https://gcc.gnu.org/bugzilla/show_bug.cgi?id=84360
Jakub Jelinek <jakub at gcc dot gnu.org> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |jakub at gcc dot gnu.org --- Comment #1 from Jakub Jelinek <jakub at gcc dot gnu.org> --- cp/error.c has code to avoid printing aka same as the first type: /* If they are identical, cut off the aka by unwinding the obstack. */ if (type_len == aka_len && memcmp (p + type_start, p+aka_start, type_len) == 0) but that doesn't trigger in this case, because p + type_start with type_len length is "using T = class std::tuple<int, int, int>" and p + aka_start with aka_len is just: "class std::tuple<int, int, int>"