The standard units data file is written in Unicode using the UTF-8 encoding. Portions of the file that are not plain ASCII begin with ‘!utf8’ and end with ‘!endutf8’. As usual, the ‘!’ must appear as the first character on the line. If a line of a data file contains byte sequences that are invalid UTF-8 or non-printing UTF-8 then units ignores the entire line.
When units runs it checks the locale to determine the character set. If UTF-8 is listed, then units reads the utf8 definitions. If any other character set is in use, then units works in plain ASCII without support for extended characters.