and mark it as finished in README. Upon further checking, expr(1) turns out to be implicitly UTF-8 compliant.