So that event loop doesn't spin needlessly; every call to queue an event wakes the lock, but we only need one lock release to process all events. So the wait loop internal counter is cleared to 0 prior to processing the events
This causes a hang (which could be resolved by acquiring the mutex but that requires attaching to haxe thread)
So that event loop doesn't spin needlessly; every call to queue an event wakes the lock, but we only need one lock release to process all events. So the wait loop internal counter is cleared to 0 prior to processing the events
This causes a hang (which could be resolved by acquiring the mutex but that requires attaching to haxe thread)