CODEWITHSUNDEEP
emacswolfram-mathematicaelisp
Joe
Joe

Reputation: 1045

How to solve character encoding issue with Emacs, mathematica.el, and Mathematica 10?

I am successfully using mathematica.el in Emacs with Mathematica 10. However, I have some kind of issue with character encoding whereby when I call mathematica-execute on a Mathematica expression the result prints lots of ^M characters to the output (I am running all processes on OS X Mavericks). I initially made sure that the $CharacterEncoding in Mathematica matched the file and process encoding of Emacs for the buffers being used as best I could: utf-8. There could still be an issue there, but that route didn't yield a solution. Next I thought, why not just create my own function that calls mathematica-execute and then removes the ^M characters from the resulting output. Here is the relevant code:

(defun delete-to-out ()
     (next-line)
     (set-mark-command nil)
     (search-forward "]=")
     (previous-line)
     (delete-region (region-beginning) (region-end)))

(defun remove-^M ()
  "Get rid of ^M characters"
  (interactive)
  (message "remove-^M called!")
  (save-excursion
    (goto-char (point-min))
    (while (re-search-forward "\r" nil t)
      (replace-match ""))))

(defun my-mathematica-execute (arg)
  "Call mathematica-execute and then clean out the ^M characters
  it inserts by calling remove-^M"
  (interactive "P")
  (save-excursion
    (mathematica-execute arg)
    (remove-^M)
    ;; (delete-to-out)
    ))

;;; Define some mathematica mode keyboard bindings
(if mathematica-mode-map
    (progn
      (define-key mathematica-mode-map [remap mathematica-execute]
        'my-mathematica-execute)
      (define-key mathematica-mode-map (kbd "C-c m")
        'remove-^M)))

Calling my-mathematica-execute works, but isn't successful at removing the ^M characters. Calling remove-^M or delete-to-out work fine in isolation subsequently. It must be a synchronization issue or timing issue? Any help would be appreciated. Alternative solutions are certainly welcome.

Upvotes: 0

Views: 166

Answers (1)

Stefan
Stefan

Reputation: 28571

The ^M indeed look like decoding issues, but it's not the character-encoding that's the problem but the EOL-encoding. Both are specified together usually in Emacs, so you probably need to use utf-8-dos instead of utf-8 somewhere.

Upvotes: 1

Related Questions