Hello, I have the following scenario: Evaluate String Stream until a certain string is found or a time has occured. The string evaluation is made not in a single function, but using a stream iterator and later an evaluator of the retrieved data. I have an iterator class as function object where a boost::shared_ptrboost::timed_mutex mutex; member is defined. It has 2 functions: operator()() which is a worker function of the time_out thread and an iteratorion function. Now I start when entered for the first time the iteration function the time_out thread. The time_out thread should block until the time_out elapses or the another function has found the desired string from the stream. my code looks like: operator()() { boost::timed_mutex::scoped_timed_lock lock(*mutex, true); boost::condition c; boost::xtime xt; boost::xtime_get(&xt, boost::TIME_UTC); xt.sec +=20; c.timed_wait(lock, xt); } My approach was to use the mutex object for syncronization. I assumed that it is locked as long as c.timed_wait(...) is waiting. So I try to lock this object in the other thread. If the mutex object can not be locked the worker thread is still waiting, otherwise time_out elapsed.

From the Rationale section of the thread docs I heavily understand which locking strategy is used for timed_mutex. Can someone point me to a the correct place with documentation and say if this approach will work or I should consider a redesign?

With Kind Regards, Ovanes Markarian