https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93231
--- Comment #3 from Jakub Jelinek <jakub at gcc dot gnu.org> --- Perhaps with the just added native_encode_initializer, rather than having separate function for CONSTRUCTOR and STRING_CST it might be better to always native_encode_initializer at certain offset with byte length of the array element type, verify it returned the passed in length and then native_interpret_expr it?
