Void variable error in fontification code

Void variable error in fontification code

[Davin Pearson]

How come with the following lines of elisp code commented out:

;;(setq d-face-m4     'd-face-m4)
;;(setq d-face-m4-dnl 'd-face-m4-dnl)

I get the following error in my *Messages* buffer:

Error during redisplay: (jit-lock-function 9575) signaled (void-variable d-face-m4-dnl) [9 times]

Putting back the lines like so:

(setq d-face-m4     'd-face-m4)
(setq d-face-m4-dnl 'd-face-m4-dnl)

gets rid of the error messages.

I have the following lines of elisp code in my .emacs file:

(defun d-font-lock-add-begin (keywords)
  (if (fboundp 'font-lock-add-keywords)
      (font-lock-add-keywords nil keywords nil)
    (setq font-lock-keywords
          (append
           keywords
           font-lock-keywords))))

(d-font-lock-add-begin
  `(
      (,(concat "\\(\\<m4_" "dnl\\)\\(.$\\)") (1 d-face-m4-dnl t) (2 font-lock-comment-face))
   )
)

Re: Void variable error in fontification code

[Michael Heerdegen]

Davin Pearson <davin.pearson@gmail.com> writes:

> (setq d-face-m4     'd-face-m4)
> (setq d-face-m4-dnl 'd-face-m4-dnl)

>       (,(concat "\\(\\<m4_" "dnl\\)\\(.$\\)") (1 d-face-m4-dnl t) (2 font-lock-comment-face))

I think you get confused by the fact that symbols are used as face
names.  For font lock keywords, faces are specified via face
expressions.  You specify a variable above, so its value has to be set
(to a face name), else you get the unbound error.

Instead, you can directly specify a face in form of a symbol, and since
this place is evaluated, you need to quote the name.

It's a bit weird in Emacs that faces effectively use a separate name
space, but they are specified as symbols, and in some cases, variables
having the same name as a face are bound to that face name (symbol).


Michael.