unicode.utf8.cfind
It is a helper function. It behaves like
unicode.utf8.find except that it treats its input offset and
expresses its output offsets in characters rather than bytes.
(The only exception are “position captures” that are still returned
expressed in bytes).