1 ;;; rfc2047.el --- functions for encoding and decoding rfc2047 messages
3 ;; Copyright (C) 1998-2011 Free Software Foundation, Inc.
5 ;; Author: Lars Magne Ingebrigtsen <larsi@gnus.org>
6 ;; MORIOKA Tomohiko <morioka@jaist.ac.jp>
7 ;; This file is part of GNU Emacs.
9 ;; GNU Emacs is free software: you can redistribute it and/or modify
10 ;; it under the terms of the GNU General Public License as published by
11 ;; the Free Software Foundation, either version 3 of the License, or
12 ;; (at your option) any later version.
14 ;; GNU Emacs is distributed in the hope that it will be useful,
15 ;; but WITHOUT ANY WARRANTY; without even the implied warranty of
16 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
17 ;; GNU General Public License for more details.
19 ;; You should have received a copy of the GNU General Public License
20 ;; along with GNU Emacs. If not, see <http://www.gnu.org/licenses/>.
24 ;; RFC 2047 is "MIME (Multipurpose Internet Mail Extensions) Part
25 ;; Three: Message Header Extensions for Non-ASCII Text".
31 (defvar message-posting-charset)
35 ;; Fixme: Avoid this (used for mail-parse-charset) mm dependence on gnus.
37 (require 'rfc2045) ;; rfc2045-encode-string
38 (autoload 'mm-body-7-or-8 "mm-bodies")
40 (defvar rfc2047-header-encoding-alist
41 '(("Newsgroups" . nil)
44 ("\\(Resent-\\)?\\(From\\|Cc\\|To\\|Bcc\\|\\(In-\\)?Reply-To\\|Sender\
45 \\|Mail-Followup-To\\|Mail-Copies-To\\|Approved\\)" . address-mime)
47 "*Header/encoding method alist.
48 The list is traversed sequentially. The keys can either be
53 1) nil, in which case no encoding is done;
54 2) `mime', in which case the header will be encoded according to RFC2047;
55 3) `address-mime', like `mime', but takes account of the rules for address
56 fields (where quoted strings and comments must be treated separately);
57 4) a charset, in which case it will be encoded as that charset;
58 5) `default', in which case the field will be encoded as the rest
61 (defvar rfc2047-charset-encoding-alist
87 "Alist of MIME charsets to RFC2047 encodings.
88 Valid encodings are nil, `Q' and `B'. These indicate binary (no) encoding,
89 quoted-printable and base64 respectively.")
91 (defvar rfc2047-encode-function-alist
92 '((Q . rfc2047-q-encode-string)
93 (B . rfc2047-b-encode-string)
95 "Alist of RFC2047 encodings to encoding functions.")
97 (defvar rfc2047-encode-encoded-words t
98 "Whether encoded words should be encoded again.")
100 (defvar rfc2047-allow-irregular-q-encoded-words t
101 "*Whether to decode irregular Q-encoded words.")
103 (eval-and-compile ;; Necessary to hard code them in `rfc2047-decode-region'.
104 (defconst rfc2047-encoded-word-regexp
105 "=\\?\\([^][\000-\040()<>@,\;:*\\\"/?.=]+\\)\\(?:\\*[^?]+\\)?\\?\
106 \\(B\\?[+/0-9A-Za-z]*=*\
109 "Regexp that matches encoded word."
110 ;; The patterns for the B encoding and the Q encoding, i.e. the ones
111 ;; beginning with "B" and "Q" respectively, are restricted into only
112 ;; the characters that those encodings may generally use.
114 (defconst rfc2047-encoded-word-regexp-loose
115 "=\\?\\([^][\000-\040()<>@,\;:*\\\"/?.=]+\\)\\(?:\\*[^?]+\\)?\\?\
116 \\(B\\?[+/0-9A-Za-z]*=*\
117 \\|Q\\?\\(?:\\?+[ -<>@-~]\\)?\\(?:[ ->@-~]+\\?+[ -<>@-~]\\)*[ ->@-~]*\\?*\
119 "Regexp that matches encoded word allowing loose Q encoding."
120 ;; The pattern for the Q encoding, i.e. the one beginning with "Q",
122 ;; "Q\\?\\(\\?+[^\n=?]\\)?\\([^\n?]+\\?+[^\n=?]\\)*[^\n?]*\\?*"
123 ;; <--------1-------><----------2,3----------><--4--><-5->
125 ;; 1. After "Q?", allow "?"s that follow a character other than "=".
126 ;; 2. Allow "=" after "Q?"; it isn't regarded as the terminator.
127 ;; 3. In the middle of an encoded word, allow "?"s that follow a
128 ;; character other than "=".
129 ;; 4. Allow any characters other than "?" in the middle of an
131 ;; 5. At the end, allow "?"s.
