Technically, byte-strings (and nested comments) mean the lexer is not regular, but the grammar is still context-free. But anyway, parsing is not typically the bottle-neck these days.