Properly handle "meta_wayland_buffer_process_damage: assertion 'buffer->resource' failed"
If a client calles wl_surface_damage_buffer()
on a surface without buffer attached (e.g. when using a shm-buffer that already got released) we run into a g_return_if_fail()
, logging the message from the title.
This is currently not explicitly forbidden by the Wayland protocol, so we should handle it regularly and not in an assertion. Also, we should probably log something about a buggy client.
See also !1143 (closed)