> Oh, the problem is that sigwait() behaviour changes after a fork: it is interrupted if an unexpected signal is received, but the signal handler is not called. It behaves correctly (the signal handler is called) without the fork.
>

Reminds me of http://bugs.python.org/issue8407#msg138066

But I think we could just remove this test: honestly, I'm not sure
that checking the behavior in case of delivery of an unblocked signal
is really useful, especially since the behavior is not clearly
defined, see POSIX man page:
"""
The sigwaitinfo() and sigtimedwait() functions may fail if:

[EINTR]
The wait was interrupted by an unblocked, caught signal. It will be
documented in system documentation whether this error will cause these
functions to fail.
"""

"may fail" [...] "It will be documented in system documentation"