Hi Byron On Wednesday 12 May 2004 01:04 am, Richard Drummond wrote: > sem_init() definitely returns "Function not implemented" here too. > sem_open() appears to work, although I don't know that it's particularly > useful. Perhaps the best course of action would be to emulate the POSIX > semaphore stuff with pthread mutexes and condition variables - like SDL > does. Rethinking this - it should be possible to implement sem_init() behaviour using sem_open(). I'll give it a try, anyway. Cheers, Rich