Re: w3m-fontify

[Katsumi Yamaoka <yamaoka@xxxxxxx>]

>>>>> In [emacs-w3m : No.10442] Olivier Klein wrote:

> Here is the log:

> Debugger entered--Lisp error: (wrong-type-argument integer-or-marker-p nil)
>   goto-char(nil)
>   (if (eq (char-after ...) 59) nil (goto-char (match-beginning 2)))
>   (unless (eq (char-after ...) 59) (goto-char (match-beginning 2)))
[...]

Ok.  The version of emacs-w3m you use is a bit old.  At the end
of April, Naohiro Aota replaced `(match-beginning 2)' with
`(match-end 1)' in the `w3m-decode-entities' function in order
to cope with an Emacs 21 bug.  He reported in the article
<http://article.gmane.org/gmane.emacs.w3m/7466> that Emacs 21
works strangely with a certain regular expression.  Although he
didn't mention that it applies to also Emacs 23, it might happen
according to the regexp engine that Emacs 23 uses.  In my Fedora
9 Linux, I see nothing strange in Emacs 21 through 23.  The code
for testing it is:

(progn
  (require 'w3m)
  (mapcar (lambda (str)
	    (list (string-match w3m-entity-regexp str)
		  (match-string 1 str)
		  (match-string 2 str)))
	  '("&lt" "&lta" "&lt@" "&lt;")))

;; Copy those lines into the *scratch* buffer, type `C-j' at the
;; end of the last line, and you will see the result.

The result should be:
((0 "lt" "") (nil "lt" "") (0 "lt" "@") (0 "lt" ";"))

Even if you get a different result, I believe it is not so serious
because the value of `w3m-entity-regexp' is much complicated and
such one will be rarely used.  Anyway I think using the latest
emacs-w3m is better.  Here it is:

http://cvs.namazu.org/emacs-w3m.tar.gz

> PS: The problem is more serious than that. Once it occurs, it messes
> up all my email headers???

> PPS: Indeed, it is not related to the content of the .emacs, it might
> have something to do with the order at which the functions are
> called. BTW: same symptoms w/ my .gnus: If I do a change to my
> dot-gnus.el, emacs recompiles the file and the error message
> disappears during the first use, but then it reappears at subsequent
> call???

Hm, I guess compiling a file and loading it might cause something
wrong in the Emacs session, because it loads the file twice.  I.e.
the first one takes place when compiling.

> PPPS: I use your message-multiple-frames.el function :-) !!!

I hope it is not the cause of your problem. ;-)

> I have included the email below: I do not notice anything special???

I tried reading it in Gnus and saw no problem.  (I use the latest
Emacs 23, the latest Gnus, the latest emacs-w3m, ...)

Regards,