[Debian-med-packaging] Thoughts on the Vienna RNA package

[Steffen Moeller]

On Wednesday 14 November 2007 16:47:09 David Vernazobres wrote:
> On Wed, Nov 14, 2007 at 02:02:03PM +0100, Steffen Moeller wrote :
> > On Wednesday 14 November 2007 12:02:04 Charles Plessy wrote:
> > > Hi all,
> > >
> > > I had a look at the Vienna package, and here are a few thoughts about
> > > the work to be done.
>
> Cool, that one of my day to day libraries....

Fantastic! Please be so kind to crosscheck what I have done.

> > > - It uses a patched version of Readseq. Maybe we should consider to
> > >   build Vienna on Debian's readseq, possibly by adopting its patches if
> > >   they are not disruptive. By the way, we could incorporate the Debian
> > >   package for readseq in our repository, if its current maintainer
> > >   agrees.
> >
> > I do not have any preference. Redundancy should be avoided. This is also
> > why I decided for the Vienna RNA package to include the RNAforest
> > application from Bielefeld rather than packaging the Bielefeld version.
>
> I will be also for packaging the application in one apt-get
> (rnavienna-all).

Not vienna-rna? We can do a split at any time. I am not worried at all about 
it.

> > > - The Vienna package provides libraries that other programs use. For
> > >   instance, MiRanda, a sofware to discover targets of micro RNAs uses
> > >   Vienna's RNAlib.
> >
> > yip, it is an essential package. I did not have a look at the libraries
> > yet. Their "make install" needs some fiddling.
>
> Well, the source code also....
> I have a series of patch that I am going to submit to the mainstream as
> the code is not compiling without gcc. (Yeah, I know, but some time on
> cluster you need to use icc or equivalent).

Fine. I'll prepare the use of quilt then.

> > > - Emboss provides some wrappers so we shoud probably package them as
> > >   well.
> >
> > Nice.
> >
> > > - There are some build issues such as the Makefile in the Kinfold
> > >   directory that has no distclean rule. Probably some autoconf mastery
> > >   can solve the problem.
> >
> > Which is beyond me. There are some static libraries being produced, which
> > is also just slightly disturbing.
>
> The RNA Vienna Libraries is a static only lib. So you will need some
> more polishing for also providing the dynamic lib as a Debian policy
> compliant package.

I was afraid someone would confirm that.

> > > - The vienna-rna source package should probably be split into multiple
> > >   binary packages, for instance RNAforester which is a distinct work.
> >
> > Right. It comes from the authors auf RNAhybrid.
>
> Hum, in this case, took the RNAlib v1.4 as a base (rnavienna-base), and
> add a package for each stuff added inside (rnavienna-forester,
> rnavienna-kinfold, ...) and the mandatory rnavienna-lib, rnavienna-dev.

These are secondary issues for me. It sounds very right what you are 
suggesting.


Best,

Steffen

-- 

Dr. Steffen Möller
University of Lübeck
Institute for Neuro- and Bioinformatics
Ratzeburger Allee 160
23538 Lübeck
Germany
T: +49 451 500 5504
F: +49 451 500 5502
moeller at inb.uni-luebeck.de