135 ;;; Functions for encoding RFC2047 messages
138 (defun rfc2047-qp-or-base64 ()
139 "Return the type with which to encode the buffer.
140 This is either `base64' or `quoted-printable'."
142 (let ((limit (min (point-max) (+ 2000 (point-min))))
144 (goto-char (point-min))
145 (skip-chars-forward "\x20-\x7f\r\n\t" limit)
146 (while (< (point) limit)
149 (skip-chars-forward "\x20-\x7f\r\n\t" limit))
150 (if (or (< (* 6 n8bit) (- limit (point-min)))
151 ;; Don't base64, say, a short line with a single
152 ;; non-ASCII char when splitting parts by charset.
157 (defun rfc2047-narrow-to-field ()
158 "Narrow the buffer to the header on the current line."
164 (if (re-search-forward "^[^ \n\t]" nil t)
167 (goto-char (point-min)))
169 (defun rfc2047-field-value ()
170 "Return the value of the field at point."
173 (rfc2047-narrow-to-field)
174 (re-search-forward ":[ \t\n]*" nil t)
175 (buffer-substring-no-properties (point) (point-max)))))
177 (defun rfc2047-quote-special-characters-in-quoted-strings (&optional
179 "Quote special characters with `\\'s in quoted strings.
180 Quoting will not be done in a quoted string if it contains characters
181 matching ENCODABLE-REGEXP or it is within parentheses."
182 (goto-char (point-min))
183 (let ((tspecials (concat "[" ietf-drums-tspecials "]"))
186 (with-syntax-table (standard-syntax-table)
190 (eq (char-before) ?\)))
192 (goto-char (point-max)))
194 (narrow-to-region start (point))
196 (while (search-forward "\"" nil t)
197 (setq beg (match-beginning 0))
198 (unless (eq (char-before beg) ?\\)
204 (setq end (1- (point)))
206 (if (and encodable-regexp
207 (re-search-forward encodable-regexp end t))
210 (narrow-to-region beg end)
211 (while (re-search-forward tspecials nil 'move)
212 (if (eq (char-before) ?\\)
213 (if (looking-at tspecials) ;; Already quoted.
216 (goto-char (match-beginning 0))
222 (goto-char (point-max)))
224 (setq start (point))))))
226 (defvar rfc2047-encoding-type 'address-mime
227 "The type of encoding done by `rfc2047-encode-region'.
228 This should be dynamically bound around calls to
229 `rfc2047-encode-region' to either `mime' or `address-mime'. See
230 `rfc2047-header-encoding-alist', for definitions.")
232 (defun rfc2047-encode-message-header ()
233 "Encode the message header according to `rfc2047-header-encoding-alist'.
234 Should be called narrowed to the head of the message."
237 (goto-char (point-min))
238 (let (alist elem method)
241 (rfc2047-narrow-to-field)
243 alist rfc2047-header-encoding-alist)
244 (while (setq elem (pop alist))
245 (when (or (and (stringp (car elem))
246 (looking-at (car elem)))
250 (if (not (rfc2047-encodable-p))
252 (when (eq method 'address-mime)
253 (rfc2047-quote-special-characters-in-quoted-strings))
254 (if (and (eq (mm-body-7-or-8) '8bit)
257 (car message-posting-charset)))
258 ;; 8 bit must be decoded.
259 (mm-encode-coding-region
260 (point-min) (point-max)
261 (mm-charset-to-coding-system
262 (car message-posting-charset))))
263 ;; No encoding necessary, but folding is nice
267 (goto-char (point-min))
268 (skip-chars-forward "^:")
269 (when (looking-at ": ")
273 ;; We found something that may perhaps be encoded.
274 (re-search-forward "^[^:]+: *" nil t)
276 ((eq method 'address-mime)
277 (rfc2047-encode-region (point) (point-max)))
279 (let ((rfc2047-encoding-type 'mime))
280 (rfc2047-encode-region (point) (point-max))))
281 ((eq method 'default)
282 (if (and (featurep 'mule)
283 (if (boundp 'enable-multibyte-characters)
284 (default-value 'enable-multibyte-characters))
286 (mm-encode-coding-region (point) (point-max)
287 mail-parse-charset)))
288 ;; We get this when CC'ing messsages to newsgroups with
289 ;; 8-bit names. The group name mail copy just got
290 ;; unconditionally encoded. Previously, it would ask
291 ;; whether to encode, which was quite confusing for the
292 ;; user. If the new behavior is wrong, tell me. I have
293 ;; left the old code commented out below.
294 ;; -- Per Abrahamsen <abraham@dina.kvl.dk> Date: 2001-10-07.
