On Sat, Mar 10, 2012 at 4:04 PM, Keith Packard <[email protected]> wrote: > On Sat, 10 Mar 2012 14:31:52 -0800, Jamey Sharp <[email protected]> wrote: >> Although, in "Restore Xlib semantics", it would be nice if you'd fix >> up the "It appears that classic Xlib respected user locks" comment. I >> think it suffices to delete the final sentence, "So we'll choose to >> let the thread that got in first consume events, despite the later >> thread's user locks." > > something like: > > - * deadlock. So we'll choose to let the thread > - * that got in first consume events, despite the > - * later thread's user locks. */ > + * deadlock. So we let the thread got in first read > + * events, then signal other waiting threads and go > + * back to check for a user lock */
Looks great; just insert "let the thread [that] got in first" and push. :-) >> Thanks for the careful review and detailed summary! > > Xlib locking code is hard. Let's go rocketing! OMG rocket science is so much easier than Xlib surgery. Jamey _______________________________________________ [email protected]: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: http://lists.x.org/mailman/listinfo/xorg-devel
