Correct invalid-sequence-coding-system spec, Roman-alphabet languages. I had been testing with the Cyrillic language environments, and I have code in my own init file that does something similar, so I hadn't noticed that this had gone wrong. lisp/ChangeLog addition: 2009-02-04 Aidan Kehoe <kehoea@parhasard.net> * mule/latin.el: Specify windows-1250 as the invalid-sequence-coding-system for the iso-8859-2 languages; actually *use* the invalid-sequence-coding-system for German and the other iso-8859-1 language environments.