Le septidi 17 frimaire, an CCXXV, Hendrik Leppkes a écrit : > I'm not sure you can safely avoid that in this design when dealing > with a fully generic library and no information or control whats going > on in other threads.
You are right, I had not realized there was a race condition here. What happens if shutdown() is performed but not close()? > The "proper" win32 way of doing non-polling IO is using the Overlapped > IO functions, which runs async and signals event objects when > something happens - but that would practically require re-implementing > the entire worker thread for win32. The receiving worker thread is less than 60 lines, the sending one is 100 lines because of the rate control code that can certainly be factored. All in all, that would not be that terrible, and if done cleanly can be reused later for more generic code.
signature.asc
Description: Digital signature
_______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
