On Mon, 2008-02-11 at 00:58 +0100, Jason Tackaberry wrote: > MainThreadCallback.set_async is deprecated (and logs a warning). Calling a > MainThreadCallback will now return an InProgress object, and the caller > can wait() on this InProgress if synchronous behaviour is needed.
I should have mentioned that although this is the new and better way, the API should be backward compatible. If set_async(False) is set, the result is always returned (regardless if it's called in the main thread or not). The new approach also fixes the problem if a MainThreadCallback is invoked in two threads at the same time. With the old way, there is a single lock and result/exception holder for MainThreadCallback so if it is called at the same time in two threads, you will get unpredictable results. (dischi, you noted this bug in a comment, which has been removed since the bug is fixed if you use cb().wait().) The new way solves this by creating a unique InProgress for each invocation, and relying on the new thread-safe semantics added to InProgress.wait(). It uses a threading.Event object, so even if the InProgress is finished before the other thread wait()s on it, it will still work as expected. Jason. ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ Freevo-devel mailing list Freevo-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/freevo-devel
