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, 2012 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
)
37 ;; Fixme: Avoid this (used for mail-parse-charset) mm dependence on gnus.
39 (require 'rfc2045
) ;; rfc2045-encode-string
40 (autoload 'mm-body-7-or-8
"mm-bodies")
42 (defvar rfc2047-header-encoding-alist
43 '(("Newsgroups" . nil
)
46 ("\\(Resent-\\)?\\(From\\|Cc\\|To\\|Bcc\\|\\(In-\\)?Reply-To\\|Sender\
47 \\|Mail-Followup-To\\|Mail-Copies-To\\|Approved\\)" . address-mime
)
49 "*Header/encoding method alist.
50 The list is traversed sequentially. The keys can either be
55 1) nil, in which case no encoding is done;
56 2) `mime', in which case the header will be encoded according to RFC2047;
57 3) `address-mime', like `mime', but takes account of the rules for address
58 fields (where quoted strings and comments must be treated separately);
59 4) a charset, in which case it will be encoded as that charset;
60 5) `default', in which case the field will be encoded as the rest
63 (defvar rfc2047-charset-encoding-alist
89 "Alist of MIME charsets to RFC2047 encodings.
90 Valid encodings are nil, `Q' and `B'. These indicate binary (no) encoding,
91 quoted-printable and base64 respectively.")
93 (defvar rfc2047-encode-function-alist
94 '((Q . rfc2047-q-encode-string
)
95 (B . rfc2047-b-encode-string
)
97 "Alist of RFC2047 encodings to encoding functions.")
99 (defvar rfc2047-encode-encoded-words t
100 "Whether encoded words should be encoded again.")
102 (defvar rfc2047-allow-irregular-q-encoded-words t
103 "*Whether to decode irregular Q-encoded words.")
105 (eval-and-compile ;; Necessary to hard code them in `rfc2047-decode-region'.
106 (defconst rfc2047-encoded-word-regexp
107 "=\\?\\([^][\000-\040()<>@,\;:*\\\"/?.=]+\\)\\(?:\\*[^?]+\\)?\\?\
108 \\(B\\?[+/0-9A-Za-z]*=*\
111 "Regexp that matches encoded word."
112 ;; The patterns for the B encoding and the Q encoding, i.e. the ones
113 ;; beginning with "B" and "Q" respectively, are restricted into only
114 ;; the characters that those encodings may generally use.
116 (defconst rfc2047-encoded-word-regexp-loose
117 "=\\?\\([^][\000-\040()<>@,\;:*\\\"/?.=]+\\)\\(?:\\*[^?]+\\)?\\?\
118 \\(B\\?[+/0-9A-Za-z]*=*\
119 \\|Q\\?\\(?:\\?+[ -<>@-~]\\)?\\(?:[ ->@-~]+\\?+[ -<>@-~]\\)*[ ->@-~]*\\?*\
121 "Regexp that matches encoded word allowing loose Q encoding."
122 ;; The pattern for the Q encoding, i.e. the one beginning with "Q",
124 ;; "Q\\?\\(\\?+[^\n=?]\\)?\\([^\n?]+\\?+[^\n=?]\\)*[^\n?]*\\?*"
125 ;; <--------1-------><----------2,3----------><--4--><-5->
127 ;; 1. After "Q?", allow "?"s that follow a character other than "=".
128 ;; 2. Allow "=" after "Q?"; it isn't regarded as the terminator.
129 ;; 3. In the middle of an encoded word, allow "?"s that follow a
130 ;; character other than "=".
131 ;; 4. Allow any characters other than "?" in the middle of an
133 ;; 5. At the end, allow "?"s.
137 ;;; Functions for encoding RFC2047 messages
140 (defun rfc2047-qp-or-base64 ()
141 "Return the type with which to encode the buffer.
142 This is either `base64' or `quoted-printable'."
144 (let ((limit (min (point-max) (+ 2000 (point-min))))
146 (goto-char (point-min))
147 (skip-chars-forward "\x20-\x7f\r\n\t" limit
)
148 (while (< (point) limit
)
151 (skip-chars-forward "\x20-\x7f\r\n\t" limit
))
152 (if (or (< (* 6 n8bit
) (- limit
(point-min)))
153 ;; Don't base64, say, a short line with a single
154 ;; non-ASCII char when splitting parts by charset.
159 (defun rfc2047-narrow-to-field ()
160 "Narrow the buffer to the header on the current line."
166 (if (re-search-forward "^[^ \n\t]" nil t
)
169 (goto-char (point-min)))
171 (defun rfc2047-field-value ()
172 "Return the value of the field at point."
