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

[stephen at xemacs.org]

Aidan Kehoe writes:

 >  > I open a file with UTF-8 coding-system.  I touch that file outside of
 >  > XEmacs and then do M-x revert-buffer RET.  The non-ASCII characters
 >  > get mangled.
 > 
 > There?s a comment from Ben in the sources about this problem: 

What does Ben's comment have to do with coding systems, other than
that they reduce the utility of replace-mode to zero?

Note that 

      if (!do_speedy_insert)
	buffer_delete_range (buf, BUF_BEG (buf), BUF_Z (buf),
			     !NILP (visit) ? INSDEL_NO_LOCKING : 0);
      else
        /* The code that the comment refers to. */

where

  int do_speedy_insert =
    coding_system_is_binary (Fget_coding_system (codesys));

Ie, for non-binary coding systems, you just delete the region, and
insert the file according the `codesys'.  Am I missing something?

I'm starting to get an inkling of why GNU had such trouble exorcising
the \201 devil. ;-)