Considering there were some hints that the wayland protocol might get
frozen soon, I think I better ask about some of my concerns with it.
In order to duplicate popular window management actions for sticking to
output edges, tiling, and maximizing windows, I think the
wl_shell_surface::configu
If dispatch_idle_sources() has a side effect of reading from one of the
epoll fd's it can cause the fd to become un-ready yet there are still
events in the internal queue.
This patch moves the (possibly mis-named) post_dispatch_check functions
so that they are run after the dispatch_idle_sourc