175 (rfc2047-narrow-to-field)
176 (re-search-forward ":[ \t\n]*" nil t
)
177 (buffer-substring-no-properties (point) (point-max)))))
179 (defun rfc2047-quote-special-characters-in-quoted-strings (&optional
181 "Quote special characters with `\\'s in quoted strings.
182 Quoting will not be done in a quoted string if it contains characters
183 matching ENCODABLE-REGEXP or it is within parentheses."
184 (goto-char (point-min))
185 (let ((tspecials (concat "[" ietf-drums-tspecials
"]"))
188 (with-syntax-table (standard-syntax-table)
192 (eq (char-before) ?\
)))
194 (goto-char (point-max)))
196 (narrow-to-region start
(point))
198 (while (search-forward "\"" nil t
)
199 (setq beg
(match-beginning 0))
200 (unless (eq (char-before beg
) ?
\\)
206 (setq end
(1- (point)))
208 (if (and encodable-regexp
209 (re-search-forward encodable-regexp end t
))
212 (narrow-to-region beg end
)
213 (while (re-search-forward tspecials nil
'move
)
214 (if (eq (char-before) ?
\\)
215 (if (looking-at tspecials
) ;; Already quoted.
218 (goto-char (match-beginning 0))
224 (goto-char (point-max)))
226 (setq start
(point))))))
228 (defvar rfc2047-encoding-type
'address-mime
229 "The type of encoding done by `rfc2047-encode-region'.
230 This should be dynamically bound around calls to
231 `rfc2047-encode-region' to either `mime' or `address-mime'. See
232 `rfc2047-header-encoding-alist', for definitions.")
234 (defun rfc2047-encode-message-header ()
235 "Encode the message header according to `rfc2047-header-encoding-alist'.
236 Should be called narrowed to the head of the message."
239 (goto-char (point-min))
240 (let (alist elem method
)
243 (rfc2047-narrow-to-field)
245 alist rfc2047-header-encoding-alist
)
246 (while (setq elem
(pop alist
))
247 (when (or (and (stringp (car elem
))
248 (looking-at (car elem
)))
252 (if (not (rfc2047-encodable-p))
254 (when (eq method
'address-mime
)
255 (rfc2047-quote-special-characters-in-quoted-strings))
256 (if (and (eq (mm-body-7-or-8) '8bit
)
259 (car message-posting-charset
)))
260 ;; 8 bit must be decoded.
261 (mm-encode-coding-region
262 (point-min) (point-max)
263 (mm-charset-to-coding-system
264 (car message-posting-charset
))))
265 ;; No encoding necessary, but folding is nice
269 (goto-char (point-min))
270 (skip-chars-forward "^:")
271 (when (looking-at ": ")
275 ;; We found something that may perhaps be encoded.
276 (re-search-forward "^[^:]+: *" nil t
)
278 ((eq method
'address-mime
)
279 (rfc2047-encode-region (point) (point-max)))
281 (let ((rfc2047-encoding-type 'mime
))
282 (rfc2047-encode-region (point) (point-max))))
283 ((eq method
'default
)
284 (if (and (featurep 'mule
)
285 (if (boundp 'enable-multibyte-characters
)
286 (default-value 'enable-multibyte-characters
))
288 (mm-encode-coding-region (point) (point-max)
289 mail-parse-charset
)))
290 ;; We get this when CC'ing messsages to newsgroups with
291 ;; 8-bit names. The group name mail copy just got
292 ;; unconditionally encoded. Previously, it would ask
293 ;; whether to encode, which was quite confusing for the
294 ;; user. If the new behavior is wrong, tell me. I have
295 ;; left the old code commented out below.
296 ;; -- Per Abrahamsen <abraham@dina.kvl.dk> Date: 2001-10-07.
297 ;; Modified by Dave Love, with the commented-out code changed
298 ;; in accordance with changes elsewhere.
300 (rfc2047-encode-region (point) (point-max)))
302 ;;; (if (or (message-options-get
303 ;;; 'rfc2047-encode-message-header-encode-any)
304 ;;; (message-options-set
305 ;;; 'rfc2047-encode-message-header-encode-any
307 ;;; "Some texts are not encoded. Encode anyway?")))
308 ;;; (rfc2047-encode-region (point-min) (point-max))
309 ;;; (error "Cannot send unencoded text")))
310 ((mm-coding-system-p method
)
311 (if (or (and (featurep 'mule
)
312 (if (boundp 'enable-multibyte-characters
)
313 (default-value 'enable-multibyte-characters
)))
314 (featurep 'file-coding
))
315 (mm-encode-coding-region (point) (point-max) method
)))
318 (goto-char (point-max)))))))
320 ;; Fixme: This, and the require below may not be the Right Thing, but
321 ;; should be safe just before release. -- fx 2001-02-08
323 (defun rfc2047-encodable-p ()
324 "Return non-nil if any characters in current buffer need encoding in headers.
