Gladly, the creators of UTF-18 did have that foresight so at least we don't have this problem at the code unit -> code point level.