* doc/lispref/strings.texi (Text Comparison): The old text was wrong: non-ASCII unibyte chars overlap part of Unicode.