The valgrind-induced race-triggered test-failure was due to the test client code, see the comment and change at l555 (vs the previous location of mir_surface_set_event_handler).
I think EventDelegate should be moved to the surface_create arguments. As it stands now mir_surface_create_sync should be considered dangerous.
The valgrind-induced race-triggered test-failure was due to the test client code, see the comment and change at l555 (vs the previous location of mir_surface_ set_event_ handler) .
I think EventDelegate should be moved to the surface_create arguments. As it stands now mir_surface_ create_ sync should be considered dangerous.