1 ;;; rfc2047.el --- functions for encoding and decoding rfc2047 messages
3 ;; Copyright (C) 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006,
4 ;; 2007, 2008, 2009, 2010, 2011 Free Software Foundation, Inc.
6 ;; Author: Lars Magne Ingebrigtsen <larsi@gnus.org>
7 ;; MORIOKA Tomohiko <morioka@jaist.ac.jp>
8 ;; This file is part of GNU Emacs.
10 ;; GNU Emacs is free software: you can redistribute it and/or modify
11 ;; it under the terms of the GNU General Public License as published by
12 ;; the Free Software Foundation, either version 3 of the License, or
13 ;; (at your option) any later version.
15 ;; GNU Emacs is distributed in the hope that it will be useful,
16 ;; but WITHOUT ANY WARRANTY; without even the implied warranty of
17 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
18 ;; GNU General Public License for more details.
20 ;; You should have received a copy of the GNU General Public License
21 ;; along with GNU Emacs. If not, see <http://www.gnu.org/licenses/>.
25 ;; RFC 2047 is "MIME (Multipurpose Internet Mail Extensions) Part
26 ;; Three: Message Header Extensions for Non-ASCII Text".
32 (defvar message-posting-charset
)
36 ;; Fixme: Avoid this (used for mail-parse-charset) mm dependence on gnus.
38 (require 'rfc2045
) ;; rfc2045-encode-string
39 (autoload 'mm-body-7-or-8
"mm-bodies")
41 (defvar rfc2047-header-encoding-alist
42 '(("Newsgroups" . nil
)
45 ("\\(Resent-\\)?\\(From\\|Cc\\|To\\|Bcc\\|\\(In-\\)?Reply-To\\|Sender\
46 \\|Mail-Followup-To\\|Mail-Copies-To\\|Approved\\)" . address-mime
)
48 "*Header/encoding method alist.
49 The list is traversed sequentially. The keys can either be
54 1) nil, in which case no encoding is done;
55 2) `mime', in which case the header will be encoded according to RFC2047;
56 3) `address-mime', like `mime', but takes account of the rules for address
57 fields (where quoted strings and comments must be treated separately);
58 4) a charset, in which case it will be encoded as that charset;
59 5) `default', in which case the field will be encoded as the rest
62 (defvar rfc2047-charset-encoding-alist
88 "Alist of MIME charsets to RFC2047 encodings.
89 Valid encodings are nil, `Q' and `B'. These indicate binary (no) encoding,
90 quoted-printable and base64 respectively.")
92 (defvar rfc2047-encode-function-alist
93 '((Q . rfc2047-q-encode-string
)
94 (B . rfc2047-b-encode-string
)
96 "Alist of RFC2047 encodings to encoding functions.")
98 (defvar rfc2047-encode-encoded-words t
99 "Whether encoded words should be encoded again.")
101 (defvar rfc2047-allow-irregular-q-encoded-words t
102 "*Whether to decode irregular Q-encoded words.")
104 (eval-and-compile ;; Necessary to hard code them in `rfc2047-decode-region'.
105 (defconst rfc2047-encoded-word-regexp
106 "=\\?\\([^][\000-\040()<>@,\;:*\\\"/?.=]+\\)\\(?:\\*[^?]+\\)?\\?\
107 \\(B\\?[+/0-9A-Za-z]*=*\
110 "Regexp that matches encoded word."
111 ;; The patterns for the B encoding and the Q encoding, i.e. the ones
112 ;; beginning with "B" and "Q" respectively, are restricted into only
113 ;; the characters that those encodings may generally use.
115 (defconst rfc2047-encoded-word-regexp-loose
116 "=\\?\\([^][\000-\040()<>@,\;:*\\\"/?.=]+\\)\\(?:\\*[^?]+\\)?\\?\
117 \\(B\\?[+/0-9A-Za-z]*=*\
118 \\|Q\\?\\(?:\\?+[ -<>@-~]\\)?\\(?:[ ->@-~]+\\?+[ -<>@-~]\\)*[ ->@-~]*\\?*\
120 "Regexp that matches encoded word allowing loose Q encoding."
121 ;; The pattern for the Q encoding, i.e. the one beginning with "Q",
123 ;; "Q\\?\\(\\?+[^\n=?]\\)?\\([^\n?]+\\?+[^\n=?]\\)*[^\n?]*\\?*"
124 ;; <--------1-------><----------2,3----------><--4--><-5->
126 ;; 1. After "Q?", allow "?"s that follow a character other than "=".
127 ;; 2. Allow "=" after "Q?"; it isn't regarded as the terminator.
128 ;; 3. In the middle of an encoded word, allow "?"s that follow a
129 ;; character other than "=".
130 ;; 4. Allow any characters other than "?" in the middle of an
132 ;; 5. At the end, allow "?"s.
136 ;;; Functions for encoding RFC2047 messages
139 (defun rfc2047-qp-or-base64 ()
140 "Return the type with which to encode the buffer.