295 ;; Modified by Dave Love, with the commented-out code changed
296 ;; in accordance with changes elsewhere.
298 (rfc2047-encode-region (point) (point-max)))
300 ;;; (if (or (message-options-get
301 ;;; 'rfc2047-encode-message-header-encode-any)
302 ;;; (message-options-set
303 ;;; 'rfc2047-encode-message-header-encode-any
305 ;;; "Some texts are not encoded. Encode anyway?")))
306 ;;; (rfc2047-encode-region (point-min) (point-max))
307 ;;; (error "Cannot send unencoded text")))
308 ((mm-coding-system-p method)
309 (if (or (and (featurep 'mule)
310 (if (boundp 'enable-multibyte-characters)
311 (default-value 'enable-multibyte-characters)))
312 (featurep 'file-coding))
313 (mm-encode-coding-region (point) (point-max) method)))
316 (goto-char (point-max)))))))
318 ;; Fixme: This, and the require below may not be the Right Thing, but
319 ;; should be safe just before release. -- fx 2001-02-08
321 (defun rfc2047-encodable-p ()
322 "Return non-nil if any characters in current buffer need encoding in headers.
323 The buffer may be narrowed."
324 (require 'message) ; for message-posting-charset
326 (mm-find-mime-charset-region (point-min) (point-max))))
327 (goto-char (point-min))
328 (or (and rfc2047-encode-encoded-words
330 (re-search-forward rfc2047-encoded-word-regexp nil t)
331 (goto-char (point-min))))
333 (not (equal charsets (list (car message-posting-charset))))))))
335 ;; Use this syntax table when parsing into regions that may need
336 ;; encoding. Double quotes are string delimiters, backslash is
337 ;; character quoting, and all other RFC 2822 special characters are
338 ;; treated as punctuation so we can use forward-sexp/forward-word to
339 ;; skip to the end of regions appropriately. Nb. ietf-drums does
340 ;; things differently.
341 (defconst rfc2047-syntax-table
342 ;; (make-char-table 'syntax-table '(2)) only works in Emacs.
343 (let ((table (make-syntax-table)))
344 ;; The following is done to work for setting all elements of the table;
345 ;; it appears to be the cleanest way.
346 ;; Play safe and don't assume the form of the word syntax entry --
348 (if (featurep 'xemacs)
349 (put-char-table t (get-char-table ?a (standard-syntax-table)) table)
350 (set-char-table-range table t (aref (standard-syntax-table) ?a)))
351 (modify-syntax-entry ?\\ "\\" table)
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)
365 (defun rfc2047-encode-region (b e)
366 "Encode words in region B to E that need encoding.
367 By default, the region is treated as containing RFC2822 addresses.
368 Dynamically bind `rfc2047-encoding-type' to change that."
370 (narrow-to-region b e)
371 (let ((encodable-regexp (if rfc2047-encode-encoded-words
372 "[^\000-\177]+\\|=\\?"
374 start ; start of current token
376 ;; Whether there's an encoded word before the current token,
377 ;; either immediately or separated by space.
379 (orig-text (buffer-substring-no-properties b e)))
380 (if (eq 'mime rfc2047-encoding-type)
381 ;; Simple case. Continuous words in which all those contain
382 ;; non-ASCII characters are encoded collectively. Encoding
383 ;; ASCII words, including `Re:' used in Subject headers, is
384 ;; avoided for interoperability with non-MIME clients and
385 ;; for making it easy to find keywords.
387 (goto-char (point-min))
388 (while (progn (skip-chars-forward " \t\n")
391 (while (and (looking-at "[ \t\n]*\\([^ \t\n]+\\)")
393 (setq end (match-end 0))
394 (re-search-forward encodable-regexp end t)))
396 (if (> (point) start)
397 (rfc2047-encode start (point))
399 ;; `address-mime' case -- take care of quoted words, comments.
400 (rfc2047-quote-special-characters-in-quoted-strings encodable-regexp)
401 (with-syntax-table rfc2047-syntax-table
402 (goto-char (point-min))
403 (condition-case err ; in case of unbalanced quotes
404 ;; Look for rfc2822-style: sequences of atoms, quoted
405 ;; strings, specials, whitespace. (Specials mustn't be
409 (skip-chars-forward " \t\n")
412 ((not (char-after))) ; eob
414 ((eq ?\" (setq csyntax (char-syntax (char-after))))
418 ;; Does it need encoding?
420 (if (re-search-forward encodable-regexp end 'move)
421 ;; It needs encoding. Strip the quotes first,
422 ;; since encoded words can't occur in quotes.
429 ;; There was a preceding quoted word. We need
430 ;; to include any separating whitespace in this
431 ;; word to avoid it getting lost.
432 (skip-chars-backward " \t")
433 ;; A space is needed between the encoded words.
