[Debian-science-sagemath] 62 failed tests and classification of remaining failures
Ximin Luo
infinity0 at debian.org
Thu Oct 13 10:45:00 UTC 2016
They are different, they were just spuriously linked to gmp due to a bug in the upstream build scripts. Also we're using 094h and not 094g as Sage are.
I fixed the spurious linking [1], but the issue still remains. I even tried building it using GMP float instead of C double [2] but the error didn't go away.
However, I see that you guys have also experienced the same issue in the past (commit 269c1e1551285566b8ba7a2b890989e5590e9f11) so I suppose it's safe to ignore it for now.
X
[1] https://anonscm.debian.org/git/debian-science/packages/cddlib.git/commit/?id=c56ad14d10953784b64ee7cee56b9fa74cf1f316
[2] https://anonscm.debian.org/git/debian-science/packages/cddlib.git/commit/?h=_experiment_gmpfloat&id=f76de112e7746877bed38b9234bfabb498e80a40
François Bissey:
> In Gentoo (and the test is fine for me):
> ldd -r /usr/bin/cdd_both_reps
> linux-vdso.so.1 (0x00007ffe1e1db000)
> libcdd.so.0 => /usr/lib64/libcdd.so.0 (0x00007fe9cb88f000)
> libc.so.6 => /lib64/libc.so.6 (0x00007fe9cb4f3000)
> /lib64/ld-linux-x86-64.so.2 (0x00007fe9cbab0000)
> Is your libcdd really different from your libcddgmp?
> ldd -r /usr/lib64/libcdd.so
> linux-vdso.so.1 (0x00007ffec3a65000)
> libc.so.6 => /lib64/libc.so.6 (0x00007f2045031000)
> /lib64/ld-linux-x86-64.so.2 (0x0000561003df2000)
>
> François
>
>> On 10/10/2016, at 23:22, Ximin Luo <infinity0 at debian.org> wrote:
>>
>> Everything seems to be fine, link wise [1]. It is the RDF test that fails:
>>
>> sage: Polyhedron(point_list)
>> A 3-dimensional polyhedron in RDF^3 defined as the convex hull of 14 vertices
>>
>> and the QQ test passes. However sage 7.3 defines:
>>
>> class Polyhedron_RDF_cdd(Polyhedron_cdd, Polyhedron_RDF):
>> _cdd_executable = 'cdd_both_reps'
>>
>> Would it be appropriate to change this to cdd_both_reps_gmp? Not sure why the test passes in vanilla sage though (or does it)...
>>
>> X
>>
>> [1]
>> $ ldd -r /usr/bin/cdd_both_reps_gmp
>> linux-vdso.so.1 (0x00007ffc896ba000)
>> libcddgmp.so.0 => /usr/lib/x86_64-linux-gnu/libcddgmp.so.0 (0x00007fd52f55a000)
>> libgmp.so.10 => /usr/lib/x86_64-linux-gnu/libgmp.so.10 (0x00007fd52f2d7000)
>> libc.so.6 => /lib/x86_64-linux-gnu/libc.so.6 (0x00007fd52ef35000)
>> /lib64/ld-linux-x86-64.so.2 (0x000055df3fd9e000)
>> $ ldd -r /usr/bin/cdd_both_reps
>> linux-vdso.so.1 (0x00007ffeaeb7a000)
>> libcdd.so.0 => /usr/lib/x86_64-linux-gnu/libcdd.so.0 (0x00007f15d83fd000)
>> libgmp.so.10 => /usr/lib/x86_64-linux-gnu/libgmp.so.10 (0x00007f15d817a000)
>> libc.so.6 => /lib/x86_64-linux-gnu/libc.so.6 (0x00007f15d7dd8000)
>> /lib64/ld-linux-x86-64.so.2 (0x000055c949fc7000)
>>
>> François Bissey:
>>> Hum, would need to check that the _gmp executable of cddlib are actually
>>> linked to libcddgmp.so and ultimately to libgmp.so.
>>>
>>> François
>>>
>>>> On 8/10/2016, at 23:08, Ximin Luo <infinity0 at debian.org> wrote:
>>>>
>>>> # One case of "Error: Numerical inconsistency is found. Use the GMP exact arithmetic."
>>>> sage -t --long src/sage/geometry/polyhedron/backend_cdd.py # 7 doctests failed
>>>
>>>
>>> _______________________________________________
>>> Debian-science-sagemath mailing list
>>> Debian-science-sagemath at lists.alioth.debian.org
>>> https://lists.alioth.debian.org/mailman/listinfo/debian-science-sagemath
>>>
>>
>>
>> --
>> GPG: ed25519/56034877E1F87C35
>> GPG: rsa4096/1318EFAC5FBBDBCE
>> https://github.com/infinity0/pubkeys.git
>>
>> _______________________________________________
>> Debian-science-sagemath mailing list
>> Debian-science-sagemath at lists.alioth.debian.org
>> https://lists.alioth.debian.org/mailman/listinfo/debian-science-sagemath
>
>
> _______________________________________________
> Debian-science-sagemath mailing list
> Debian-science-sagemath at lists.alioth.debian.org
> https://lists.alioth.debian.org/mailman/listinfo/debian-science-sagemath
>
--
GPG: ed25519/56034877E1F87C35
GPG: rsa4096/1318EFAC5FBBDBCE
https://github.com/infinity0/pubkeys.git
More information about the Debian-science-sagemath
mailing list