A solution loops clock thread ⌚️, main thread and philosophers threads 👴. Main thread are constantly comparing all thread timestamp and current time aritmetic difference against - time to die 💀. Clock thread ⌚️ is updating current time, needed, for threads 👴, for aritmetic logic.
In main.c the allocate_philosophers() returns a pointer array of t_philosopher structs. Structs are used to pass into start_philosopher_threads() where each initiated 🚦 thread has a proper struct.
wait_for_dead() checks if any of threads timestamps and current time aritmetic difference is bigger that death time 💀 and checks if all philosophers have eaten number of times 🍴.
lock_all stops all philosopher threads by mutex lock mutex->global 🛑. Global mutex lock are in each philosopher thread, for purpose of preventing data races and for having control from the main thread.
lock_all will wait for all threads to jump into global mutex lock 🛑, because this methon does not use thread_join(). Yes this method does not use thread_join(). Experiments proven 🔬 that we can end child threads by locking them and quiting main.
So before proceeding to freeing memory it is important that threads, would be locked🛑. lock_all then locks clock thread ⌚️🛑.
Eventualy we proceed to terminate_all 🏁, which will delocate all allocated memory including mutexes, structs and contents of structs, pointer array. We believe that return does terminate process, clean, and those philosopher thread souls ✨, disapears.
Here a simplified version displaying mutex lock in action
Faster
Example of structs sharing same memory