[Debian-science-sagemath] 7.3 testing

Jerome BENOIT calculus at rezozer.net
Thu Aug 11 20:46:03 UTC 2016


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA512

Hi,

I have just performed a test with the current material:
I still have a zombie at the end, but it a jave zombie (not a python): some progress here.

On 10/08/16 23:00, Ximin Luo wrote:
> Ximin Luo:
>> Right now as of commit 9a574672, for ptestlong I am getting quite a lot of segfaults with a stack trace that mentions sage/rings/polynomial/pbori.pyx, so this might be worth looking at next. (Yes, I did uninstall polybori and only have the new brial installed.)
>>
> 
> This is ptestlong after a successful docbuild, with the pari local symlink workaround described in README.Debian.
> 
> https://people.debian.org/~infinity0/res/sagemath-7.3+git9a574672+localparipatches+rebuiltlinbox_ptestlong.txt (view in browser)
> https://people.debian.org/~infinity0/res/sagemath-7.3+git9a574672+localparipatches+rebuiltlinbox_ptestlong.log (download)
> 
> 319 tests failed, not counting tests skipped due to segfaults and aborts:
> 
> $ grep failed sage/logs/ptestlong.log.summary | grep -oP '# \d+' | cut -b3- | awk '{s+=$1} END {print s}'

- From where does sage/logs/ptestlong.log.summary come ?


> 316
> $ grep -v failed sage/logs/ptestlong.log.summary 
> ----------------------------------------------------------------------
> sage -t --long src/sage/rings/polynomial/multi_polynomial_sequence.py  # Killed due to segmentation fault
> sage -t --long src/sage/rings/polynomial/pbori.pyx  # Timed out (with hangup after interrupt)
> sage -t --long src/sage/structure/sage_object.pyx  # Killed due to abort
> ----------------------------------------------------------------------
> Total time for all tests: 5817.3 seconds
>     cpu time: 16328.7 seconds
>     cumulative wall time: 20911.6 seconds
> $ grep "Failed example" sage/logs/ptestlong.log | wc -l
> 319
> 
> Some notes, from analysing the logs:
> 
> - all the segfaults are due to pbori.pyx/pbori.cpp. this is actually cysignals working properly
> - other cysignals issues are due to path issues, we can either add a patch or just package+upload cysignals
> - lots of "NameError: " which is quite concerning but also means they could probably all be fixed at once
> - some sagenb errors, probably due to our older version (0.11.2 vs 0.13)

Will packaging cysignals ease  the checking process ?
((I guess yes, but I want other point of view to convince me to package.))

> 
> I'll take a break from this and resume in a few weeks.
> 
Enjoy you break.

Cheers,
Jerome


> X
> 

- -- 
Jerome BENOIT | calculus+at-rezozer^dot*net
https://qa.debian.org/developer.php?login=calculus@rezozer.net
AE28 AE15 710D FF1D 87E5  A762 3F92 19A6 7F36 C68B
-----BEGIN PGP SIGNATURE-----
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=
=lPI+
-----END PGP SIGNATURE-----



More information about the Debian-science-sagemath mailing list