http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56815
--- Comment #11 from Paolo Carlini <paolo.carlini at oracle dot com> 2013-04-03 11:35:55 UTC --- In that case we can have a pedwarn instead of a permerror in the C++ front-end and restore some consistency. Still I find weird that those people in the wild currently must always use p++ and avoid p + 1. Maybe the compromise I was suggesting, thus change in C++ the pedwarn in the common code to permerror but still protected with (pedantic || warn_pointer_arith) is a decent compromise.