@chucker I feel quite confident that any correction will be towards the UTF-16 interpretation, for “compatibility”