325 The buffer may be narrowed."
326 (require 'message
) ; for message-posting-charset
328 (mm-find-mime-charset-region (point-min) (point-max))))
329 (goto-char (point-min))
330 (or (and rfc2047-encode-encoded-words
332 (re-search-forward rfc2047-encoded-word-regexp nil t
)
333 (goto-char (point-min))))
335 (not (equal charsets
(list (car message-posting-charset
))))))))
337 ;; Use this syntax table when parsing into regions that may need
338 ;; encoding. Double quotes are string delimiters, backslash is
339 ;; character quoting, and all other RFC 2822 special characters are
340 ;; treated as punctuation so we can use forward-sexp/forward-word to
341 ;; skip to the end of regions appropriately. Nb. ietf-drums does
342 ;; things differently.
343 (defconst rfc2047-syntax-table
344 ;; (make-char-table 'syntax-table '(2)) only works in Emacs.
345 (let ((table (make-syntax-table)))
346 ;; The following is done to work for setting all elements of the table
347 ;; in Emacs 21-23 and XEmacs; it appears to be the cleanest way.
348 ;; Play safe and don't assume the form of the word syntax entry --
350 (if (fboundp 'set-char-table-range
) ; Emacs
351 (funcall (intern "set-char-table-range")
352 table t
(aref (standard-syntax-table) ?a
))
353 (if (fboundp 'put-char-table
)
354 (if (fboundp 'get-char-table
) ; warning avoidance
355 (put-char-table t
(get-char-table ?a
(standard-syntax-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
)
364 (modify-syntax-entry ?\
] "." table
)
365 (modify-syntax-entry ?
: "." table
)
366 (modify-syntax-entry ?\
; "." table)
367 (modify-syntax-entry ?
, "." table
)
368 (modify-syntax-entry ?
@ "." table
)
371 (defun rfc2047-encode-region (b e
)
372 "Encode words in region B to E that need encoding.
373 By default, the region is treated as containing RFC2822 addresses.
374 Dynamically bind `rfc2047-encoding-type' to change that."
376 (narrow-to-region b e
)
377 (let ((encodable-regexp (if rfc2047-encode-encoded-words
378 "[^\000-\177]+\\|=\\?"
380 start
; start of current token
382 ;; Whether there's an encoded word before the current token,
383 ;; either immediately or separated by space.
385 (orig-text (buffer-substring-no-properties b e
)))
386 (if (eq 'mime rfc2047-encoding-type
)
387 ;; Simple case. Continuous words in which all those contain
388 ;; non-ASCII characters are encoded collectively. Encoding
389 ;; ASCII words, including `Re:' used in Subject headers, is
390 ;; avoided for interoperability with non-MIME clients and
391 ;; for making it easy to find keywords.
393 (goto-char (point-min))
394 (while (progn (skip-chars-forward " \t\n")
397 (while (and (looking-at "[ \t\n]*\\([^ \t\n]+\\)")
399 (setq end
(match-end 0))
400 (re-search-forward encodable-regexp end t
)))
402 (if (> (point) start
)
403 (rfc2047-encode start
(point))
405 ;; `address-mime' case -- take care of quoted words, comments.
406 (rfc2047-quote-special-characters-in-quoted-strings encodable-regexp
)
407 (with-syntax-table rfc2047-syntax-table
408 (goto-char (point-min))
409 (condition-case err
; in case of unbalanced quotes
410 ;; Look for rfc2822-style: sequences of atoms, quoted
411 ;; strings, specials, whitespace. (Specials mustn't be
415 (skip-chars-forward " \t\n")
418 ((not (char-after))) ; eob
420 ((eq ?
\" (setq csyntax
(char-syntax (char-after))))
424 ;; Does it need encoding?
426 (if (re-search-forward encodable-regexp end
'move
)
427 ;; It needs encoding. Strip the quotes first,
428 ;; since encoded words can't occur in quotes.
431 (delete-backward-char 1)
435 ;; There was a preceding quoted word. We need
436 ;; to include any separating whitespace in this
437 ;; word to avoid it getting lost.
438 (skip-chars-backward " \t")
439 ;; A space is needed between the encoded words.
443 ;; Adjust the end position for the deleted quotes.
444 (rfc2047-encode start
(- end
2))
445 (setq last-encoded t
)) ; record that it was encoded
446 (setq last-encoded nil
)))
448 ;; Skip other delimiters, but record that they've
449 ;; potentially separated quoted words.
