unfold MIME header lines

[Katsumi Yamaoka <yamaoka@jpl.org>]

Hi,

This article has the following Subject header and you may see
four lines just as it is.

Subject: =?us-ascii?Q?unfold?=
 MIME
 =?us-ascii?Q?header?=
 lines

Things like this take place frequently in Japanse (or any other
non-ascii) messages and they are quite ugly.  Gnus uses the
function `mail-decode-encoded-word-region' by default to decode
MIME headers which only removes a newline and whitespaces
between encoded words.  It is correct in the eyes of RFC2231 but
it is unsatisfactory for Gnus as a living utensil.

Here is a hack to remove a newline and whitespaces between
encoded word and non-encoded word as well.  It can be merged in
the function `rfc2047-decode-region' for the Gnus' default.

(defadvice mail-decode-encoded-word-region
  (before remove-newline-and-whitespaces (start end) activate)
  "Remove excessive newline and whitespaces around encoded words."
  (save-excursion
    (save-restriction
      (narrow-to-region start end)
      (let ((regexp (concat "\\(\\([\t ]*\n\\)?[\t ]+\\)"
			    rfc2047-encoded-word-regexp)))
	(goto-char (point-min))
	(while (re-search-forward rfc2047-encoded-word-regexp nil t)
	  (while (looking-at regexp)
	    (goto-char (match-end 0))
	    (delete-region (match-beginning 1) (match-end 1)))
	  (if (looking-at "\\([\t ]*\n\\)?[\t ]+")
	      (progn
		(delete-region (match-beginning 0) (match-end 0))
		(insert " "))))
	(goto-char (point-min))
	(while (re-search-forward regexp nil t)
	  (goto-char (match-beginning 1))
	  (delete-region (point) (match-end 1))
	  (insert " ")))
      (setq end (point-max)))))
-- 
Katsumi Yamaoka <yamaoka@jpl.org>