Re: ispell no longer works as expected (process-send-string problem?)

[Kenichi Handa]

In article <address@hidden>, James Felix Black <address@hidden> writes:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1

> % emacs -Q
> M-x load-library ispell
> foo
> M-$

> Yields the following in the minibuffer:
> Wrong type argument: integer-or-marker-p, "^such
> "

> And, in *Messages*:
> Checking spelling of FOO...
> process-send-string: Wrong type argument: integer-or-marker-p, "^foo
> "

> This is a build from the repository, taken at about midnight on 7  
> March 2007:

If you do that after M-x toggle-debug-on-error RET, and
*Backtrace* buffer shows that the error is signaled in the
function utf-8-pre-write-conversion, please check if
lisp/international/utf-8.el has this code for
utf-8-pre-write-conversion:

(defun utf-8-pre-write-conversion (beg end)
  "Prepare for `utf-translate-cjk-mode' to encode text between BEG and END.
This is used as a post-read-conversion of utf-8 coding system."
  (if (and utf-translate-cjk-mode
           (not utf-translate-cjk-lang-env)
           (if (stringp beg)
               (string-match "\\cc\\|\\cj\\|\\ch" beg)
             (save-excursion
               (goto-char beg)
               (re-search-forward "\\cc\\|\\cj\\|\\ch" end t))))
      (utf-translate-cjk-load-tables))
  nil)

and it is byte-compiled BEFORE you built Emacs.

---
Kenichi Handa
address@hidden