451 (setq last-encoded nil
))
453 (error "Unbalanced parentheses"))
455 ;; Look for the end of parentheses.
457 ;; Encode text as an unstructured field.
458 (let ((rfc2047-encoding-type 'mime
))
459 (rfc2047-encode-region (1+ start
) (1- (point))))
460 (skip-chars-forward ")"))
461 (t ; normal token/whitespace sequence
463 ;; Skip one ASCII word, or encode continuous words
464 ;; in which all those contain non-ASCII characters.
466 (while (not (or end
(eobp)))
467 (when (looking-at "[\000-\177]+")
471 (while (and (or (re-search-forward
472 "[ \t\n]\\|\\Sw" end
'move
)
474 (eq ?
\\ (char-syntax (char-before))))
475 ;; Skip backslash-quoted characters.
478 (setq end
(match-beginning 0))
479 (if rfc2047-encode-encoded-words
482 (when (search-forward "=?" end
'move
)
483 (goto-char (match-beginning 0))
486 ;; Where the value nil of `end' means there may be
487 ;; text to have to be encoded following the point.
488 ;; Otherwise, the point reached to the end of ASCII
489 ;; words separated by whitespace or a special char.
491 (when (looking-at encodable-regexp
)
492 (goto-char (setq begin
(match-end 0)))
493 (while (and (looking-at "[ \t\n]+\\([^ \t\n]+\\)")
494 (setq end
(match-end 0))
496 (while (re-search-forward
497 encodable-regexp end t
))
500 (or (not (re-search-forward "\\Sw" end t
))
502 (goto-char (match-beginning 0))
505 (when (looking-at "[^ \t\n]+")
506 (setq end
(match-end 0))
507 (if (re-search-forward "\\Sw+" end t
)
508 ;; There are special characters better
509 ;; to be encoded so that MTAs may parse
511 (cond ((= end
(point)))
512 ((looking-at (concat "\\sw*\\("
517 (goto-char (1- (match-end 0)))
518 (unless (= (point) (match-beginning 0))
519 ;; Separate encodable text and
523 (skip-chars-forward " \t\n")
524 (if (and (looking-at "[^ \t\n]+")
525 (string-match encodable-regexp
528 (goto-char end
)))))))
529 (skip-chars-backward " \t\n")
532 (if (re-search-forward encodable-regexp end
'move
)
534 (unless (memq (char-before start
) '(nil ?
\t ?
))
537 (skip-chars-backward "^ \t\n")
538 (and (looking-at "\\Sw+")
539 (= (match-end 0) start
)))
540 ;; Also encode bogus delimiters.
542 ;; Separate encodable text and delimiter.
545 (setq start
(1+ start
)
547 (rfc2047-encode start end
)
548 (setq last-encoded t
))
549 (setq last-encoded nil
)))))
551 (if (or debug-on-quit debug-on-error
)
552 (signal (car err
) (cdr err
))
553 (error "Invalid data for rfc2047 encoding: %s"
554 (mm-replace-in-string orig-text
"[ \t\n]+" " "))))))))
555 (rfc2047-fold-region b
(point))
556 (goto-char (point-max))))
558 (defun rfc2047-encode-string (string)
559 "Encode words in STRING.
560 By default, the string is treated as containing addresses (see
561 `rfc2047-encoding-type')."
562 (mm-with-multibyte-buffer
564 (rfc2047-encode-region (point-min) (point-max))
568 ;; 2. Syntax of encoded-words
570 ;; While there is no limit to the length of a multiple-line header
571 ;; field, each line of a header field that contains one or more
572 ;; 'encoded-word's is limited to 76 characters.
574 ;; In `rfc2047-encode-parameter' it is bound to nil, so don't defconst it.
575 (defvar rfc2047-encode-max-chars
76
576 "Maximum characters of each header line that contain encoded-words.
577 According to RFC 2047, it is 76. If it is nil, encoded-words
578 will not be folded. Too small value may cause an error. You
579 should not change this value.")
581 (defun rfc2047-encode-1 (column string cs encoder start crest tail
583 "Subroutine used by `rfc2047-encode'."
584 (cond ((string-equal string
"")
586 ((not rfc2047-encode-max-chars
)
588 (funcall encoder
(if cs
589 (mm-encode-coding-string string cs
)
592 ((>= column rfc2047-encode-max-chars
)
594 (cond ((string-match "\n[ \t]+\\'" eword
)
595 ;; Reomove a superfluous empty line.
596 (setq eword
(substring eword
0 (match-beginning 0))))
597 ((string-match "(+\\'" eword
)
598 ;; Break the line before the open parenthesis.
