Re: font-lock question

[Javier Oviedo]

 Hi Alex.

Actually, I implemented something very similar to this exact code. My
motivation is that the occur mode buffer does not display the matches in the
original buffers faces. I was just curious if I could hack something and
decided to implement this myself...though an obviously vicious hack(see
below). I decided to do this more for fun then anything else...I just wanted
to see if I could figure it out. (Kevin Rogers suggests that I should submit
a bug report.)

This little hack works when running occur on a emacs-lisp-mode buffer...I
was going to add highlighting using overlays next, assuming that I could
figure that out as well. :-)

(defvar joviedo-occur-font-lock-default nil)
(defadvice occur (after joviedo-occur last activate)
  (setq joviedo-occur-font-lock-default font-lock-defaults)
  (pop-to-buffer "*Occur*")
  (set (make-local-variable 'font-lock-defaults)
joviedo-occur-font-lock-default)
  (font-lock-mode nil)
  (font-lock-fontify-buffer)

;;;still need to add highlights using overlays.

  )


One issue, however, is that this does not work for a c-mode buffer. I get
some sort of strange error that I can't figure out. I thought the principle
was relatively sound but I guess not. The error in the *Message* is the
following:

Wrong type argument: stringp, nil
Error in post-command-hook: (wrong-type-argument stringp nil)

When I examine the value for font-lock-defaults(after my defadvice runs) I
see the following:

font-lock-defaults's value is
((c-font-lock-keywords c-font-lock-keywords-1 c-font-lock-keywords-2
c-font-lock-keywords-3)
 nil nil
 ((95 . "w")
  (36 . "w"))
 c-beginning-of-syntax
 (font-lock-mark-block-function . c-mark-function))


This is exactly what I get if I examine the value of font-lock-defaults
before ever running occur. For some reason, there is something wrong with
this value(list) but I can't figure it out.

Any thoughts?????? help???

-- 
Javier


"Alan Mackenzie" <none@example.invalid> wrote in message
thi80c.q5.ln@acm.acm">news:thi80c.q5.ln@acm.acm...
> Javier Oviedo <email_joviedo@yahoo.com> wrote on Mon, 9 Feb 2004 11:18:00
> -0500:
> > Hello all:
>
> > This will probably sound like a rather bizarre question but here goes...
>
> Bizarre indeed!
>
> > Is there a way to fontify a buffer based on a different mode but still
> > retain all of the original/base modes attributes(expect for the face
> > attributes, of course)?
>
> Yes.  The answer is _always_ yes in Emacs.  :-)
>
> > Example:  I have c-file open in emacs and the buffer is using c-mode.
> > I'd like to use the faces for emacs-lisp-mode(this just an example, so
> > please humor me) but keep every other attribute of the c-mode. Can this
> > be done? Is there some other way I could accomplish this? Perhaps using
> > a minor mode and it's faces but keeping the attributes of the major
> > mode? I'm open to any and all suggestions.
>
> How about something like this (not tested):
>
> (defun jo-funny-font ()
>   (font-lock-mode -1)
> ;; The next line is taken from lisp-mode-variables in lisp-mode.el
>   (setq font-lock-defaults
>         '((lisp-font-lock-keywords
>            lisp-font-lock-keywords-1 lisp-font-lock-keywords-2)
>           nil nil (("+-*/.<>=!?$%_&~^:" . "w")) beginning-of-defun
>           (font-lock-mark-block-function . mark-defun)))
>   (font-lock-mode 1))
> (add-hook 'c-mode-hook 'jo-funny-font)
>
> But why do you want to do this?
>
> > Thanks in advance!
>
> > Javier
>
> -- 
> Alan Mackenzie (Munich, Germany)
> Email: aacm@muuc.dee; to decode, wherever there is a repeated letter
> (like "aa"), remove half of them (leaving, say, "a").
>