[Debian-science-sagemath] sagemath testing migration

[Tobias Hansen]

On 12/23/2016 10:21 PM, Ximin Luo wrote:
> Ximin Luo:
>> Ximin Luo:
>>> Tobias Hansen:
>>>> - fixes the piuparts checks, if necessary
>>>
>>> All good on that front, at least:
>>>
>>> https://piuparts.debian.org/sid/source/s/sagemath.html
>>>
>>
>> 7.4-4 is now installed in unstable, i386 and amd64:
>>
>> https://buildd.debian.org/status/package.php?p=sagemath
>>
>> I looked into the brial segfault a bit more, got a minimal test-case here without Sage:
>>
>> $ python
>>>>> from brial import groebner_basis
>>>>> from brial.blocks import declare_ring
>>>>> r = declare_ring(['x', 'y'], globals())
>>>>> id = [x + y]
>>>>> groebner_basis(id)
>> [.. shit explodes ..]
>>
> 
> Hm, the above example only works in the presence of sage (e.g. after using `. <(debian/rules export-build-env)`, it doesn't work for a plain brial:i386 install. If I run it without Sage, I get:
> 
> $ python -c 'from brial import groebner_basis'
> Traceback (most recent call last):
>   File "<string>", line 1, in <module>
>   File "/usr/lib/python2.7/dist-packages/brial/__init__.py", line 31, in <module>
>     from .PyPolyBoRi import *
>   File "/usr/lib/python2.7/dist-packages/brial/PyPolyBoRi.py", line 125, in <module>
>     from .dynamic.PyPolyBoRi import *
> ImportError: No module named dynamic.PyPolyBoRi
> 
> This thread from upstream https://sourceforge.net/p/polybori/mailman/message/31089269/ says:
> "this means, that PolyBoRi's python binding were not actually build."
> 
> Presumably, this means the non-Sage bindings.
> 
> X
> 

Yes, python-polybori 0.8.3 had these files while python-brial does not:

/usr/lib/python2.7/dist-packages/polybori/dynamic/PyPolyBoRi.so
/usr/lib/python2.7/dist-packages/polybori/dynamic/__init__.py

I just checked that after a brial build these files are not in
debian/tmp so it's not just an oversight in installing the files.

Merry Christmas!
Tobias