signal/bsd_signal

[Roland McGrath]

> If the mask is returned and the bit is not set something might be
> handled differently than expected.  I don't know of a case, I'm just
> looking throught the specs.  If some piece in the test suite is
> testing this it will complain.  Now it won't anymore.

What I have been saying is that I have no reason to think that seeing that
bit set is the "expected behavior".  I haven't checked BSD, but my offhand
guess would be that the bit would not be set in sa_mask as reported by
sigaction after a signal call.