Now those methods use lock for querying queue size, like

    def qsize(self):
        with self.mutex:
            return self._qsize()

The lock is not necessary because thread context switch may be done *after* returning from mutex protected code but *before* getting result by calling side.

All three methods (qsize(), empty() and full()) gives *approximated value*, so getting rid of lock doesn't make it less stringent.

It may seem pointless to hold the lock, but it is guaranteed behavior (and has been so for a very, very long time).

'''
    # Override these methods to implement other queue organizations                                                               
    # (e.g. stack or priority queue).                                                                                             
    # These will only be called with appropriate locks held   

    # Initialize the queue representation                                                                                         
    def _init(self, maxsize):
        self.queue = deque()

    def _qsize(self):
        return len(self.queue)

    # Put a new item in the queue                                                                                                 
    def _put(self, item):
        self.queue.append(item)

    # Get an item from the queue                                                                                                  
    def _get(self):
        return self.queue.popleft()
 
'''

Raymond, is there known custom third-party queue class derived from queue.Queue? I believe all those are in stdlib only.

Also locking guarantee is promised by comment in source code only, documentation says nothing about it.

I believe proposed change will not make any harm, but ok with closing as "wontfix" if you fill it's dangerous.

Yes, there are.

https://code.openhub.net/search?s=%22def%20_qsize%22%20%22import%20Queue%22&pp=0&ff=1&mp=1&ml=1&me=1&md=1&filterChecked=true