still codings problem (saving) with babel.

[Stephen J. Turnbull]

Uwe Brauer writes:

 > Char: é (U+2000E9 jit-ucs-charset-0 35 56) point=467 of 473(99%) column 2 

Maybe Aidan can explain what looks like a Latin-1 character is being
stuffed into a jit-ucs charset.  I would think that shouldn't happen,
but maybe Babelfish is weirdly presenting you with a non-Unicode
character or there's some other upstream reason (ie, in whatever code
is fetching and presenting the result of our query) why this is
happening.

In the meantime, try running this function over the buffer:

(defun force-latin-1 ()
  (interactive)
  (save-excursion
    (goto-char (point-min))
    (while (not (eobp))
      (let ((code (char-to-unicode (char-after))))
        (when (and (> code (* 17 65536))
                   (< (% code 65536) 256))
          (insert (unicode-to-char (% code 65536)))
          (delete-char))
        (forward-char)))))

I don't guarantee this won't corrupt things.