coding system in ChangeLog text

[DJ Delorie]

What coding system are we following in ChangeLogs?  We're starting to
collect various 8-bit characters in people's names (which is a good
sign, I suppose) but Emacs complained that it couldn't figure out the
encoding this last time around.  A quick egrep shows a mix of
obviously iso-latin-1 characters and things that look like wide
characters.

For example, just today:

Rafael \201<A-tilde>\201vila de Esp\201<A-tilde>\201­ndola <rafael.espindola@gmail.com>

And compare:

1998: Martin von L<A-tilde><paragraph>wis  <loewis@informatik.hu-berlin.de>
2000: Martin v. L<o-umlaut>wis  <loewis@informatik.hu-berlin.de>