141 This is either `base64' or `quoted-printable'."
143 (let ((limit (min (point-max) (+ 2000 (point-min))))
145 (goto-char (point-min))
146 (skip-chars-forward "\x20-\x7f\r\n\t" limit
)
147 (while (< (point) limit
)
150 (skip-chars-forward "\x20-\x7f\r\n\t" limit
))
151 (if (or (< (* 6 n8bit
) (- limit
(point-min)))
152 ;; Don't base64, say, a short line with a single
153 ;; non-ASCII char when splitting parts by charset.
158 (defun rfc2047-narrow-to-field ()
159 "Narrow the buffer to the header on the current line."
165 (if (re-search-forward "^[^ \n\t]" nil t
)
168 (goto-char (point-min)))
170 (defun rfc2047-field-value ()
171 "Return the value of the field at point."
174 (rfc2047-narrow-to-field)
175 (re-search-forward ":[ \t\n]*" nil t
)
176 (buffer-substring-no-properties (point) (point-max)))))
178 (defun rfc2047-quote-special-characters-in-quoted-strings (&optional
180 "Quote special characters with `\\'s in quoted strings.
181 Quoting will not be done in a quoted string if it contains characters
182 matching ENCODABLE-REGEXP or it is within parentheses."
183 (goto-char (point-min))
184 (let ((tspecials (concat "[" ietf-drums-tspecials
"]"))
187 (with-syntax-table (standard-syntax-table)
191 (eq (char-before) ?\
)))
193 (goto-char (point-max)))
195 (narrow-to-region start
(point))
197 (while (search-forward "\"" nil t
)
198 (setq beg
(match-beginning 0))
199 (unless (eq (char-before beg
) ?
\\)
205 (setq end
(1- (point)))
207 (if (and encodable-regexp
208 (re-search-forward encodable-regexp end t
))
211 (narrow-to-region beg end
)
212 (while (re-search-forward tspecials nil
'move
)
213 (if (eq (char-before) ?
\\)
214 (if (looking-at tspecials
) ;; Already quoted.
217 (goto-char (match-beginning 0))
223 (goto-char (point-max)))
225 (setq start
(point))))))
227 (defvar rfc2047-encoding-type
'address-mime
228 "The type of encoding done by `rfc2047-encode-region'.
229 This should be dynamically bound around calls to
230 `rfc2047-encode-region' to either `mime' or `address-mime'. See
231 `rfc2047-header-encoding-alist', for definitions.")
233 (defun rfc2047-encode-message-header ()
234 "Encode the message header according to `rfc2047-header-encoding-alist'.
235 Should be called narrowed to the head of the message."
238 (goto-char (point-min))
239 (let (alist elem method
)
242 (rfc2047-narrow-to-field)
244 alist rfc2047-header-encoding-alist
)
245 (while (setq elem
(pop alist
))
246 (when (or (and (stringp (car elem
))
247 (looking-at (car elem
)))
251 (if (not (rfc2047-encodable-p))
253 (when (eq method
'address-mime
)
254 (rfc2047-quote-special-characters-in-quoted-strings))
255 (if (and (eq (mm-body-7-or-8) '8bit
)
258 (car message-posting-charset
)))
259 ;; 8 bit must be decoded.
260 (mm-encode-coding-region
261 (point-min) (point-max)
262 (mm-charset-to-coding-system
263 (car message-posting-charset
))))
264 ;; No encoding necessary, but folding is nice
268 (goto-char (point-min))
269 (skip-chars-forward "^:")
270 (when (looking-at ": ")
274 ;; We found something that may perhaps be encoded.
275 (re-search-forward "^[^:]+: *" nil t
)
277 ((eq method
'address-mime
)
278 (rfc2047-encode-region (point) (point-max)))
280 (let ((rfc2047-encoding-type 'mime
))
281 (rfc2047-encode-region (point) (point-max))))
282 ((eq method
'default
)
283 (if (and (featurep 'mule
)
284 (if (boundp 'enable-multibyte-characters
)
285 (default-value 'enable-multibyte-characters
))
287 (mm-encode-coding-region (point) (point-max)
288 mail-parse-charset
)))
289 ;; We get this when CC'ing messsages to newsgroups with
290 ;; 8-bit names. The group name mail copy just got
291 ;; unconditionally encoded. Previously, it would ask
292 ;; whether to encode, which was quite confusing for the
293 ;; user. If the new behavior is wrong, tell me. I have
294 ;; left the old code commented out below.
295 ;; -- Per Abrahamsen <abraham@dina.kvl.dk> Date: 2001-10-07.
296 ;; Modified by Dave Love, with the commented-out code changed
297 ;; in accordance with changes elsewhere.
