[Python-modules-team] Bug#860663: Bug#860663: pytest: FTBFS on i386: segfault during tests

Thu Apr 20 07:40:34 UTC 2017

Control: tags -1 - moreinfo
Control: reassign -1 pypy 5.6.0+dfsg-4
Control: severity -1 important
Control: affects -1 src:pytest

On 2017-04-20 09:14:08, Lucas Nussbaum wrote:
> On 19/04/17 at 22:15 +0200, Lucas Nussbaum wrote:
> > On 19/04/17 at 20:42 +0200, Sebastian Ramacher wrote:
> > > Control: tags -1 + moreinfo unreproducible
> > > 
> > > On 2017-04-19 09:26:48, Lucas Nussbaum wrote:
> > > > Source: pytest
> > > > Version: 3.0.6-1
> > > > Severity: serious
> > > > Tags: stretch sid
> > > > User: debian-qa at lists.debian.org
> > > > Usertags: qa-ftbfs-20170418-i386 qa-ftbfs
> > > > Justification: FTBFS in stretch on i386
> > > > 
> > > > Hi,
> > > > 
> > > > During a rebuild of all packages in stretch (in a stretch chroot, not a
> > > > sid chroot), your package failed to build on i386.
> > > > 
> > > > Relevant part (hopefully):
> > > > > ../../../testing/test_pluginmanager.py ...........................
> > > > > ../../../testing/test_pytester.py x.......
> > > > > ../../../testing/test_recwarn.py ......................
> > > > > ../../../testing/test_resultlog.py ...........
> > > > > ../../../testing/test_runner.py ...................sssssss...ssss......x.......................
> > > > > ../../../testing/test_runner_xunit.py ...............
> > > > > ../../../testing/test_session.py .................
> > > > > ../../../testing/test_skipping.py ..........................................................x..........
> > > > > ../../../testing/test_terminal.py ..........s..........................s.............................................................
> > > > > ../../../testing/test_tmpdir.py ...........s
> > > > > ../../../testing/test_unittest.py ....................ssssssss................
> > > > > ../../../testing/code/test_code.py ..............
> > > > > ../../../testing/code/test_excinfo.py ............Segmentation fault
> > > > > E: pybuild pybuild:283: test: plugin custom failed with: exit code=139: cd debian/tmp/test-working-directory && pypy -m pytest --lsof -rfsxX --ignore=/<<PKGBUILDDIR>>/testing/freeze --ignore=/<<PKGBUILDDIR>>/testing/test_entry_points.py --ignore=/<<PKGBUILDDIR>>/testing/test_pdb.py /<<PKGBUILDDIR>>/testing
> > > > > dh_auto_test: pybuild --test -i pypy -p 5.6 --system=custom --test-args=cd debian/tmp/test-working-directory && {interpreter} -m pytest --lsof -rfsxX --ignore={dir}/testing/freeze --ignore={dir}/testing/test_entry_points.py --ignore={dir}/testing/test_pdb.py {dir}/testing returned exit code 13
> > > > > debian/rules:20: recipe for target 'override_dh_auto_test' failed
> > > 
> > > I'm unable to reproduce the crash. Could you please provide a backtrace?
> > 
> > Can you provide a build log and diff it with mine?

FWIW, the diff is attached.

> I dug a bit more, and can say that:
> - it also fails on a smaller EC2 instance, so it's not related to the
>   fact that the first instance was quite powerful, many cores, etc.
> - it also fails on a completely up-to-date jessie host (still with a
>   stretch/i386 chroot) -- my kernel wasn't entirely up-to-date
> 
> I've generated a core file and put it at:
> http://aws-logs.debian.net/2017/04/18/pypy-segfault.core.gz

Thanks, but it fails with 403.

> The backtrace looks like:
> #0  0xf553ba24 in pypy_g_handle_bytecode__AccessDirect_None () from /usr/lib/pypy/bin/libpypy-c.so
> #1  0xf5b3282f in pypy_g_portal_5 () from /usr/lib/pypy/bin/libpypy-c.so
> #2  0xf5e68bb4 in pypy_g_ll_portal_runner__Unsigned_Bool_pypy_interpreter () from /usr/lib/pypy/bin/libpypy-c.so
> #3  0xf5b327d8 in pypy_g_PyFrame_dispatch () from /usr/lib/pypy/bin/libpypy-c.so
> #4  0xf54e954e in pypy_g_execute_frame () from /usr/lib/pypy/bin/libpypy-c.so
> #5  0xf636b801 in pypy_g_execute_frame_rvmprof.star_3 () from /usr/lib/pypy/bin/libpypy-c.so
> #6  0xf54bc07c in pypy_g_PyFrame_run () from /usr/lib/pypy/bin/libpypy-c.so
> #7  0xf5533573 in pypy_g_CALL_FUNCTION__AccessDirect_None () from /usr/lib/pypy/bin/libpypy-c.so
> #8  0xf5537eba in pypy_g_dispatch_bytecode__AccessDirect_None () from /usr/lib/pypy/bin/libpypy-c.so
> #9  0xf553ba29 in pypy_g_handle_bytecode__AccessDirect_None () from /usr/lib/pypy/bin/libpypy-c.so
> #10 0xf5b3282f in pypy_g_portal_5 () from /usr/lib/pypy/bin/libpypy-c.so
> #11 0xf5e68bb4 in pypy_g_ll_portal_runner__Unsigned_Bool_pypy_interpreter () from /usr/lib/pypy/bin/libpypy-c.so
> #12 0xf5b327d8 in pypy_g_PyFrame_dispatch () from /usr/lib/pypy/bin/libpypy-c.so
> #13 0xf54e954e in pypy_g_execute_frame () from /usr/lib/pypy/bin/libpypy-c.so
> #14 0xf636b801 in pypy_g_execute_frame_rvmprof.star_3 () from /usr/lib/pypy/bin/libpypy-c.so
> #15 0xf54bc07c in pypy_g_PyFrame_run () from /usr/lib/pypy/bin/libpypy-c.so
> #16 0xf5533573 in pypy_g_CALL_FUNCTION__AccessDirect_None () from /usr/lib/pypy/bin/libpypy-c.so
> #17 0xf5537eba in pypy_g_dispatch_bytecode__AccessDirect_None () from /usr/lib/pypy/bin/libpypy-c.so
> #18 0xf553ba29 in pypy_g_handle_bytecode__AccessDirect_None () from /usr/lib/pypy/bin/libpypy-c.so

In any case, reassigning to pypy. Currently my best explanation is that pypy
crashes during pytest's excinfo tests when running with a jessie kernel, but not
with a stretch kernel.

Cheers
-- 
Sebastian Ramacher
-------------- next part --------------
A non-text attachment was scrubbed...
Name: pytest.diff
Type: text/x-diff
Size: 243672 bytes
Desc: not available
URL: <http://lists.alioth.debian.org/pipermail/python-modules-team/attachments/20170420/8cd63de3/attachment-0001.diff>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: not available
URL: <http://lists.alioth.debian.org/pipermail/python-modules-team/attachments/20170420/8cd63de3/attachment-0001.sig>