599 (setq crest
(concat crest
(match-string 0 eword
))
600 eword
(substring eword
0 (match-beginning 0))))))
601 (rfc2047-encode-1 (length crest
) string cs encoder start
" " tail
602 (concat eword
"\n" crest
)))
605 (limit (1- (length string
)))
610 (setq next
(concat start
613 (mm-encode-coding-string
614 (substring string
0 (1+ index
))
616 (substring string
0 (1+ index
))))
618 len
(+ column
(length next
)))
619 (if (> len rfc2047-encode-max-chars
)
622 (if (or (< index limit
)
623 (<= (+ len
(or (string-match "\n" tail
)
625 rfc2047-encode-max-chars
))
628 (if (string-match "\\`)+" tail
)
629 ;; Break the line after the close parenthesis.
630 (setq tail
(concat (substring tail
0 (match-end 0))
632 (substring tail
(match-end 0)))
638 (concat eword next tail
)
641 (string-match "(+\\'" eword
))
642 (setq crest
(concat crest
(match-string 0 eword
))
643 eword
(substring eword
0 (match-beginning 0)))
644 (setq eword
(concat eword next
)))
646 eword
(concat eword next
)))
647 (when (string-match "\n[ \t]+\\'" eword
)
648 ;; Reomove a superfluous empty line.
649 (setq eword
(substring eword
0 (match-beginning 0))))
650 (rfc2047-encode-1 (length crest
) (substring string index
)
651 cs encoder start
" " tail
652 (concat eword
"\n" crest
)))))))
654 (defun rfc2047-encode (b e
)
655 "Encode the word(s) in the region B to E.
656 Point moves to the end of the region."
657 (let ((mime-charset (or (mm-find-mime-charset-region b e
) (list 'us-ascii
)))
658 cs encoding tail crest eword
)
659 (cond ((> (length mime-charset
) 1)
660 (error "Can't rfc2047-encode `%s'"
661 (buffer-substring-no-properties b e
)))
662 ((= (length mime-charset
) 1)
663 (setq mime-charset
(car mime-charset
)
664 cs
(mm-charset-to-coding-system mime-charset
))
665 (unless (and (mm-multibyte-p)
666 (mm-coding-system-p cs
))
669 (narrow-to-region b e
)
671 (or (cdr (assq mime-charset
672 rfc2047-charset-encoding-alist
))
673 ;; For the charsets that don't have a preferred
674 ;; encoding, choose the one that's shorter.
675 (if (eq (rfc2047-qp-or-base64) 'base64
)
680 (skip-chars-forward "^ \t\n")
681 ;; `tail' may contain a close parenthesis.
682 (setq tail
(buffer-substring-no-properties e
(point)))
684 (setq b
(point-marker)
685 e
(set-marker (make-marker) e
))
686 (rfc2047-fold-region (point-at-bol) b
)
688 (skip-chars-backward "^ \t\n")
689 (unless (= 0 (skip-chars-backward " \t"))
690 ;; `crest' may contain whitespace and an open parenthesis.
691 (setq crest
(buffer-substring-no-properties (point) b
)))
692 (setq eword
(rfc2047-encode-1
694 (mm-replace-in-string
695 (buffer-substring-no-properties b e
)
696 "\n\\([ \t]?\\)" "\\1")
698 (or (cdr (assq encoding
699 rfc2047-encode-function-alist
))
701 (concat "=?" (downcase (symbol-name mime-charset
))
702 "?" (upcase (symbol-name encoding
)) "?")
705 (delete-region (if (eq (aref eword
0) ?
\n)
707 ;; The line was folded before encoding.
712 ;; `eword' contains `crest' and `tail'.
716 (unless (or (/= 0 (length tail
))
718 (looking-at "[ \t\n)]"))
723 (defun rfc2047-fold-field ()
724 "Fold the current header field."
727 (rfc2047-narrow-to-field)
728 (rfc2047-fold-region (point-min) (point-max)))))
730 (defun rfc2047-fold-region (b e
)
731 "Fold long lines in region B to E."
733 (narrow-to-region b e
)
734 (goto-char (point-min))
738 (bol (save-restriction
742 (when (and (or break qword-break
)
743 (> (- (point) bol
) 76))
744 (goto-char (or break qword-break
))
747 (skip-chars-backward " \t")
748 (if (looking-at "[ \t]")
751 (setq bol
(1- (point)))
752 ;; Don't break before the first non-LWSP characters.
753 (skip-chars-forward " \t")
757 ((eq (char-after) ?
\n)
762 (skip-chars-forward " \t")
763 (unless (or (eobp) (eq (char-after) ?
\n))
765 ((eq (char-after) ?
\r)
767 ((memq (char-after) '(? ?