299 (rfc2047-encode-region (point) (point-max)))
301 ;;; (if (or (message-options-get
302 ;;; 'rfc2047-encode-message-header-encode-any)
303 ;;; (message-options-set
304 ;;; 'rfc2047-encode-message-header-encode-any
306 ;;; "Some texts are not encoded. Encode anyway?")))
307 ;;; (rfc2047-encode-region (point-min) (point-max))
308 ;;; (error "Cannot send unencoded text")))
309 ((mm-coding-system-p method
)
310 (if (or (and (featurep 'mule
)
311 (if (boundp 'enable-multibyte-characters
)
312 (default-value 'enable-multibyte-characters
)))
313 (featurep 'file-coding
))
314 (mm-encode-coding-region (point) (point-max) method
)))
317 (goto-char (point-max)))))))
319 ;; Fixme: This, and the require below may not be the Right Thing, but
320 ;; should be safe just before release. -- fx 2001-02-08
322 (defun rfc2047-encodable-p ()
323 "Return non-nil if any characters in current buffer need encoding in headers.
324 The buffer may be narrowed."
325 (require 'message
) ; for message-posting-charset
327 (mm-find-mime-charset-region (point-min) (point-max))))
328 (goto-char (point-min))
329 (or (and rfc2047-encode-encoded-words
331 (re-search-forward rfc2047-encoded-word-regexp nil t
)
332 (goto-char (point-min))))
334 (not (equal charsets
(list (car message-posting-charset
))))))))
336 ;; Use this syntax table when parsing into regions that may need
337 ;; encoding. Double quotes are string delimiters, backslash is
338 ;; character quoting, and all other RFC 2822 special characters are
339 ;; treated as punctuation so we can use forward-sexp/forward-word to
340 ;; skip to the end of regions appropriately. Nb. ietf-drums does
341 ;; things differently.
342 (defconst rfc2047-syntax-table
343 ;; (make-char-table 'syntax-table '(2)) only works in Emacs.
344 (let ((table (make-syntax-table)))
345 ;; The following is done to work for setting all elements of the table;
346 ;; it appears to be the cleanest way.
347 ;; Play safe and don't assume the form of the word syntax entry --
349 (if (featurep 'xemacs
)
350 (put-char-table t
(get-char-table ?a
(standard-syntax-table)) table
)
351 (set-char-table-range table t
(aref (standard-syntax-table) ?a
)))
352 (modify-syntax-entry ?
\\ "\\" table
)
353 (modify-syntax-entry ?
\" "\"" table
)
354 (modify-syntax-entry ?\
( "(" table
)
355 (modify-syntax-entry ?\
) ")" table
)
356 (modify-syntax-entry ?\
< "." table
)
357 (modify-syntax-entry ?\
> "." table
)
358 (modify-syntax-entry ?\
[ "." table
)
359 (modify-syntax-entry ?\
] "." table
)
360 (modify-syntax-entry ?
: "." table
)
361 (modify-syntax-entry ?\
; "." table)
362 (modify-syntax-entry ?
, "." table
)
363 (modify-syntax-entry ?
@ "." table
)
366 (defun rfc2047-encode-region (b e
)
367 "Encode words in region B to E that need encoding.
368 By default, the region is treated as containing RFC2822 addresses.
369 Dynamically bind `rfc2047-encoding-type' to change that."
371 (narrow-to-region b e
)
372 (let ((encodable-regexp (if rfc2047-encode-encoded-words
373 "[^\000-\177]+\\|=\\?"
375 start
; start of current token
377 ;; Whether there's an encoded word before the current token,
378 ;; either immediately or separated by space.
380 (orig-text (buffer-substring-no-properties b e
)))
381 (if (eq 'mime rfc2047-encoding-type
)
382 ;; Simple case. Continuous words in which all those contain
383 ;; non-ASCII characters are encoded collectively. Encoding
384 ;; ASCII words, including `Re:' used in Subject headers, is
385 ;; avoided for interoperability with non-MIME clients and
386 ;; for making it easy to find keywords.
388 (goto-char (point-min))
389 (while (progn (skip-chars-forward " \t\n")
392 (while (and (looking-at "[ \t\n]*\\([^ \t\n]+\\)")
394 (setq end
(match-end 0))
395 (re-search-forward encodable-regexp end t
)))
397 (if (> (point) start
)
398 (rfc2047-encode start
(point))
400 ;; `address-mime' case -- take care of quoted words, comments.
401 (rfc2047-quote-special-characters-in-quoted-strings encodable-regexp
)
402 (with-syntax-table rfc2047-syntax-table
403 (goto-char (point-min))
404 (condition-case err
; in case of unbalanced quotes
405 ;; Look for rfc2822-style: sequences of atoms, quoted
406 ;; strings, specials, whitespace. (Specials mustn't be
410 (skip-chars-forward " \t\n")
413 ((not (char-after))) ; eob
415 ((eq ?
\" (setq csyntax
(char-syntax (char-after))))
419 ;; Does it need encoding?
