Warning reference will be nullptred bug

The disappearance of the issue was indeed a false positive caused by our sanitiser - it is publicly available again (and still marked unresolved because of the issue unrelated to this regression):

Steve