[Debian-med-packaging] Convenience code copy of samtools and tabix in r-bioc-rsamtools.

[Andreas Tille]

Hi Charles,

On Wed, Oct 30, 2013 at 06:43:03PM +0900, Charles Plessy wrote:
> Le Wed, Oct 30, 2013 at 06:38:39PM +0900, Charles Plessy a écrit :
> > 
> > Ah, and I see libbcf.a as well.  This one can be build from samtools.
> > Quick-and-dirty way would be to put it in libbam-dev, otherwise we need to make
> > a new package with a trip to the NEW queue.
> 
> Hmmm, and I see samtools 0.2.0~rc3 on GitHub, with the following in the README:
> 
>     The original samtools package has been splitted into three separate but tightly coordinated projects:
>     
>         htslib: C-library for handling high-throughput sequencing data

Links:
   http://vcftools.sourceforge.net/htslib.html
   https://github.com/samtools/htslib

>         samtools: mpileup and other tools for handling SAM, BAM, CRAM

Should we apply

diff --git a/debian/watch b/debian/watch
index 3d2e134..c920547 100644
--- a/debian/watch
+++ b/debian/watch
@@ -1,2 +1,2 @@
 version=3
-https://github.com/samtools/samtools/tags .*/v?(\d[\d\.]+)\.tar\.gz
+https://github.com/samtools/samtools/tags .*/v?(\d[-\d\.rc]+)\.tar\.gz


to also observe RC releases from

  https://github.com/samtools/samtools

or should we intentionally leave the watch file as is.  I'm in big
favour of using releases declared as stable by upstream in general -
just wondering what you might think.

>         bcftools: calling and other tools for handling VCF, BCF

https://github.com/samtools/bcftools

> If there is anybody (newcomers welcome !) who would like to contribute on this:
> the more the merrier !

Considering that my MoM offer remains quite unanswered I did lost some
hope that there are so much newcomers just waiting for challenges to
tackle. :-(

What would you think should be the starting point:  Packaging htslib and
bcftools now, wait until it might pass new queue and update samtools
once this has happened?

Back to the Rsamtools package.  I did some experiments to try convincing
the R Build-System to use -lbam instead of its own libbam.a but failed.
For me this automatism that is called is bit mysterious and I have not
found any hook to drop in some alternative library.  Since it obviously
takes a lot of time to work down the whole BioConductor dependency tree
to get r-bioc-cummerbund updated to its latest version (which was simply
the reason for all my BioC work and which turned out way harder than
expected) I'm tempted to push the package to new with some remark that
we are aware about the code copy and will try to find a reasonable
solution in an upgrade / later version of this package.  Do you have any
better idea?

Kind regards

       Andreas.

-- 
http://fam-tille.de