\t))
768 (skip-chars-forward " \t")
769 (unless first
;; Don't break just after the header name.
770 (setq break
(point))))
772 (if (not (looking-at "=\\?[^=]"))
773 (if (eq (char-after) ?
=)
775 (skip-chars-forward "^ \t\n\r="))
776 ;; Don't break at the start of the field.
777 (unless (= (point) b
)
778 (setq qword-break
(point)))
779 (skip-chars-forward "^ \t\n\r")))
781 (skip-chars-forward "^ \t\n\r")))
783 (when (and (or break qword-break
)
784 (> (- (point) bol
) 76))
785 (goto-char (or break qword-break
))
788 (if (or (> 0 (skip-chars-backward " \t"))
789 (looking-at "[ \t]"))
792 (setq bol
(1- (point)))
793 ;; Don't break before the first non-LWSP characters.
794 (skip-chars-forward " \t")
796 (forward-char 1))))))
798 (defun rfc2047-unfold-field ()
799 "Fold the current line."
802 (rfc2047-narrow-to-field)
803 (rfc2047-unfold-region (point-min) (point-max)))))
805 (defun rfc2047-unfold-region (b e
)
806 "Unfold lines in region B to E."
808 (narrow-to-region b e
)
809 (goto-char (point-min))
810 (let ((bol (save-restriction
813 (eol (point-at-eol)))
816 (if (and (looking-at "[ \t]")
817 (< (- (point-at-eol) bol
) 76))
818 (delete-region eol
(progn
820 (skip-chars-forward "\r\n")
822 (setq bol
(point-at-bol)))
823 (setq eol
(point-at-eol))
826 (defun rfc2047-b-encode-string (string)
827 "Base64-encode the header contained in STRING."
828 (base64-encode-string string t
))
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 replacement for the `rfc2231-encode-string' function.
852 When attaching files as MIME parts, we should use the RFC2231 encoding
853 to specify the file names containing non-ASCII characters. However,
854 many mail softwares don't support it in practice and recipients won't
855 be able to extract files with correct names. Instead, the RFC2047-like
856 encoding is acceptable generally. This function provides the very
857 RFC2047-like encoding, resigning to such a regrettable trend. To use
858 it, put the following line in your ~/.gnus.el file:
860 \(defalias 'mail-header-encode-parameter 'rfc2047-encode-parameter)
862 (let ((rfc2047-encoding-type 'mime
)
863 (rfc2047-encode-max-chars nil
))
864 (rfc2045-encode-string param
(rfc2047-encode-string value
))))
867 ;;; Functions for decoding RFC2047 messages
870 (defvar rfc2047-quote-decoded-words-containing-tspecials nil
871 "If non-nil, quote decoded words containing special characters.")
873 (defvar rfc2047-allow-incomplete-encoded-text t
874 "*Non-nil means allow incomplete encoded-text in successive encoded-words.
875 Dividing of encoded-text in the place other than character boundaries
876 violates RFC2047 section 5, while we have a capability to decode it.
877 If it is non-nil, the decoder will decode B- or Q-encoding in each
878 encoded-word, concatenate them, and decode it by charset. Otherwise,
879 the decoder will fully decode each encoded-word before concatenating
882 (defun rfc2047-strip-backslashes-in-quoted-strings ()
883 "Strip backslashes in quoted strings. `\\\"' remains."
884 (goto-char (point-min))
886 (with-syntax-table (standard-syntax-table)
887 (while (search-forward "\"" nil t
)
888 (unless (eq (char-before) ?
\\)
889 (setq beg
(match-end 0))
890 (goto-char (match-beginning 0))
895 (narrow-to-region beg
(1- (point)))
897 (while (search-forward "\\" nil
'move
)
898 (unless (memq (char-after) '(?
\"))
899 (delete-backward-char 1))
903 (goto-char beg
))))))))
905 (defun rfc2047-charset-to-coding-system (charset &optional allow-override
)
906 "Return coding-system corresponding to MIME CHARSET.
907 If your Emacs implementation can't decode CHARSET, return nil.
909 If allow-override is given, use `mm-charset-override-alist' to
910 map undesired charset names to their replacement. This should
911 only be used for decoding, not for encoding."
