Hi Alex, A mutex object contains all the data structures for locking. A lock object locks the given mutex object in the constructor and unlocks it in its destructor. Example: class ThreadSafeSomething { private: boost::mutex mutex; public: void doSomething() { boost::mutex::scoped_lock(mutex); something(); } // mutex is unlocked here }; If there would be no lock object you would have to write something like (welcome to java world): class ThreadSafeSomething { private: my::special::mutex mutex; public: void doSomething() { mutex.lock(); something(); mutex.unlock(); } }; Final question: What happens when an exception is thrown in something()? Kind regards, Markus Am Mittwoch, den 01.07.2009, 17:08 +0200 schrieb Axel:

Hello

This post may not be well suited to this mailing list. I have some questions about the lockable concept provided in boost. There are many "lock" types, and many "mutex" types. My main question is then : when should I chose a given lock type rather a given mutex type ? Given the fact that some of lock types are not thread-safe, if only one thread needs locks, why should I use one of these lock types and why shouldnt I use some mutex type ?

Finally, what were the goals of providing so much different ways to lock/unlock resources ? (I may not be aware of all the concepts related to secure resource access)

Since the "lock" et "mutex" types provide ways to lock/unlock access to resources, why are not these types defined under the same name ? Are the concepts they implements much more different than I think ?

Thanks Axel _______________________________________________ Boost-users mailing list Boost-users@lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users