Mule bugs: misidentification (Latin-1 vs. Chinese), revert issues

[Nix]

On 23 Oct 2006, stephen at xemacs.org uttered the following:
> AFAIK you should never get "Auto" in the modeline; that should always
> be replaced with a real coding system ("Auto" is short for
> "autodetect").

If that `never' is `never ever ever', well, I have a *lot* of Gnus
Article buffers here with `Auto' in the modeline.

Nothing goes noticeably wrong as a result (probably because they're all
boring as sin ISO-8859-1 buffers).

My .emacs does contain:

(set-coding-priority-list '(utf-8 iso-8-1 utf-8-bom)) ; Use UTF-8 or ISO-8859-x
(set-language-environment "English")            ; A reasonable default

... but that shouldn't cause this, should it?

-- 
`When we are born we have plenty of Hydrogen but as we age our
 Hydrogen pool becomes depleted.'