Hi, On Tue, May 15, 2018 at 2:36 PM, <l...@protonmail.ch> wrote:
> Hi again, > > Yes I propose removing the persistent lifetime, removing the concept of > different lifetimes as a whole, and just keeping the oneshot lifetime as > the implicit lifetime. > > I initially did not consider any of the race conditions, but even then, I > do not think it justifies the complex persistent lifetime and the large > increase in complexity it brings with it. > > I think the approach where the client has to rerequest it each time it > enters the region wherein the constraint has to be active, although more > prone to slightly incorrect behavior, is superior because of the decreased > complexity it burdens the compositor with. > FWIW, both SDL and Xwayland uses “persistent” and we have working implementations of the protocol in weston/libweston and mutter, so it might be hard to convince people (both compositor and clients developers) to drop their working implementation to simplify the development of other compositors... Cheers, Olivier
_______________________________________________ wayland-devel mailing list wayland-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/wayland-devel