updated coding style on character encoding, etc.

[Tom Yu]

I added what I hope is an uncontroversial rule to our coding style to
require UTF-8 encoding and to avoid characters outside US-ASCII except
for uses like spelling proper names.

    http://k5wiki.kerberos.org/wiki/Coding_style/Formatting

This was motivated by doing some copyright notice cleanup and
observing some inconsistent character encodings (ISO-8859-1 vs UTF-8
for some proper names, curly quotation marks, etc.) while doing so.