912 (when (stringp charset
)
913 (setq charset
(intern (downcase charset
))))
914 (when (or (not charset
)
915 (eq 'gnus-all mail-parse-ignored-charsets
)
916 (memq 'gnus-all mail-parse-ignored-charsets
)
917 (memq charset mail-parse-ignored-charsets
))
918 (setq charset mail-parse-charset
))
919 (let ((cs (mm-charset-to-coding-system charset nil allow-override
)))
920 (cond ((eq cs
'ascii
)
921 (setq cs
(or (mm-charset-to-coding-system mail-parse-charset
)
923 ((mm-coding-system-p cs
))
925 (listp mail-parse-ignored-charsets
)
926 (memq 'gnus-unknown mail-parse-ignored-charsets
))
927 (setq cs
(mm-charset-to-coding-system mail-parse-charset
))))
932 (defun rfc2047-decode-encoded-words (words)
933 "Decode successive encoded-words in WORDS and return a decoded string.
934 Each element of WORDS looks like (CHARSET ENCODING ENCODED-TEXT
936 (let (word charset cs encoding text rest
)
938 (setq word
(pop words
))
939 (if (and (setq cs
(rfc2047-charset-to-coding-system
940 (setq charset
(car word
)) t
))
942 (cond ((char-equal ?B
(nth 1 word
))
943 (setq text
(base64-decode-string
944 (rfc2047-pad-base64 (nth 2 word
)))))
945 ((char-equal ?Q
(nth 1 word
))
946 (setq text
(quoted-printable-decode-string
947 (mm-subst-char-in-string
948 ?_ ?
(nth 2 word
) t
)))))
950 (message "%s" (error-message-string code
))
952 (if (and rfc2047-allow-incomplete-encoded-text
954 ;; Concatenate text of which the charset is the same.
955 (setcdr (car rest
) (concat (cdar rest
) text
))
956 (push (cons cs text
) rest
))
957 ;; Don't decode encoded-word.
958 (push (cons nil
(nth 3 word
)) rest
)))
961 (or (and (setq cs
(caar rest
))
963 (mm-decode-coding-string (cdar rest
) cs
)
965 (message "%s" (error-message-string code
))
967 (concat (when (cdr rest
) " ")
970 (not (eq (string-to-char words
) ?
)))
976 ;; Fixme: This should decode in place, not cons intermediate strings.
977 ;; Also check whether it needs to worry about delimiting fields like
980 ;; In fact it's reported that (invalid) encoding of mailboxes in
981 ;; addr-specs is in use, so delimiting fields might help. Probably
982 ;; not decoding a word which isn't properly delimited is good enough
983 ;; and worthwhile (is it more correct or not?), e.g. something like
984 ;; `=?iso-8859-1?q?foo?=@'.
986 (defun rfc2047-decode-region (start end
&optional address-mime
)
987 "Decode MIME-encoded words in region between START and END.
988 If ADDRESS-MIME is non-nil, strip backslashes which precede characters
989 other than `\"' and `\\' in quoted strings."
991 (let ((case-fold-search t
)
993 (if rfc2047-allow-irregular-q-encoded-words
995 (concat "[\n\t ]*\\(" rfc2047-encoded-word-regexp-loose
"\\)"))
997 (concat "[\n\t ]*\\(" rfc2047-encoded-word-regexp
"\\)"))))
1001 (narrow-to-region start end
)
1003 (rfc2047-strip-backslashes-in-quoted-strings))
1004 (goto-char (setq b start
))
1005 ;; Look for the encoded-words.
1006 (while (setq match
(re-search-forward eword-regexp nil t
))
1007 (setq e
(match-beginning 1)
1011 (push (list (match-string 2) ;; charset
1012 (char-after (match-beginning 3)) ;; encoding
1013 (substring (match-string 3) 2) ;; encoded-text
1014 (match-string 1)) ;; encoded-word
1016 ;; Look for the subsequent encoded-words.
1017 (when (setq match
(looking-at eword-regexp
))
1018 (goto-char (setq end
(match-end 0)))))
1019 ;; Replace the encoded-words with the decoded one.
1020 (delete-region e end
)
1021 (insert (rfc2047-decode-encoded-words (nreverse words
)))
1023 (narrow-to-region e
(point))
1025 ;; Remove newlines between decoded words, though such
1026 ;; things essentially must not be there.
1027 (while (re-search-forward "[\n\r]+" nil t
)
1028 (replace-match " "))
1029 (setq end
(point-max))
1030 ;; Quote decoded words if there are special characters
1031 ;; which might violate RFC2822.
1032 (when (and rfc2047-quote-decoded-words-containing-tspecials
1033 (let ((regexp (car (rassq
1035 rfc2047-header-encoding-alist
))))
1040 ;; Don't quote words if already quoted.
1041 (not (and (eq (char-before e
) ?
\")
1042 (eq (char-after end
) ?
\")))
1045 (while (and (memq (char-after) '(? ?
\t))
1046 (zerop (forward-line -
1))))
1047 (looking-at regexp
)))))))
1050 (skip-chars-forward " \t")
1051 (setq start
(point))
1052 (setq quoted
(eq (char-after) ?
