Calling C$FILEINFO and C$COPY using a file with a name that includes a diacritical character does not work.
The data item contains the correct file name but neither C$FILEINFO or C$COPY will find the file.
The UNICODE changes in 10.1 were strictly for certain aspects of the runtime:
Not all aspects of the runtime handle UNICODE. For example, reference modification still works byte-by-byte, and not character-by-character.
Although the filename may appear correct in the program, that is no guarantee that the standard library routines will support it.