437 ;; Adjust the end position for the deleted quotes.
438 (rfc2047-encode start (- end 2))
439 (setq last-encoded t)) ; record that it was encoded
440 (setq last-encoded nil)))
442 ;; Skip other delimiters, but record that they've
443 ;; potentially separated quoted words.
445 (setq last-encoded nil))
447 (error "Unbalanced parentheses"))
449 ;; Look for the end of parentheses.
451 ;; Encode text as an unstructured field.
452 (let ((rfc2047-encoding-type 'mime))
453 (rfc2047-encode-region (1+ start) (1- (point))))
454 (skip-chars-forward ")"))
455 (t ; normal token/whitespace sequence
457 ;; Skip one ASCII word, or encode continuous words
458 ;; in which all those contain non-ASCII characters.
460 (while (not (or end (eobp)))
461 (when (looking-at "[\000-\177]+")
465 (while (and (or (re-search-forward
466 "[ \t\n]\\|\\Sw" end 'move)
468 (eq ?\\ (char-syntax (char-before))))
469 ;; Skip backslash-quoted characters.
472 (setq end (match-beginning 0))
473 (if rfc2047-encode-encoded-words
476 (when (search-forward "=?" end 'move)
477 (goto-char (match-beginning 0))
480 ;; Where the value nil of `end' means there may be
481 ;; text to have to be encoded following the point.
482 ;; Otherwise, the point reached to the end of ASCII
483 ;; words separated by whitespace or a special char.
485 (when (looking-at encodable-regexp)
486 (goto-char (setq begin (match-end 0)))
487 (while (and (looking-at "[ \t\n]+\\([^ \t\n]+\\)")
488 (setq end (match-end 0))
490 (while (re-search-forward
491 encodable-regexp end t))
494 (or (not (re-search-forward "\\Sw" end t))
496 (goto-char (match-beginning 0))
499 (when (looking-at "[^ \t\n]+")
500 (setq end (match-end 0))
501 (if (re-search-forward "\\Sw+" end t)
502 ;; There are special characters better
503 ;; to be encoded so that MTAs may parse
505 (cond ((= end (point)))
506 ((looking-at (concat "\\sw*\\("
511 (goto-char (1- (match-end 0)))
512 (unless (= (point) (match-beginning 0))
513 ;; Separate encodable text and
517 (skip-chars-forward " \t\n")
518 (if (and (looking-at "[^ \t\n]+")
519 (string-match encodable-regexp
522 (goto-char end)))))))
523 (skip-chars-backward " \t\n")
526 (if (re-search-forward encodable-regexp end 'move)
528 (unless (memq (char-before start) '(nil ?\t ? ))
531 (skip-chars-backward "^ \t\n")
532 (and (looking-at "\\Sw+")
533 (= (match-end 0) start)))
534 ;; Also encode bogus delimiters.
536 ;; Separate encodable text and delimiter.
539 (setq start (1+ start)
541 (rfc2047-encode start end)
542 (setq last-encoded t))
543 (setq last-encoded nil)))))
545 (if (or debug-on-quit debug-on-error)
546 (signal (car err) (cdr err))
547 (error "Invalid data for rfc2047 encoding: %s"
548 (mm-replace-in-string orig-text "[ \t\n]+" " "))))))))
549 (rfc2047-fold-region b (point))
550 (goto-char (point-max))))
552 (defun rfc2047-encode-string (string)
553 "Encode words in STRING.
554 By default, the string is treated as containing addresses (see
555 `rfc2047-encoding-type')."
556 (mm-with-multibyte-buffer
558 (rfc2047-encode-region (point-min) (point-max))
562 ;; 2. Syntax of encoded-words
564 ;; While there is no limit to the length of a multiple-line header
565 ;; field, each line of a header field that contains one or more
566 ;; 'encoded-word's is limited to 76 characters.
568 ;; In `rfc2047-encode-parameter' it is bound to nil, so don't defconst it.
569 (defvar rfc2047-encode-max-chars 76
570 "Maximum characters of each header line that contain encoded-words.
571 According to RFC 2047, it is 76. If it is nil, encoded-words
572 will not be folded. Too small value may cause an error. You
573 should not change this value.")
575 (defun rfc2047-encode-1 (column string cs encoder start crest tail
577 "Subroutine used by `rfc2047-encode'."
578 (cond ((string-equal string "")
580 ((not rfc2047-encode-max-chars)
582 (funcall encoder (if cs
583 (mm-encode-coding-string string cs)
586 ((>= column rfc2047-encode-max-chars)
588 (cond ((string-match "\n[ \t]+\\'" eword)
589 ;; Reomove a superfluous empty line.