421 (if (re-search-forward encodable-regexp end
'move
)
422 ;; It needs encoding. Strip the quotes first,
423 ;; since encoded words can't occur in quotes.
430 ;; There was a preceding quoted word. We need
431 ;; to include any separating whitespace in this
432 ;; word to avoid it getting lost.
433 (skip-chars-backward " \t")
434 ;; A space is needed between the encoded words.
438 ;; Adjust the end position for the deleted quotes.
439 (rfc2047-encode start
(- end
2))
440 (setq last-encoded t
)) ; record that it was encoded
441 (setq last-encoded nil
)))
443 ;; Skip other delimiters, but record that they've
444 ;; potentially separated quoted words.
446 (setq last-encoded nil
))
448 (error "Unbalanced parentheses"))
450 ;; Look for the end of parentheses.
452 ;; Encode text as an unstructured field.
453 (let ((rfc2047-encoding-type 'mime
))
454 (rfc2047-encode-region (1+ start
) (1- (point))))
455 (skip-chars-forward ")"))
456 (t ; normal token/whitespace sequence
458 ;; Skip one ASCII word, or encode continuous words
459 ;; in which all those contain non-ASCII characters.
461 (while (not (or end
(eobp)))
462 (when (looking-at "[\000-\177]+")
466 (while (and (or (re-search-forward
467 "[ \t\n]\\|\\Sw" end
'move
)
469 (eq ?
\\ (char-syntax (char-before))))
470 ;; Skip backslash-quoted characters.
473 (setq end
(match-beginning 0))
474 (if rfc2047-encode-encoded-words
477 (when (search-forward "=?" end
'move
)
478 (goto-char (match-beginning 0))
481 ;; Where the value nil of `end' means there may be
482 ;; text to have to be encoded following the point.
483 ;; Otherwise, the point reached to the end of ASCII
484 ;; words separated by whitespace or a special char.
486 (when (looking-at encodable-regexp
)
487 (goto-char (setq begin
(match-end 0)))
488 (while (and (looking-at "[ \t\n]+\\([^ \t\n]+\\)")
489 (setq end
(match-end 0))
491 (while (re-search-forward
492 encodable-regexp end t
))
495 (or (not (re-search-forward "\\Sw" end t
))
497 (goto-char (match-beginning 0))
500 (when (looking-at "[^ \t\n]+")
501 (setq end
(match-end 0))
502 (if (re-search-forward "\\Sw+" end t
)
503 ;; There are special characters better
504 ;; to be encoded so that MTAs may parse
506 (cond ((= end
(point)))
507 ((looking-at (concat "\\sw*\\("
512 (goto-char (1- (match-end 0)))
513 (unless (= (point) (match-beginning 0))
514 ;; Separate encodable text and
518 (skip-chars-forward " \t\n")
519 (if (and (looking-at "[^ \t\n]+")
520 (string-match encodable-regexp
523 (goto-char end
)))))))
524 (skip-chars-backward " \t\n")
527 (if (re-search-forward encodable-regexp end
'move
)
529 (unless (memq (char-before start
) '(nil ?
\t ?
))
532 (skip-chars-backward "^ \t\n")
533 (and (looking-at "\\Sw+")
534 (= (match-end 0) start
)))
535 ;; Also encode bogus delimiters.
537 ;; Separate encodable text and delimiter.
540 (setq start
(1+ start
)
542 (rfc2047-encode start end
)
543 (setq last-encoded t
))
544 (setq last-encoded nil
)))))
546 (if (or debug-on-quit debug-on-error
)
547 (signal (car err
) (cdr err
))
548 (error "Invalid data for rfc2047 encoding: %s"
549 (mm-replace-in-string orig-text
"[ \t\n]+" " "))))))))
550 (rfc2047-fold-region b
(point))
551 (goto-char (point-max))))
553 (defun rfc2047-encode-string (string)
554 "Encode words in STRING.
555 By default, the string is treated as containing addresses (see
556 `rfc2047-encoding-type')."
557 (mm-with-multibyte-buffer
559 (rfc2047-encode-region (point-min) (point-max))
563 ;; 2. Syntax of encoded-words
565 ;; While there is no limit to the length of a multiple-line header
566 ;; field, each line of a header field that contains one or more
567 ;; 'encoded-word's is limited to 76 characters.
569 ;; In `rfc2047-encode-parameter' it is bound to nil, so don't defconst it.
570 (defvar rfc2047-encode-max-chars
76
571 "Maximum characters of each header line that contain encoded-words.
572 According to RFC 2047, it is 76. If it is nil, encoded-words
573 will not be folded. Too small value may cause an error. You
574 should not change this value.")
576 (defun rfc2047-encode-1 (column string cs encoder start crest tail
578 "Subroutine used by `rfc2047-encode'."
579 (cond ((string-equal string
"")
581 ((not rfc2047-encode-max-chars
)
583 (funcall encoder
(if cs
584 (mm-encode-coding-string string cs
)
587 ((>= column rfc2047-encode-max-chars
)
589 (cond ((string-match "\n[ \t]+\\'" eword
)
590 ;; Reomove a superfluous empty line.
