expect_protocol_error() ensures that the connection has failed in the expected way. To satisfy ASan leak detection, we still need to tear down everything, including call client_destroy(). client_destroy() needs to check that tear-down does not cause protocol errors, so it does one last roundtrip that checks that is succeeds. But if the connection is already down in an expected way, this roundtrip cannot succeed and must be skipped. Also moves the roundtrip under 'if (wl_display)', assuming the 'if' is there for a reason - but obviously that reason was never used as it would have crashed. Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>dev
parent
6ae243f91e
commit
ed6df8ed1c
Loading…
Reference in new issue