Code review comment for lp:~alan-griffiths/mir/drag-and-drop-II

> 18:39:25 [ RUN ]
> DragAndDrop.when_drag_moves_from_window_leave_event_contains_cookie
> ...
> 18:39:25 /<<BUILDDIR>>/mir-0.26.0+zesty4293bzr4086/tests/acceptance-
> tests/drag_and_drop.cpp:256: Failure
> 18:39:25 Value of: have_cookie.wait_for(receive_event_timeout)
> 18:39:25 Expected: is equal to true
> 18:39:25 Actual: false (of type bool)
> 18:39:25 ==11713== Invalid read of size 8
>
> The memory errors and FD leaks that follow result from following the resulting
> nullptr.

Hmm wrong paste buffer, but still I don't see how this can happen:

18:35:34 11: [ RUN ] DragAndDrop.when_drag_moves_from_window_leave_event_contains_cookie
...
18:37:06 11: /<<BUILDDIR>>/mir-0.26.0+zesty4293bzr4086/tests/acceptance-tests/drag_and_drop.cpp:256: Failure
18:37:06 11: Value of: have_cookie.wait_for(receive_event_timeout)
18:37:06 11: Expected: is equal to true
18:37:06 11: Actual: false (of type bool)
18:37:06 11: ==11713== Invalid read of size 8

Basically, the click isn't seen by the window. And on this branch the window has definitely received {mir_window_attrib_focus, true}