590 (setq eword (substring eword 0 (match-beginning 0))))
591 ((string-match "(+\\'" eword)
592 ;; Break the line before the open parenthesis.
593 (setq crest (concat crest (match-string 0 eword))
594 eword (substring eword 0 (match-beginning 0))))))
595 (rfc2047-encode-1 (length crest) string cs encoder start " " tail
596 (concat eword "\n" crest)))
599 (limit (1- (length string)))
604 (setq next (concat start
607 (mm-encode-coding-string
608 (substring string 0 (1+ index))
610 (substring string 0 (1+ index))))
612 len (+ column (length next)))
613 (if (> len rfc2047-encode-max-chars)
616 (if (or (< index limit)
617 (<= (+ len (or (string-match "\n" tail)
619 rfc2047-encode-max-chars))
622 (if (string-match "\\`)+" tail)
623 ;; Break the line after the close parenthesis.
624 (setq tail (concat (substring tail 0 (match-end 0))
626 (substring tail (match-end 0)))
632 (concat eword next tail)
635 (string-match "(+\\'" eword))
636 (setq crest (concat crest (match-string 0 eword))
637 eword (substring eword 0 (match-beginning 0)))
638 (setq eword (concat eword next)))
640 eword (concat eword next)))
641 (when (string-match "\n[ \t]+\\'" eword)
642 ;; Reomove a superfluous empty line.
643 (setq eword (substring eword 0 (match-beginning 0))))
644 (rfc2047-encode-1 (length crest) (substring string index)
645 cs encoder start " " tail
646 (concat eword "\n" crest)))))))
648 (defun rfc2047-encode (b e)
649 "Encode the word(s) in the region B to E.
650 Point moves to the end of the region."
651 (let ((mime-charset (or (mm-find-mime-charset-region b e) (list 'us-ascii)))
652 cs encoding tail crest eword)
653 ;; Use utf-8 as a last resort if determining charset of text fails.
654 (if (memq nil mime-charset)
655 (setq mime-charset (list 'utf-8)))
656 (cond ((> (length mime-charset) 1)
657 (error "Can't rfc2047-encode `%s'"
658 (buffer-substring-no-properties b e)))
659 ((= (length mime-charset) 1)
660 (setq mime-charset (car mime-charset)
661 cs (mm-charset-to-coding-system mime-charset))
662 (unless (and (mm-multibyte-p)
663 (mm-coding-system-p cs))
666 (narrow-to-region b e)
668 (or (cdr (assq mime-charset
669 rfc2047-charset-encoding-alist))
670 ;; For the charsets that don't have a preferred
671 ;; encoding, choose the one that's shorter.
672 (if (eq (rfc2047-qp-or-base64) 'base64)
677 (skip-chars-forward "^ \t\n")
678 ;; `tail' may contain a close parenthesis.
679 (setq tail (buffer-substring-no-properties e (point)))
681 (setq b (point-marker)
682 e (set-marker (make-marker) e))
683 (rfc2047-fold-region (point-at-bol) b)
685 (skip-chars-backward "^ \t\n")
686 (unless (= 0 (skip-chars-backward " \t"))
687 ;; `crest' may contain whitespace and an open parenthesis.
688 (setq crest (buffer-substring-no-properties (point) b)))
689 (setq eword (rfc2047-encode-1
691 (mm-replace-in-string
692 (buffer-substring-no-properties b e)
693 "\n\\([ \t]?\\)" "\\1")
695 (or (cdr (assq encoding
696 rfc2047-encode-function-alist))
698 (concat "=?" (downcase (symbol-name mime-charset))
699 "?" (upcase (symbol-name encoding)) "?")
702 (delete-region (if (eq (aref eword 0) ?\n)
704 ;; The line was folded before encoding.
709 ;; `eword' contains `crest' and `tail'.
713 (unless (or (/= 0 (length tail))
715 (looking-at "[ \t\n)]"))
720 (defun rfc2047-fold-field ()
721 "Fold the current header field."
724 (rfc2047-narrow-to-field)
725 (rfc2047-fold-region (point-min) (point-max)))))
727 (defun rfc2047-fold-region (b e)
728 "Fold long lines in region B to E."
730 (narrow-to-region b e)
731 (goto-char (point-min))
735 (bol (save-restriction
739 (when (and (or break qword-break)
740 (> (- (point) bol) 76))
741 (goto-char (or break qword-break))
744 (skip-chars-backward " \t")
745 (if (looking-at "[ \t]")
748 (setq bol (1- (point)))
749 ;; Don't break before the first non-LWSP characters.
750 (skip-chars-forward " \t")
754 ((eq (char-after) ?\n)
759 (skip-chars-forward " \t")
760 (unless (or (eobp) (eq (char-after) ?\n))
762 ((eq (char-after) ?\r)
764 ((memq (char-after) '(? ?\t))
765 (skip-chars-forward " \t")
766 (unless first ;; Don't break just after the header name.
