On Tue 13 Sep 2011 12:19:35 CEST, Roland Mas wrote:
Hi,
I started having a look at the bug in the sid chroot of merulo (the ia64
porterbox). After a failed debuild, I re-ran the last command of the
debian/rules, and tried to separate the different tests to see which
were failing. My results:
(sid)lolando@merulo:~/subvertpy-0.8.5/build/lib.linux-ia64-2.7$ failed="";for i in $(python2.7
-m testtools.run -v -f -l subvertpy.tests.test_suite); do python2.7 -m testtools.run $i&>/dev/null
|| failed="$failed $i"; done;echo $failed
Segmentation fault
Segmentation fault
Segmentation fault
subvertpy.tests.test_repos.TestClient.test_is_dir
subvertpy.tests.test_repos.TestClient.test_is_file
subvertpy.tests.test_repos.TestClient.test_paths_changed
(sid)lolando@merulo:~/subvertpy-0.8.5/build/lib.linux-ia64-2.7$
The three tests that fail are therefore test_is_dir, test_is_file and
test_paths_changed in the TestClient class. More precisely, the
segfault occurs in the
self.assertEquals(False,root.is_dir("nonexistant")) clause for
test_is_dir and test_is_file, and in
self.assertEquals({},root.paths_changed()) for test_paths_changed.
Commenting out these three lines ends in all the individual tests
passing, but running the full testsuite in a single invocation still
gets a segfault for some reason.
I haven't dug deeper, but hopefully this can help isolate the problem.
Thanks for looking into this! Looking at the code quickly I don't see
anything obvious we're doing wrong. I'll have a closer look later this
week.
If you still have this setup somewhere, can you perhaps paste the gdb
backtrace of one of these segfaults?
Cheers,
Jelmer
--
To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
