Peter Dimov wrote: [...] > Double calls to pthread_mutex_lock are the equivalent of doing > > scoped_lock l1(m); > scoped_lock l2(m); > > This is already undefined behavior when m is boost::mutex according to the > mutex requirements; m is not required to check for deadlocks.
Okay. > We are discussing the > > scoped_lock l1(m); > l1.lock(); > > case. Since l1 is required to know whether it's been locked or not (it has a > public locked() query) it can easily check. What for? http://www.terekhov.de/DESIGN-futex-CV.cpp (see mutex::guard and mutex::release_guard). I guess that I'm just missing and/or misunderstanding something... oder? regards, alexander. _______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