767 (setq break (point))))
769 (if (not (looking-at "=\\?[^=]"))
770 (if (eq (char-after) ?=)
772 (skip-chars-forward "^ \t\n\r="))
773 ;; Don't break at the start of the field.
774 (unless (= (point) b)
775 (setq qword-break (point)))
776 (skip-chars-forward "^ \t\n\r")))
778 (skip-chars-forward "^ \t\n\r")))
780 (when (and (or break qword-break)
781 (> (- (point) bol) 76))
782 (goto-char (or break qword-break))
785 (if (or (> 0 (skip-chars-backward " \t"))
786 (looking-at "[ \t]"))
789 (setq bol (1- (point)))
790 ;; Don't break before the first non-LWSP characters.
791 (skip-chars-forward " \t")
793 (forward-char 1))))))
795 (defun rfc2047-unfold-field ()
796 "Fold the current line."
799 (rfc2047-narrow-to-field)
800 (rfc2047-unfold-region (point-min) (point-max)))))
802 (defun rfc2047-unfold-region (b e)
803 "Unfold lines in region B to E."
805 (narrow-to-region b e)
806 (goto-char (point-min))
807 (let ((bol (save-restriction
810 (eol (point-at-eol)))
813 (if (and (looking-at "[ \t]")
814 (< (- (point-at-eol) bol) 76))
815 (delete-region eol (progn
817 (skip-chars-forward "\r\n")
819 (setq bol (point-at-bol)))
820 (setq eol (point-at-eol))
823 (defun rfc2047-b-encode-string (string)
824 "Base64-encode the header contained in STRING."
825 (base64-encode-string string t))
827 (autoload 'quoted-printable-encode-region "qp")
829 (defun rfc2047-q-encode-string (string)
830 "Quoted-printable-encode the header in STRING."
831 (mm-with-unibyte-buffer
833 (quoted-printable-encode-region
834 (point-min) (point-max) nil
835 ;; = (\075), _ (\137), ? (\077) are used in the encoded word.
836 ;; Avoid using 8bit characters.
837 ;; This list excludes `especials' (see the RFC2047 syntax),
838 ;; meaning that some characters in non-structured fields will
839 ;; get encoded when they con't need to be. The following is
840 ;; what it used to be.
841 ;;; ;; Equivalent to "^\000-\007\011\013\015-\037\200-\377=_?"
842 ;;; "\010\012\014\040-\074\076\100-\136\140-\177")
843 "-\b\n\f !#-'*+0-9A-Z\\^`-~\d")
844 (subst-char-in-region (point-min) (point-max) ? ?_)
847 (defun rfc2047-encode-parameter (param value)
848 "Return and PARAM=VALUE string encoded in the RFC2047-like style.
849 This is a substitution for the `rfc2231-encode-string' function, that
850 is the standard but many mailers don't support it."
851 (let ((rfc2047-encoding-type 'mime)
852 (rfc2047-encode-max-chars nil))
853 (rfc2045-encode-string param (rfc2047-encode-string value))))
856 ;;; Functions for decoding RFC2047 messages
859 (defvar rfc2047-quote-decoded-words-containing-tspecials nil
860 "If non-nil, quote decoded words containing special characters.")
862 (defvar rfc2047-allow-incomplete-encoded-text t
863 "*Non-nil means allow incomplete encoded-text in successive encoded-words.
864 Dividing of encoded-text in the place other than character boundaries
865 violates RFC2047 section 5, while we have a capability to decode it.
866 If it is non-nil, the decoder will decode B- or Q-encoding in each
867 encoded-word, concatenate them, and decode it by charset. Otherwise,
868 the decoder will fully decode each encoded-word before concatenating
871 (defun rfc2047-strip-backslashes-in-quoted-strings ()
872 "Strip backslashes in quoted strings. `\\\"' remains."
873 (goto-char (point-min))
875 (with-syntax-table (standard-syntax-table)
876 (while (search-forward "\"" nil t)
877 (unless (eq (char-before) ?\\)
878 (setq beg (match-end 0))
879 (goto-char (match-beginning 0))
884 (narrow-to-region beg (1- (point)))
886 (while (search-forward "\\" nil 'move)
887 (unless (memq (char-after) '(?\"))
892 (goto-char beg))))))))
894 (defun rfc2047-charset-to-coding-system (charset &optional allow-override)
895 "Return coding-system corresponding to MIME CHARSET.
896 If your Emacs implementation can't decode CHARSET, return nil.
898 If allow-override is given, use `mm-charset-override-alist' to
899 map undesired charset names to their replacement. This should
900 only be used for decoding, not for encoding."
