[Biopython-dev] Bug#751277: python-biopython: FTBFS on mips* powerpc s390x

Michiel de Hoon mjldehoon at yahoo.com
Mon Aug 18 07:35:53 UTC 2014


Hi Andreas,

Without access to powerpc, I have no way to test the code.
Can you try recompiling Biopython and checking what exactly happens inside the distance_converter function in Bio/Cluster/clustermodule.c ?
For example, I am really wondering what strlen(data) inside this function returns on powerpc.

Best,
-Michiel.

--------------------------------------------
On Sat, 8/16/14, Andreas Tille <andreas at an3as.eu> wrote:

 Subject: Re: Bug#751277: python-biopython: FTBFS on mips* powerpc s390x
 To: "Peter Cock" <p.j.a.cock at googlemail.com>
 Cc: "Dejan Latinovic" <Dejan.Latinovic at imgtec.com>, "Michiel de Hoon" <mjldehoon at yahoo.com>, "Biopython discussion list" <biopython at lists.open-bio.org>, "751277 at bugs.debian.org" <751277 at bugs.debian.org>, "biopython-dev at biopython.org" <biopython-dev at biopython.org>
 Date: Saturday, August 16, 2014, 5:37 AM
 
 Hi Peter,
 
 On Thu, Aug 14, 2014 at
 09:52:40AM +0100, Peter Cock wrote:
 >
 >
 > >   1. waiting for
 your confirmation / patch
 >
 >   2. deactivating the specific test
 > >   3. exclude mips for
 biopython
 > >   4. ? any
 better idea ?
 > >
 >
 > In the current state all the work we spent in biopython
 over the last
 > > monthes will not
 migrate to testing for the simple reason that the
 > > current package in testing just does
 not run the test suite at build
 > >
 time and moreover python3 is not supported.
 > >
 > > Kind
 regards
 > >
 >
 >      Andreas.
 > 
 > I would suggest (2), deactivate this test
 (at least for for mips) as
 > the most
 practical short term solution for the Debian packages.
 > Or if you prefer (3), don't target
 mips for the Biopython package
 >
 (yet).
 > 
 > Medium
 term, I hope we can fix the C code to handle either
 > Endian platform - option (1).
 
 It seems after having fixed
 the issue caused by wise we have one
 remaining problem:
 
   On powerpc[1] and s390x[2] test_Cluster
 fails even with Python 2.7 with:
 
 ======================================================================
 ERROR: test_clusterdistance
 (test_Cluster.TestCluster)
 ----------------------------------------------------------------------
 Traceback (most recent call last):
   File
 "/«BUILDDIR»/python-biopython-1.64+dfsg/.pybuild/pythonX.Y_3.4/build/Tests/test_Cluster.py",
 line 212, in test_clusterdistance
    
 method='a', transpose=0)
 ValueError:
 method should be a single character
 
 ======================================================================
 ERROR: test_kcluster
 (test_Cluster.TestCluster)
 ----------------------------------------------------------------------
 Traceback (most recent call last):
   File
 "/«BUILDDIR»/python-biopython-1.64+dfsg/.pybuild/pythonX.Y_3.4/build/Tests/test_Cluster.py",
 line 141, in test_kcluster
    
 method='a', dist='e')
 ValueError: method should be a single
 character
 
 ======================================================================
 ERROR: test_somcluster
 (test_Cluster.TestCluster)
 ----------------------------------------------------------------------
 Traceback (most recent call last):
   File
 "/«BUILDDIR»/python-biopython-1.64+dfsg/.pybuild/pythonX.Y_3.4/build/Tests/test_Cluster.py",
 line 557, in test_somcluster
    
 inittau=0.02, niter=100, dist='e')
 ValueError: distance should be a single
 character
 
 ======================================================================
 ERROR: test_treecluster
 (test_Cluster.TestCluster)
 ----------------------------------------------------------------------
 Traceback (most recent call last):
   File
 "/«BUILDDIR»/python-biopython-1.64+dfsg/.pybuild/pythonX.Y_3.4/build/Tests/test_Cluster.py",
 line 290, in test_treecluster
    
 transpose=0, method='a', dist='e')
 ValueError: method should be a single
 character
 
 ----------------------------------------------------------------------
 Ran 210 tests in 293.712 seconds
 
 FAILED (failures = 1)
 
 
 On sparc[3]
 there is a problem with dialign but sparc is no release
 architecture and wie might ignore this.  It
 might be a helpful hint
 anyway.
 
 Any hint for the test_Cluster
 problem?  If not I would also consider to
 hide it cowardly under the carpet for the
 moment.  The new package is so
 much better
 tested than the one in the testing distribution which
 does
 not even dare about any unit tests and
 only for this reason reached the
 testing
 distribution.
 
 What do you
 think?
 
 Kind regards
 
    
    Andreas.
 
 scrool these links to the end to see the
 problem:
 
 [1] https://buildd.debian.org/status/fetch.php?pkg=python-biopython&arch=powerpc&ver=1.64%2Bdfsg-3&stamp=1408116532
 [2] https://buildd.debian.org/status/fetch.php?pkg=python-biopython&arch=s390x&ver=1.64%2Bdfsg-3&stamp=1408107524
 [3]
 https://buildd.debian.org/status/fetch.php?pkg=python-biopython&arch=sparc&ver=1.64%2Bdfsg-3&stamp=1408130792
 
 -- 
 http://fam-tille.de
 



More information about the Biopython-dev mailing list