591 (setq eword
(substring eword
0 (match-beginning 0))))
592 ((string-match "(+\\'" eword
)
593 ;; Break the line before the open parenthesis.
594 (setq crest
(concat crest
(match-string 0 eword
))
595 eword
(substring eword
0 (match-beginning 0))))))
596 (rfc2047-encode-1 (length crest
) string cs encoder start
" " tail
597 (concat eword
"\n" crest
)))
600 (limit (1- (length string
)))
605 (setq next
(concat start
608 (mm-encode-coding-string
609 (substring string
0 (1+ index
))
611 (substring string
0 (1+ index
))))
613 len
(+ column
(length next
)))
614 (if (> len rfc2047-encode-max-chars
)
617 (if (or (< index limit
)
618 (<= (+ len
(or (string-match "\n" tail
)
620 rfc2047-encode-max-chars
))
623 (if (string-match "\\`)+" tail
)
624 ;; Break the line after the close parenthesis.
625 (setq tail
(concat (substring tail
0 (match-end 0))
627 (substring tail
(match-end 0)))
633 (concat eword next tail
)
636 (string-match "(+\\'" eword
))
637 (setq crest
(concat crest
(match-string 0 eword
))
638 eword
(substring eword
0 (match-beginning 0)))
639 (setq eword
(concat eword next
)))
641 eword
(concat eword next
)))
642 (when (string-match "\n[ \t]+\\'" eword
)
643 ;; Reomove a superfluous empty line.
644 (setq eword
(substring eword
0 (match-beginning 0))))
645 (rfc2047-encode-1 (length crest
) (substring string index
)
646 cs encoder start
" " tail
647 (concat eword
"\n" crest
)))))))
649 (defun rfc2047-encode (b e
)
650 "Encode the word(s) in the region B to E.
651 Point moves to the end of the region."
652 (let ((mime-charset (or (mm-find-mime-charset-region b e
) (list 'us-ascii
)))
653 cs encoding tail crest eword
)
654 ;; Use utf-8 as a last resort if determining charset of text fails.
655 (if (memq nil mime-charset
)
656 (setq mime-charset
(list 'utf-8
)))
657 (cond ((> (length mime-charset
) 1)
658 (error "Can't rfc2047-encode `%s'"
659 (buffer-substring-no-properties b e
)))
660 ((= (length mime-charset
) 1)
661 (setq mime-charset
(car mime-charset
)
662 cs
(mm-charset-to-coding-system mime-charset
))
663 (unless (and (mm-multibyte-p)
664 (mm-coding-system-p cs
))
667 (narrow-to-region b e
)
669 (or (cdr (assq mime-charset
670 rfc2047-charset-encoding-alist
))
671 ;; For the charsets that don't have a preferred
672 ;; encoding, choose the one that's shorter.
673 (if (eq (rfc2047-qp-or-base64) 'base64
)
678 (skip-chars-forward "^ \t\n")
679 ;; `tail' may contain a close parenthesis.
680 (setq tail
(buffer-substring-no-properties e
(point)))
682 (setq b
(point-marker)
683 e
(set-marker (make-marker) e
))
684 (rfc2047-fold-region (point-at-bol) b
)
686 (skip-chars-backward "^ \t\n")
687 (unless (= 0 (skip-chars-backward " \t"))
688 ;; `crest' may contain whitespace and an open parenthesis.
689 (setq crest
(buffer-substring-no-properties (point) b
)))
690 (setq eword
(rfc2047-encode-1
692 (mm-replace-in-string
693 (buffer-substring-no-properties b e
)
694 "\n\\([ \t]?\\)" "\\1")
696 (or (cdr (assq encoding
697 rfc2047-encode-function-alist
))
699 (concat "=?" (downcase (symbol-name mime-charset
))
700 "?" (upcase (symbol-name encoding
)) "?")
703 (delete-region (if (eq (aref eword
0) ?
\n)
705 ;; The line was folded before encoding.
710 ;; `eword' contains `crest' and `tail'.
714 (unless (or (/= 0 (length tail
))
716 (looking-at "[ \t\n)]"))
721 (defun rfc2047-fold-field ()
722 "Fold the current header field."
725 (rfc2047-narrow-to-field)
726 (rfc2047-fold-region (point-min) (point-max)))))
728 (defun rfc2047-fold-region (b e
)
729 "Fold long lines in region B to E."
731 (narrow-to-region b e
)
732 (goto-char (point-min))
736 (bol (save-restriction
740 (when (and (or break qword-break
)
741 (> (- (point) bol
) 76))
742 (goto-char (or break qword-break
))
745 (skip-chars-backward " \t")
746 (if (looking-at "[ \t]")
749 (setq bol
(1- (point)))
750 ;; Don't break before the first non-LWSP characters.
