https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65168
--- Comment #12 from Manuel López-Ibáñez <manu at gcc dot gnu.org> --- (In reply to Jonathan Wakely from comment #11) > Great, then my suggestion would be to use exactly the same wording for > references. I guess the Clang devs added their wording because someone mistakenly assumed (like in Jan's original testcase) that applying & to a reference bounded to *p, when p is NULL may return false. That said, I find their wording very confusing. Also, above you said that "&r is not the address of a reference, it's the address of whatever the reference is bound to", thus using the same working seems confusing. But I trust the person that submits a proper patch to choose whatever wording they feel appropriate.