901 (when (stringp charset)
902 (setq charset (intern (downcase charset))))
903 (when (or (not charset)
904 (eq 'gnus-all mail-parse-ignored-charsets)
905 (memq 'gnus-all mail-parse-ignored-charsets)
906 (memq charset mail-parse-ignored-charsets))
907 (setq charset mail-parse-charset))
908 (let ((cs (mm-charset-to-coding-system charset nil allow-override)))
909 (cond ((eq cs 'ascii)
910 (setq cs (or (mm-charset-to-coding-system mail-parse-charset)
912 ((mm-coding-system-p cs))
914 (listp mail-parse-ignored-charsets)
915 (memq 'gnus-unknown mail-parse-ignored-charsets))
916 (setq cs (mm-charset-to-coding-system mail-parse-charset))))
921 (autoload 'quoted-printable-decode-string "qp")
923 (defun rfc2047-decode-encoded-words (words)
924 "Decode successive encoded-words in WORDS and return a decoded string.
925 Each element of WORDS looks like (CHARSET ENCODING ENCODED-TEXT
927 (let (word charset cs encoding text rest)
929 (setq word (pop words))
930 (if (and (setq cs (rfc2047-charset-to-coding-system
931 (setq charset (car word)) t))
933 (cond ((char-equal ?B (nth 1 word))
934 (setq text (base64-decode-string
935 (rfc2047-pad-base64 (nth 2 word)))))
936 ((char-equal ?Q (nth 1 word))
937 (setq text (quoted-printable-decode-string
938 (mm-subst-char-in-string
939 ?_ ? (nth 2 word) t)))))
941 (message "%s" (error-message-string code))
943 (if (and rfc2047-allow-incomplete-encoded-text
945 ;; Concatenate text of which the charset is the same.
946 (setcdr (car rest) (concat (cdar rest) text))
947 (push (cons cs text) rest))
948 ;; Don't decode encoded-word.
949 (push (cons nil (nth 3 word)) rest)))
952 (or (and (setq cs (caar rest))
954 (mm-decode-coding-string (cdar rest) cs)
956 (message "%s" (error-message-string code))
958 (concat (when (cdr rest) " ")
961 (not (eq (string-to-char words) ? )))
967 ;; Fixme: This should decode in place, not cons intermediate strings.
968 ;; Also check whether it needs to worry about delimiting fields like
971 ;; In fact it's reported that (invalid) encoding of mailboxes in
972 ;; addr-specs is in use, so delimiting fields might help. Probably
973 ;; not decoding a word which isn't properly delimited is good enough
974 ;; and worthwhile (is it more correct or not?), e.g. something like
975 ;; `=?iso-8859-1?q?foo?=@'.
977 (defun rfc2047-decode-region (start end &optional address-mime)
978 "Decode MIME-encoded words in region between START and END.
979 If ADDRESS-MIME is non-nil, strip backslashes which precede characters
980 other than `\"' and `\\' in quoted strings."
982 (let ((case-fold-search t)
984 (if rfc2047-allow-irregular-q-encoded-words
986 (concat "[\n\t ]*\\(" rfc2047-encoded-word-regexp-loose "\\)"))
988 (concat "[\n\t ]*\\(" rfc2047-encoded-word-regexp "\\)"))))
992 (narrow-to-region start end)
994 (rfc2047-strip-backslashes-in-quoted-strings))
995 (goto-char (setq b start))
996 ;; Look for the encoded-words.
997 (while (setq match (re-search-forward eword-regexp nil t))
998 (setq e (match-beginning 1)
1002 (push (list (match-string 2) ;; charset
1003 (char-after (match-beginning 3)) ;; encoding
1004 (substring (match-string 3) 2) ;; encoded-text
1005 (match-string 1)) ;; encoded-word
1007 ;; Look for the subsequent encoded-words.
1008 (when (setq match (looking-at eword-regexp))
1009 (goto-char (setq end (match-end 0)))))
1010 ;; Replace the encoded-words with the decoded one.
1011 (delete-region e end)
1012 (insert (rfc2047-decode-encoded-words (nreverse words)))
1014 (narrow-to-region e (point))
1016 ;; Remove newlines between decoded words, though such
1017 ;; things essentially must not be there.
1018 (while (re-search-forward "[\n\r]+" nil t)
1019 (replace-match " "))
1020 (setq end (point-max))
1021 ;; Quote decoded words if there are special characters
1022 ;; which might violate RFC2822.
1023 (when (and rfc2047-quote-decoded-words-containing-tspecials
1024 (let ((regexp (car (rassq
1026 rfc2047-header-encoding-alist))))
1031 ;; Don't quote words if already quoted.
