as a debian package,

What's the license of REBASE (URL)?

and I have good contacts with the
upstream author of primer3, who is asking his employer (the copyright
holder) to relicence his work under the BSD licence. It was recently
confirmed on -devel that the current licence of primer3 (BSD with
advertisement clause) is officially DFSG-free, mostly for historical
reasons. But this licence is just too easy to infringe by accident by
distributors, so I think that is is wiser to wait the answer from the
copyright holder of primer3 before uploading it to main (there is a
package almost ready made by Steffen Moeller, and I am working on it to
upgrade to the newest version).

That's great.  As the sponsor of Steffen I know that he worked on
this package but his new job seems to occupy more of his time and
thus he did not finished primer3.

Who is intersted to paricipate to the packaging of EMBOSS?

You can see the list of interested people at

    https://alioth.debian.org/projects/pkg-emboss/

and you will notice that I just added you.  Short summary of the
things the other members did:

  Matt:    Inofficial packages, now MIA
  Steffen: Start packaging some dependencies
  Andreas: Creating the Alioth project and bothering Matt

The Alioth project was started after the mail

   http://lists.debian.org/debian-med/2005/06/msg00041.html

hit the Deian-Med list but until now nothing more has been (efectively)
done.

I have guests in the two last weeks of september, so it means that I
would mostly work on this in the next two weeks, and the first week of
october. Then I do not know how the freeze process works: it the
packages have to be in testing on the 18th of october, it means that we
must upload emboss before the 8th. Otherwise, we may have some time.

Well, my personal view is that it makes not really sense to race
for the freece date.  Lets build a reasonable EMBOSS package and
if it will not be ready for stable Etch than people can install
it from testing or we might offer a backport if necessary.  Look,
if our users would have wanted EMBOSS that much from us they would
have done more than sit and wait.  If it is more for us developers
for testing and verifying that everything works smooth the future
testing is the right target.

This schedule is very tight. If there are unexpected difficulties, I
think that it is perfectly OK to release Etch without EMBOSS. But I
would be a bit fustrated if I did not at least try.

Thanks for your effort, Charles.  My short summary is:

  1. If the "working after your patches" package of Matt's old
     packaged version is worth uploading and users of stable
     might be happy about it - just go for the upload (I can
     offer sponsoring if needed, if someone else can do it it
     is even better)

  2. Racing for a freece date makes no sense especially for
     complex packages.  Just start working on its dependencies.
     If these would reach stable we have also won something.
     Use as much time as you need to make good packages for
     EMBOSS and test them in the future testing.

Kind regards

        Andreas.

--
http://fam-tille.de



Reply to: