homepage

http://www.python.org/dev/buildbot/all/builders/x86 FreeBSD 6.4 3.x/builds/1570/steps/test/logs/stdio
"""
[ 29/356] test_signal
Fatal error 'mutex is on list' at line 540 in file /usr/src/lib/libpthread/thread/thr_mutex.c (errno = 22)
test test_signal failed -- Traceback (most recent call last):
 File "/usr/home/db3l/buildarea/3.x.bolen-freebsd/build/Lib/test/test_signal.py", line 669, in test_sigwait_thread
 self.check_sigwait(test, signal.SIGUSR1)
 File "/usr/home/db3l/buildarea/3.x.bolen-freebsd/build/Lib/test/test_signal.py", line 632, in check_sigwait
 self.assertEqual(os.waitpid(pid, 0), (pid, 0))
AssertionError: Tuples differ: (50592, 134) != (50592, 0)
First differing element 1:
134
0
- (50592, 134)
? ^^^
+ (50592, 0)
? ^
"""
Fatal error 'mutex is on list' at line 540 in file /usr/src/lib/libpthread/thread/thr_mutex.c (errno = 22)
In a multi-threaded program, we're only allowed to call async-safe functions between fork() and exec() in the child process (well, taken literaly this means we shouldn't run Python at all after fork...). It looks like FreeBSD 6.4 doesn't like new threads after fork() :
http://unix.derkeiler.com/Mailing-Lists/FreeBSD/hackers/2008-03/msg00108.html
test_sigwait_thread should either be skipped on FreeBSD 6 (it seems to pass on FreeBSD 7), or removed altogether.

> In a multi-threaded program, we're only allowed to call async-safe
> functions between fork() and exec() in the child process
Oh, I don't know that. The test should be modified to create a new clean Python process using subprocess to run the test.

Patch attached:
- a new process is spawned (using assert_python_ok()) to avoid
undefined behavior (and crash on FreeBSD)
- I've kept SIGUSR1 default handler, because, as noted in
http://bugs.python.org/issue8407#msg138066 , there can be subtle
differences between default handlers and user-installed ones which can
mask bugs
- I've fixed a comment in test_sigwait

New changeset d07d0afea9a7 by Victor Stinner in branch 'default':
Issue #12316: Fix sigwait() test using threads
http://hg.python.org/cpython/rev/d07d0afea9a7 

> Patch attached
The patch looks correct, but only FreeBSD 6.4 can tell us if it "works" or not, so I commited your new version of the test.
I inlined check_sigwait() because the function was only called by test_sigwait().

test_signal passed in build #1577 of x86 FreeBSD 6.4 3.x: I close this issue. Thanks neologix!