1032 (not (and (eq (char-before e) ?\")
1033 (eq (char-after end) ?\")))
1036 (while (and (memq (char-after) '(? ?\t))
1037 (zerop (forward-line -1))))
1038 (looking-at regexp)))))))
1041 (skip-chars-forward " \t")
1042 (setq start (point))
1043 (setq quoted (eq (char-after) ?\"))
1044 (goto-char (point-max))
1045 (skip-chars-backward " \t" start)
1046 (if (setq quoted (and quoted
1047 (> (point) (1+ start))
1048 (eq (char-before) ?\")))
1051 (setq start (1+ start)
1052 end (point-marker)))
1053 (setq end (point-marker)))
1055 (while (search-forward "\"" end t)
1058 (zerop (% (skip-chars-backward "\\\\") 2))
1059 (goto-char (match-beginning 0)))
1062 (when (and (not quoted)
1066 (concat "[" ietf-drums-tspecials "]")
1072 (set-marker end nil)))
1073 (goto-char (point-max)))
1074 (when (and (mm-multibyte-p)
1076 (not (eq mail-parse-charset 'us-ascii))
1077 (not (eq mail-parse-charset 'gnus-decoded)))
1078 (mm-decode-coding-region b e mail-parse-charset))
1080 (when (and (mm-multibyte-p)
1082 (not (eq mail-parse-charset 'us-ascii))
1083 (not (eq mail-parse-charset 'gnus-decoded)))
1084 (mm-decode-coding-region b (point-max) mail-parse-charset))))))
1086 (defun rfc2047-decode-address-region (start end)
1087 "Decode MIME-encoded words in region between START and END.
1088 Backslashes which precede characters other than `\"' and `\\' in quoted
1089 strings are stripped."
1090 (rfc2047-decode-region start end t))
1092 (defun rfc2047-decode-string (string &optional address-mime)
1093 "Decode MIME-encoded STRING and return the result.
1094 If ADDRESS-MIME is non-nil, strip backslashes which precede characters
1095 other than `\"' and `\\' in quoted strings."
1096 ;; (let ((m (mm-multibyte-p)))
1097 (if (string-match "=\\?" string)
1099 ;; We used to only call mm-enable-multibyte if `m' is non-nil,
1100 ;; but this can't be the right criterion. Don't just revert this
1101 ;; change if it encounters a bug. Please help me fix it
1102 ;; right instead. --Stef
1103 ;; The string returned should always be multibyte in a multibyte
1104 ;; session, i.e. the buffer should be multibyte before
1105 ;; `buffer-string' is called.
1106 (mm-enable-multibyte)
1109 (rfc2047-decode-region (point-min) (point-max) address-mime))
1114 (when (mm-multibyte-string-p string)
1115 (mm-enable-multibyte))
1117 (rfc2047-strip-backslashes-in-quoted-strings)
1119 ;; Fixme: As above, `m' here is inappropriate.
1122 (not (eq mail-parse-charset 'us-ascii))
1123 (not (eq mail-parse-charset 'gnus-decoded)))
1124 ;; `decode-coding-string' in Emacs offers a third optional
1125 ;; arg NOCOPY to avoid consing a new string if the decoding
1126 ;; is "trivial". Unfortunately it currently doesn't
1127 ;; consider anything else than a `nil' coding system
1129 ;; `rfc2047-decode-string' is called multiple times for each
1130 ;; article during summary buffer generation, and we really
1131 ;; want to avoid unnecessary consing. So we bypass
1132 ;; `decode-coding-string' if the string is purely ASCII.
1133 (if (and (fboundp 'detect-coding-string)
1134 ;; string is purely ASCII
1135 (eq (detect-coding-string string t) 'undecided))
1137 (mm-decode-coding-string string mail-parse-charset))
1138 (mm-string-to-multibyte string)))) ;; )
1140 (defun rfc2047-decode-address-string (string)
1141 "Decode MIME-encoded STRING and return the result.
1142 Backslashes which precede characters other than `\"' and `\\' in quoted
1143 strings are stripped."
1144 (rfc2047-decode-string string t))
1146 (defun rfc2047-pad-base64 (string)
1147 "Pad STRING to quartets."
1148 ;; Be more liberal to accept buggy base64 strings. If
1149 ;; base64-decode-string accepts buggy strings, this function could
1150 ;; be aliased to identity.
1151 (if (= 0 (mod (length string) 4))
1153 (when (string-match "=+$" string)
1154 (setq string (substring string 0 (match-beginning 0))))
1155 (case (mod (length string) 4)
1157 (1 string) ;; Error, don't pad it.
1158 (2 (concat string "=="))
1159 (3 (concat string "=")))))
1163 ;;; rfc2047.el ends here