Subj : Re: thread suspend using pthreads? To : comp.programming.threads From : David Hopwood Date : Wed Aug 31 2005 05:41 pm Joe Seigh wrote: > The best way to implement suspend is to not do it. Every threading > api that has implemented suspend/resume has either deprecated it or > has warnings not to use it. It's basically useless except as a > mechanism to introduce race conditions. And there are a lot easier > ways to create race conditions. ITYM deadlock, not race conditions. Suspend can't generate a race condition that wasn't possible already. I tend to agree about suspend being a bad idea. It can be useful for debugging, though. In that case we don't care about the temporary deadlocks that may occur until the thread is manually resumed. (The user can't deadlock!) -- David Hopwood .