Since you reported the crashes upstream, and did not respond to the request from comment #6, maybe a fix may come from upstream.