[Debian-med-packaging] Bug#813262: python-biopython: FTBFS: ApplicationError: Non-zero return code 255 from 'phyml -i Phylip/interlaced2.phy -d aa'

[Andreas Tille]

Hi Peter,

On Mon, Feb 01, 2016 at 11:25:13AM +0000, Peter Cock wrote:
> I spotted this via an automated @DebianBug Twitter posting:
> https://twitter.com/DebianBug/status/693569490655776768
> https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=813262

Thanks a lot for watching!

> Andreas has dealt with other issues like this in the past, but
> I wanted to make this explicit on the issue tracker:
> 
> This isn't a problem in Biopython itself,

Yes.  BioPython is OK - I need to reassign the bug - probably to
libhmsbeagle-dev which is possibly not featuring the correct
dependencies.  I just wanted to investigate the issue before I
do the reassigning since for the moment it is just the BioPython
*package* *in* *debian* that fails.

> but a Biopython test
> case calling phyml indicates there is a problem with the Debian
> phyml package.
> 
> Should the phyml GPU problem be filed under the phyml Debian
> package, perhaps adding a simple basic check to its tests?:
> 
> phyml -i Phylip/interlaced2.phy -d aa

I agree that we could add this check even to the phyml package.
 
> This sample input test file is available here, also used as an example
> in the EMBOSS suite:
> 
> https://github.com/biopython/biopython/blob/master/Tests/Phylip/interlaced2.phy
> 
> I suspect any PHYLIP input file would work for checking the
> binary works and might help catch this apparent compilation
> issue earlier. Looking over the upstream Makefile it was not
> obvious that there is any testing like this being done at the
> moment, but there are example files likely to be useful for
> adding such a test:
> 
> https://github.com/stephaneguindon/phyml

My plan for this years GSoC is to find a student who could add as much
testing to Debian Med packages as possible.  In this sense BioPython is
a great test suite since it somehow adds tests to those packages that
come without one. :-)
 
Thanks again for getting involved into this bug report.  I think
I will use the upcoming Debian Med sprint[1] to sort these things
out.

Kind regards

        Andreas.

[1] https://wiki.debian.org/Sprints/2016/DebianMed2016

-- 
http://fam-tille.de