751 (skip-chars-forward " \t")
755 ((eq (char-after) ?
\n)
760 (skip-chars-forward " \t")
761 (unless (or (eobp) (eq (char-after) ?
\n))
763 ((eq (char-after) ?
\r)
765 ((memq (char-after) '(? ?
\t))
766 (skip-chars-forward " \t")
767 (unless first
;; Don't break just after the header name.
768 (setq break
(point))))
770 (if (not (looking-at "=\\?[^=]"))
771 (if (eq (char-after) ?
=)
773 (skip-chars-forward "^ \t\n\r="))
774 ;; Don't break at the start of the field.
775 (unless (= (point) b
)
776 (setq qword-break
(point)))
777 (skip-chars-forward "^ \t\n\r")))
779 (skip-chars-forward "^ \t\n\r")))
781 (when (and (or break qword-break
)
782 (> (- (point) bol
) 76))
783 (goto-char (or break qword-break
))
786 (if (or (> 0 (skip-chars-backward " \t"))
787 (looking-at "[ \t]"))
790 (setq bol
(1- (point)))
791 ;; Don't break before the first non-LWSP characters.
792 (skip-chars-forward " \t")
794 (forward-char 1))))))
796 (defun rfc2047-unfold-field ()
797 "Fold the current line."
800 (rfc2047-narrow-to-field)
801 (rfc2047-unfold-region (point-min) (point-max)))))
803 (defun rfc2047-unfold-region (b e
)
804 "Unfold lines in region B to E."
806 (narrow-to-region b e
)
807 (goto-char (point-min))
808 (let ((bol (save-restriction
811 (eol (point-at-eol)))
814 (if (and (looking-at "[ \t]")
815 (< (- (point-at-eol) bol
) 76))
816 (delete-region eol
(progn
818 (skip-chars-forward "\r\n")
820 (setq bol
(point-at-bol)))
821 (setq eol
(point-at-eol))
824 (defun rfc2047-b-encode-string (string)
825 "Base64-encode the header contained in STRING."
826 (base64-encode-string string t
))
828 (autoload 'quoted-printable-encode-region
"qp")
830 (defun rfc2047-q-encode-string (string)
831 "Quoted-printable-encode the header in STRING."
832 (mm-with-unibyte-buffer
834 (quoted-printable-encode-region
835 (point-min) (point-max) nil
836 ;; = (\075), _ (\137), ? (\077) are used in the encoded word.
837 ;; Avoid using 8bit characters.
838 ;; This list excludes `especials' (see the RFC2047 syntax),
839 ;; meaning that some characters in non-structured fields will
840 ;; get encoded when they con't need to be. The following is
841 ;; what it used to be.
842 ;;; ;; Equivalent to "^\000-\007\011\013\015-\037\200-\377=_?"
843 ;;; "\010\012\014\040-\074\076\100-\136\140-\177")
844 "-\b\n\f !#-'*+0-9A-Z\\^`-~\d")
845 (subst-char-in-region (point-min) (point-max) ? ?_
)
848 (defun rfc2047-encode-parameter (param value
)
849 "Return and PARAM=VALUE string encoded in the RFC2047-like style.
850 This is a substitution for the `rfc2231-encode-string' function, that
851 is the standard but many mailers don't support it."
852 (let ((rfc2047-encoding-type 'mime
)
853 (rfc2047-encode-max-chars nil
))
854 (rfc2045-encode-string param
(rfc2047-encode-string value
))))
857 ;;; Functions for decoding RFC2047 messages
860 (defvar rfc2047-quote-decoded-words-containing-tspecials nil
861 "If non-nil, quote decoded words containing special characters.")
863 (defvar rfc2047-allow-incomplete-encoded-text t
864 "*Non-nil means allow incomplete encoded-text in successive encoded-words.
865 Dividing of encoded-text in the place other than character boundaries
866 violates RFC2047 section 5, while we have a capability to decode it.
867 If it is non-nil, the decoder will decode B- or Q-encoding in each
868 encoded-word, concatenate them, and decode it by charset. Otherwise,
869 the decoder will fully decode each encoded-word before concatenating
872 (defun rfc2047-strip-backslashes-in-quoted-strings ()
873 "Strip backslashes in quoted strings. `\\\"' remains."
874 (goto-char (point-min))
876 (with-syntax-table (standard-syntax-table)
877 (while (search-forward "\"" nil t
)
878 (unless (eq (char-before) ?
\\)
879 (setq beg
(match-end 0))
880 (goto-char (match-beginning 0))
885 (narrow-to-region beg
(1- (point)))
887 (while (search-forward "\\" nil
'move
)
888 (unless (memq (char-after) '(?
\"))
893 (goto-char beg
))))))))
895 (defun rfc2047-charset-to-coding-system (charset &optional allow-override
)
896 "Return coding-system corresponding to MIME CHARSET.
