On Friday, March 1, 2002, at 01:11 , Mark Asbach wrote: please excuse if I send this question to the wrong list - I'm absolutely new to the Darwin project. This particular question crosses the kernel/userland boundary, so this list should be fine. darwin-userlevel might have been slightly better though. Currently I'm trying to port a realtime CORBA ORB to Mac OS X. While it compiles and runs fine on Solaris machines, I've got certain troubles with X's pthread implementation. As I was told on the student developer's list, the Darwin kernel currently doesn't support per-thread signals. So I'd tried to replace the unimplemented calls to pthread_kill() with pthread_cancel() or pthread_detach(). Both result in 'zombie' threads that are stuck in a "semaphore_wait_trap" within pthread_exit() as the gdb call chain tells me. pthread_kill() doesn't merely kill the target thread; it sends a signal. Signals in some cases are used for legacy or portability reasons to communicate among threads, so unless you understand the application you are porting well enough to be sure what it is using pthread_kill() for (and you understand the program flow to know that pthread_cancel() won't change things in a damaging way), pthread_cancel() probably won't do the job for you. pthread_detach() surely won't help by itself, since all it does is make a thread's exit state immediately reapable by the threads implementation, (as opposed to keeping it around until pthread_join() is called for the thread). Since someone also stated, that this is a known issue and already corrected in "the current builds of 10.2 (expected July or so)", I wonder, if I can replace this specific part of Mac OS X's kernel with a current build of Darwin (that hopefully got rid of this issue already) - and how to do that in detail. You could potentially get a Darwin system working with an updated xnu and Libsystem, but this isn't for the faint of heart. Unless you're already familiar with how Apple projects are built, you're in for several days of learning. As for replacing those components on an OS X 10.1 box, I wouldn' t even try it, since it is bound to break many other things. Jason _______________________________________________ darwin-kernel mailing list | darwin-kernel@lists.apple.com Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/darwin-kernel Do not post admin requests to the list. They will be ignored.