[PATCH] wayland-client: Treat EOF when reading the wayland socket as an error

[Neil Roberts]

> the only thing that confused me was the past tense sentence […]

Ok, that sounds sensible. Here is an updated patch.

> However we'd have to change every client / toolkit and so using a
> fatal display error is nicer since the toolkits should already be
> handling that.

Well, that would be ideal but it looks like actually GTK and
Clutter/Cogl don't currently handle any errors so they will both end
up constantly waking up the main loop and hitting 100% CPU. Maybe we
should make them both abort on error like what would happen if you
were using Xlib by default.

Regards,
- Neil

-- >8 --

If EOF is encountered while reading from the Wayland socket, make
wl_display_read_events() return -1 so that it will be treated as an
error. The documentation for this function states that it will set
errno when there is an error so it additionally makes up an errno of
EPIPE.

If we don't do this then when the compositor quits the Wayland socket
will be become ready for reading but wl_display_dispatch will do
nothing which typically makes the application take up 100% CPU. In
particular eglSwapBuffers will likely get stuck in an infinite busy
loop because it repeatedly calls wl_display_dispatch_queue while it
waits for the frame callback.
---
 src/wayland-client.c | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/src/wayland-client.c b/src/wayland-client.c
index cb091ab..9400bcd 100644
--- a/src/wayland-client.c
+++ b/src/wayland-client.c
@@ -877,6 +877,15 @@ read_events(struct wl_display *display)
 				display_fatal_error(display, errno);
 			return -1;
 		}
+		else if (total == 0) {
+			/* The compositor has closed the socket. This
+			 * should be considered an error so we'll fake
+			 * an errno */
+			errno = EPIPE;
+			display_fatal_error(display, errno);
+			return -1;
+		}
+
 		for (rem = total; rem >= 8; rem -= size) {
 			size = queue_event(display, rem);
 			if (size == -1) {
-- 
1.7.11.3.g3c3efa5