\"))
1053 (goto-char (point-max))
1054 (skip-chars-backward " \t" start
)
1055 (if (setq quoted
(and quoted
1056 (> (point) (1+ start
))
1057 (eq (char-before) ?
\")))
1060 (setq start
(1+ start
)
1061 end
(point-marker)))
1062 (setq end
(point-marker)))
1064 (while (search-forward "\"" end t
)
1067 (zerop (%
(skip-chars-backward "\\\\") 2))
1068 (goto-char (match-beginning 0)))
1071 (when (and (not quoted
)
1075 (concat "[" ietf-drums-tspecials
"]")
1081 (set-marker end nil
)))
1082 (goto-char (point-max)))
1083 (when (and (mm-multibyte-p)
1085 (not (eq mail-parse-charset
'us-ascii
))
1086 (not (eq mail-parse-charset
'gnus-decoded
)))
1087 (mm-decode-coding-region b e mail-parse-charset
))
1089 (when (and (mm-multibyte-p)
1091 (not (eq mail-parse-charset
'us-ascii
))
1092 (not (eq mail-parse-charset
'gnus-decoded
)))
1093 (mm-decode-coding-region b
(point-max) mail-parse-charset
))))))
1095 (defun rfc2047-decode-address-region (start end
)
1096 "Decode MIME-encoded words in region between START and END.
1097 Backslashes which precede characters other than `\"' and `\\' in quoted
1098 strings are stripped."
1099 (rfc2047-decode-region start end t
))
1101 (defun rfc2047-decode-string (string &optional address-mime
)
1102 "Decode MIME-encoded STRING and return the result.
1103 If ADDRESS-MIME is non-nil, strip backslashes which precede characters
1104 other than `\"' and `\\' in quoted strings."
1105 ;; (let ((m (mm-multibyte-p)))
1106 (if (string-match "=\\?" string
)
1108 ;; We used to only call mm-enable-multibyte if `m' is non-nil,
1109 ;; but this can't be the right criterion. Don't just revert this
1110 ;; change if it encounters a bug. Please help me fix it
1111 ;; right instead. --Stef
1112 ;; The string returned should always be multibyte in a multibyte
1113 ;; session, i.e. the buffer should be multibyte before
1114 ;; `buffer-string' is called.
1115 (mm-enable-multibyte)
1118 (rfc2047-decode-region (point-min) (point-max) address-mime
))
1123 (when (mm-multibyte-string-p string
)
1124 (mm-enable-multibyte))
1126 (rfc2047-strip-backslashes-in-quoted-strings)
1128 ;; Fixme: As above, `m' here is inappropriate.
1131 (not (eq mail-parse-charset
'us-ascii
))
1132 (not (eq mail-parse-charset
'gnus-decoded
)))
1133 ;; `decode-coding-string' in Emacs offers a third optional
1134 ;; arg NOCOPY to avoid consing a new string if the decoding
1135 ;; is "trivial". Unfortunately it currently doesn't
1136 ;; consider anything else than a `nil' coding system
1138 ;; `rfc2047-decode-string' is called multiple times for each
1139 ;; article during summary buffer generation, and we really
1140 ;; want to avoid unnecessary consing. So we bypass
1141 ;; `decode-coding-string' if the string is purely ASCII.
1142 (if (and (fboundp 'detect-coding-string
)
1143 ;; string is purely ASCII
1144 (eq (detect-coding-string string t
) 'undecided
))
1146 (mm-decode-coding-string string mail-parse-charset
))
1147 (mm-string-to-multibyte string
)))) ;; )
1149 (defun rfc2047-decode-address-string (string)
1150 "Decode MIME-encoded STRING and return the result.
1151 Backslashes which precede characters other than `\"' and `\\' in quoted
1152 strings are stripped."
1153 (rfc2047-decode-string string t
))
1155 (defun rfc2047-pad-base64 (string)
1156 "Pad STRING to quartets."
1157 ;; Be more liberal to accept buggy base64 strings. If
1158 ;; base64-decode-string accepts buggy strings, this function could
1159 ;; be aliased to identity.
1160 (if (= 0 (mod (length string
) 4))
1162 (when (string-match "=+$" string
)
1163 (setq string
(substring string
0 (match-beginning 0))))
1164 (case (mod (length string
) 4)
1166 (1 string
) ;; Error, don't pad it.
1167 (2 (concat string
"=="))
1168 (3 (concat string
"=")))))
1172 ;; arch-tag: a07fe3d4-22b5-4c4a-bd89-b1f82d5d36f6
1173 ;;; rfc2047.el ends here