897 If your Emacs implementation can't decode CHARSET, return nil.
899 If allow-override is given, use `mm-charset-override-alist' to
900 map undesired charset names to their replacement. This should
901 only be used for decoding, not for encoding."
902 (when (stringp charset
)
903 (setq charset
(intern (downcase charset
))))
904 (when (or (not charset
)
905 (eq 'gnus-all mail-parse-ignored-charsets
)
906 (memq 'gnus-all mail-parse-ignored-charsets
)
907 (memq charset mail-parse-ignored-charsets
))
908 (setq charset mail-parse-charset
))
909 (let ((cs (mm-charset-to-coding-system charset nil allow-override
)))
910 (cond ((eq cs
'ascii
)
911 (setq cs
(or (mm-charset-to-coding-system mail-parse-charset
)
913 ((mm-coding-system-p cs
))
915 (listp mail-parse-ignored-charsets
)
916 (memq 'gnus-unknown mail-parse-ignored-charsets
))
917 (setq cs
(mm-charset-to-coding-system mail-parse-charset
))))
922 (autoload 'quoted-printable-decode-string
"qp")
924 (defun rfc2047-decode-encoded-words (words)
925 "Decode successive encoded-words in WORDS and return a decoded string.
926 Each element of WORDS looks like (CHARSET ENCODING ENCODED-TEXT
928 (let (word charset cs encoding text rest
)
930 (setq word
(pop words
))
931 (if (and (setq cs
(rfc2047-charset-to-coding-system
932 (setq charset
(car word
)) t
))
934 (cond ((char-equal ?B
(nth 1 word
))
935 (setq text
(base64-decode-string
936 (rfc2047-pad-base64 (nth 2 word
)))))
937 ((char-equal ?Q
(nth 1 word
))
938 (setq text
(quoted-printable-decode-string
939 (mm-subst-char-in-string
940 ?_ ?
(nth 2 word
) t
)))))
942 (message "%s" (error-message-string code
))
944 (if (and rfc2047-allow-incomplete-encoded-text
946 ;; Concatenate text of which the charset is the same.
947 (setcdr (car rest
) (concat (cdar rest
) text
))
948 (push (cons cs text
) rest
))
949 ;; Don't decode encoded-word.
950 (push (cons nil
(nth 3 word
)) rest
)))
953 (or (and (setq cs
(caar rest
))
955 (mm-decode-coding-string (cdar rest
) cs
)
957 (message "%s" (error-message-string code
))
959 (concat (when (cdr rest
) " ")
962 (not (eq (string-to-char words
) ?
)))
968 ;; Fixme: This should decode in place, not cons intermediate strings.
969 ;; Also check whether it needs to worry about delimiting fields like
972 ;; In fact it's reported that (invalid) encoding of mailboxes in
973 ;; addr-specs is in use, so delimiting fields might help. Probably
974 ;; not decoding a word which isn't properly delimited is good enough
975 ;; and worthwhile (is it more correct or not?), e.g. something like
976 ;; `=?iso-8859-1?q?foo?=@'.
978 (defun rfc2047-decode-region (start end
&optional address-mime
)
979 "Decode MIME-encoded words in region between START and END.
980 If ADDRESS-MIME is non-nil, strip backslashes which precede characters
981 other than `\"' and `\\' in quoted strings."
983 (let ((case-fold-search t
)
985 (if rfc2047-allow-irregular-q-encoded-words
987 (concat "[\n\t ]*\\(" rfc2047-encoded-word-regexp-loose
"\\)"))
989 (concat "[\n\t ]*\\(" rfc2047-encoded-word-regexp
"\\)"))))
993 (narrow-to-region start end
)
995 (rfc2047-strip-backslashes-in-quoted-strings))
996 (goto-char (setq b start
))
997 ;; Look for the encoded-words.
998 (while (setq match
(re-search-forward eword-regexp nil t
))
999 (setq e
(match-beginning 1)
1003 (push (list (match-string 2) ;; charset
1004 (char-after (match-beginning 3)) ;; encoding
1005 (substring (match-string 3) 2) ;; encoded-text
1006 (match-string 1)) ;; encoded-word
1008 ;; Look for the subsequent encoded-words.
1009 (when (setq match
(looking-at eword-regexp
))
1010 (goto-char (setq end
(match-end 0)))))
1011 ;; Replace the encoded-words with the decoded one.
1012 (delete-region e end
)
1013 (insert (rfc2047-decode-encoded-words (nreverse words
)))
1015 (narrow-to-region e
(point))
1017 ;; Remove newlines between decoded words, though such
1018 ;; things essentially must not be there.
1019 (while (re-search-forward "[\n\r]+" nil t
)
1020 (replace-match " "))
1021 (setq end
(point-max))
1022 ;; Quote decoded words if there are special characters
1023 ;; which might violate RFC2822.
1024 (when (and rfc2047-quote-decoded-words-containing-tspecials
1025 (let ((regexp (car (rassq
1027 rfc2047-header-encoding-alist
))))
1032 ;; Don't quote words if already quoted.
