Date: 2001-04-03 16:53:44
--- In boost_at_y..., Lie-Quan Lee <llee1_at_l...> wrote:
> At Tue, 03 Apr 2001 21:08:31 -0000,
> William Kempf wrote:
> > > "undefined" could be an alias of "guaranteed dead lock"
> > > on the implementation in pthreads.
> > No, undefined does _not_ mean "guaranteed dead lock" by the POSIX
> > definition (so what specific implementations do is irrelevant).
> Here is what IEEE-1003.1-200x says:
> If the mutex type is PTHREAD_MUTEX_DEFAULT, attempting to
> lock the mutex results in undefined behavior. Attempting to unlock
> the mutex if it was not locked by the calling thread results in
> undefined behavior. Attempting to unlock the mutex if it is not
> results in undefined behavior.
> According to that, the pthread implementor is legal to make
> PTHREAD_MUTEX_DEFAULT type mutex to be an alias of
> PTHREAD_MUTEX_NORMAL type mutex which is "guaranteed dead lock" on
> recursive locking.
Yes, but the policy remains "undefined". They are still not
synonyms, regardless of whether or not the implementation causes both
to behave identically.
> I guess that my question in that message is: if does and should the
> "undefined" defined in Boost.Threads have different meaning than
> IEEE-1003.1-200x said?
It doesn't have a different meaning, and the two policies are quite
different. POSIX and Boost.Threads are in agreement here.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk