https://gcc.gnu.org/bugzilla/show_bug.cgi?id=80944
--- Comment #3 from Martin Sebor <msebor at gcc dot gnu.org> --- Good point! It never occurred to me. That explains why no compiler does it. That suggests that the optimization opportunity is up to the user. I.e., a class like std::string could optimize this case by adding a constraint/assert that s->s doesn't point at s. Let me look into that.