1033 (not (and (eq (char-before e
) ?
\")
1034 (eq (char-after end
) ?
\")))
1037 (while (and (memq (char-after) '(? ?
\t))
1038 (zerop (forward-line -
1))))
1039 (looking-at regexp
)))))))
1042 (skip-chars-forward " \t")
1043 (setq start
(point))
1044 (setq quoted
(eq (char-after) ?
\"))
1045 (goto-char (point-max))
1046 (skip-chars-backward " \t" start
)
1047 (if (setq quoted
(and quoted
1048 (> (point) (1+ start
))
1049 (eq (char-before) ?
\")))
1052 (setq start
(1+ start
)
1053 end
(point-marker)))
1054 (setq end
(point-marker)))
1056 (while (search-forward "\"" end t
)
1059 (zerop (%
(skip-chars-backward "\\\\") 2))
1060 (goto-char (match-beginning 0)))
1063 (when (and (not quoted
)
1067 (concat "[" ietf-drums-tspecials
"]")
1073 (set-marker end nil
)))
1074 (goto-char (point-max)))
1075 (when (and (mm-multibyte-p)
1077 (not (eq mail-parse-charset
'us-ascii
))
1078 (not (eq mail-parse-charset
'gnus-decoded
)))
1079 (mm-decode-coding-region b e mail-parse-charset
))
1081 (when (and (mm-multibyte-p)
1083 (not (eq mail-parse-charset
'us-ascii
))
1084 (not (eq mail-parse-charset
'gnus-decoded
)))
1085 (mm-decode-coding-region b
(point-max) mail-parse-charset
))))))
1087 (defun rfc2047-decode-address-region (start end
)
1088 "Decode MIME-encoded words in region between START and END.
1089 Backslashes which precede characters other than `\"' and `\\' in quoted
1090 strings are stripped."
1091 (rfc2047-decode-region start end t
))
1093 (defun rfc2047-decode-string (string &optional address-mime
)
1094 "Decode MIME-encoded STRING and return the result.
1095 If ADDRESS-MIME is non-nil, strip backslashes which precede characters
1096 other than `\"' and `\\' in quoted strings."
1097 ;; (let ((m (mm-multibyte-p)))
1098 (if (string-match "=\\?" string
)
1100 ;; We used to only call mm-enable-multibyte if `m' is non-nil,
1101 ;; but this can't be the right criterion. Don't just revert this
1102 ;; change if it encounters a bug. Please help me fix it
1103 ;; right instead. --Stef
1104 ;; The string returned should always be multibyte in a multibyte
1105 ;; session, i.e. the buffer should be multibyte before
1106 ;; `buffer-string' is called.
1107 (mm-enable-multibyte)
1110 (rfc2047-decode-region (point-min) (point-max) address-mime
))
1115 (when (mm-multibyte-string-p string
)
1116 (mm-enable-multibyte))
1118 (rfc2047-strip-backslashes-in-quoted-strings)
1120 ;; Fixme: As above, `m' here is inappropriate.
1123 (not (eq mail-parse-charset
'us-ascii
))
1124 (not (eq mail-parse-charset
'gnus-decoded
)))
1125 ;; `decode-coding-string' in Emacs offers a third optional
1126 ;; arg NOCOPY to avoid consing a new string if the decoding
1127 ;; is "trivial". Unfortunately it currently doesn't
1128 ;; consider anything else than a `nil' coding system
1130 ;; `rfc2047-decode-string' is called multiple times for each
1131 ;; article during summary buffer generation, and we really
1132 ;; want to avoid unnecessary consing. So we bypass
1133 ;; `decode-coding-string' if the string is purely ASCII.
1134 (if (and (fboundp 'detect-coding-string
)
1135 ;; string is purely ASCII
1136 (eq (detect-coding-string string t
) 'undecided
))
1138 (mm-decode-coding-string string mail-parse-charset
))
1139 (mm-string-to-multibyte string
)))) ;; )
1141 (defun rfc2047-decode-address-string (string)
1142 "Decode MIME-encoded STRING and return the result.
1143 Backslashes which precede characters other than `\"' and `\\' in quoted
1144 strings are stripped."
1145 (rfc2047-decode-string string t
))
1147 (defun rfc2047-pad-base64 (string)
1148 "Pad STRING to quartets."
1149 ;; Be more liberal to accept buggy base64 strings. If
1150 ;; base64-decode-string accepts buggy strings, this function could
1151 ;; be aliased to identity.
1152 (if (= 0 (mod (length string
) 4))
1154 (when (string-match "=+$" string
)
1155 (setq string
(substring string
0 (match-beginning 0))))
1156 (case (mod (length string
) 4)
1158 (1 string
) ;; Error, don't pad it.
1159 (2 (concat string
"=="))
1160 (3 (concat string
"=")))))
1164 ;;; rfc2047.el ends here