https://gcc.gnu.org/bugzilla/show_bug.cgi?id=87957
--- Comment #32 from Jan Hubicka <hubicka at ucw dot cz> --- > I guess we might end up streaming stuff we don't need. Can't we simply > remove the assert? We do build the copy using the main variant type > so this seems to be just a consistency check. The consistency check prevents code from creating duplicated copies of TYPE_DECL (if multiple types could reffer to one TYPE_DECL we would copy it each time we copy a type). We really should rewrite it to IDENTIFIER_TYPE for middle-end purposes because TYPE_DECLs are useful only for C++ ODR types. It just appeared that tings works otherwise (i.e. we do not try to produce mangled name for Ada types) because need_assembler_name_p return false or Ada frontend makes assembler names of TYPE_DECLs NULL anyway. I will try adding that check. Honza
