https://gcc.gnu.org/bugzilla/show_bug.cgi?id=119429
--- Comment #13 from Jonathan Wakely <redi at gcc dot gnu.org> --- (In reply to Desmond Rhodes from comment #7) > If the construct of `size_t __nargs = -1` isn't fixed, it will prevent > people from using the standard library. A bold claim! std::string has been relying on the same conversion since the first C++ standard in 1998, for the npos constant.
