Re: meaning of "the process is blocking [a specific signal]"

[Roland McGrath <roland@xxxxxxxxxx>]

It's certainly the case that this wording predates threads and was written
to talk about the old-fashioned single-threaded view.  This is the same
case as all SIGTTOU/SIGTTIN cases, in read/write, etc.  Ideally they would
all refer to some common language in one place to be clear.

I think the only way to take this case is as s/process/thread/.
It's the only thing that's sensible for what this case is about.
It's probably what all real implementations actually do.  The call
itself is what can generate the signal.  If the calling thread is
blocking the signal then the call will complete in specified ways
despite being in a background pgrp.  It makes no sense to generate
the signal and have it affect other threads when the instigating
call itself is not affected.


Thanks,
Roland