[Debian-science-sagemath] 7.4 with the latest shiny new packages: 4000 failed tests

Ximin Luo infinity0 at debian.org
Thu Nov 3 11:39:00 UTC 2016


I added some primitive stats-scripts to debian/rules. You can use it to see which files, error types, or stack traces, have the most failures.

For example, it looks like 1247 of these are (perhaps) linbox related:

$ debian/rules failed-tests-by-cause | tail -n5 | tr '\t' '\n'
     23 return func(*args, **kwds)
      File "sage/matrix/matrix_modn_sparse.pyx", line 725, in sage.matrix.matrix_modn_sparse.Matrix_modn_sparse._rank_linbox (/home/infinity0/sage/sagemath/sage/src/build/cythonized/sage/matrix/matrix_modn_sparse.c:8579)
        sig_on()
      File "src/cysignals/signals.pyx", line 109, in cysignals.signals.sig_raise_exception (build/pari/src/cysignals/signals.c:1478)
    SignalError: Segmentation fault

     26 return self._poly_linbox(var=var, typ='minpoly')
      File "sage/matrix/matrix_integer_dense.pyx", line 1324, in sage.matrix.matrix_integer_dense.Matrix_integer_dense._poly_linbox (/home/infinity0/sage/sagemath/sage/src/build/cythonized/sage/matrix/matrix_integer_dense.cpp:13914)
        sig_on()
      File "src/cysignals/signals.pyx", line 109, in cysignals.signals.sig_raise_exception (build/pari/src/cysignals/signals.c:1478)
    SignalError: Segmentation fault

    274 r = self._rank_linbox()
      File "sage/matrix/matrix_integer_dense.pyx", line 3468, in sage.matrix.matrix_integer_dense.Matrix_integer_dense._rank_linbox (/home/infinity0/sage/sagemath/sage/src/build/cythonized/sage/matrix/matrix_integer_dense.cpp:30469)
        sig_on()
      File "src/cysignals/signals.pyx", line 109, in cysignals.signals.sig_raise_exception (build/pari/src/cysignals/signals.c:1478)
    SignalError: Segmentation fault

    320 with opt_ctx: # we are preserving the global options state here
      File "sage/libs/singular/function.pyx", line 1502, in sage.libs.singular.function.call_function (/home/infinity0/sage/sagemath/sage/src/build/cythonized/sage/libs/singular/function.cpp:16366)
        sig_on()
      File "src/cysignals/signals.pyx", line 109, in cysignals.signals.sig_raise_exception (build/pari/src/cysignals/signals.c:1478)
    SignalError: Segmentation fault

   1247 return self._poly_linbox(var=var, typ='charpoly')
      File "sage/matrix/matrix_integer_dense.pyx", line 1328, in sage.matrix.matrix_integer_dense.Matrix_integer_dense._poly_linbox (/home/infinity0/sage/sagemath/sage/src/build/cythonized/sage/matrix/matrix_integer_dense.cpp:13973)
        sig_on()
      File "src/cysignals/signals.pyx", line 109, in cysignals.signals.sig_raise_exception (build/pari/src/cysignals/signals.c:1478)
    SignalError: Segmentation fault

I also got access to a 23-core machine that (with ccaching) runs the whole build plus tests in ~2 hours, so no need to be too scared of the 4000 figure, we are definitely in a better position than we were with Sage 7.3 back in August.

X

Tobias Hansen:
> Hi,
> 
> Jeroen, it would be nice if you could comment on this. We get these
> ~4000 failing tests most likely because we are building sage 7.4 with
> pari 2.9. You already started updating pari in sage to 2.8.1. Do you
> plan to update to pari 2.9 as well?
> 
> I already tried applying your patches from sage trac #21765 and #21703
> which did not help. Do the failures in the test log from Ximins mail
> below look like it's caused by pari 2.9?
> 
> Best,
> Tobias
> 
> On 11/01/2016 11:17 AM, Ximin Luo wrote:
>> The good news is that the build (including docbuild) works without any other changes.
>>
>> Also the creation of debs works, if you override the test failures and patch dh_sphinxdoc as described in README.Debian.
>>
>> Test logs here:
>>
>> https://people.debian.org/~infinity0/res/sagemath-7.3+gite606a24a_ptestlong.log.gz
>>
> 
> 
> _______________________________________________
> 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