| blob | e881256f386ef04b3d23c09927881f6f4f25ae2b |
1 ;;; rfc2047.el --- functions for encoding and decoding rfc2047 messages
3 ;; Copyright (C) 1998-2012 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/>.
22 ;;; Commentary:
24 ;; RFC 2047 is "MIME (Multipurpose Internet Mail Extensions) Part
25 ;; Three: Message Header Extensions for Non-ASCII Text".
27 ;;; Code:
35 ;; Fixme: Avoid this (used for mail-parse-charset) mm dependence on gnus.
47 "*Header/encoding method alist.
48 The list is traversed sequentially. The keys can either be
49 header regexps or t.
51 The values can 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
87 "Alist of MIME charsets to RFC2047 encodings.
88 Valid encodings are nil, `Q' and `B'. These indicate binary (no) encoding,
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.
113 )
119 "Regexp that matches encoded word allowing loose Q encoding."
120 ;; The pattern for the Q encoding, i.e. the one beginning with "Q",
121 ;; is similar to:
122 ;; "Q\\?\\(\\?+[^\n=?]\\)?\\([^\n?]+\\?+[^\n=?]\\)*[^\n?]*\\?*"
123 ;; <--------1-------><----------2,3----------><--4--><-5->
124 ;; They mean:
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
130 ;; encoded word.
131 ;; 5. At the end, allow "?"s.
132 ))
134 ;;;
135 ;;; Functions for encoding RFC2047 messages
136 ;;;
139 "Return the type with which to encode the buffer.
140 This is either `base64' or `quoted-printable'."
151 ;; Don't base64, say, a short line with a single
152 ;; non-ASCII char when splitting parts by charset.
154 'quoted-printable
158 "Narrow the buffer to the header on the current line."
170 "Return the value of the field at point."
178 encodable-regexp)
180 Quoting will not be done in a quoted string if it contains characters
181 matching ENCODABLE-REGEXP or it is within parentheses."
185 beg end)
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
233 "Encode the message header according to `rfc2047-header-encoding-alist'.
234 Should be called narrowed to the head of the message."
243 alist rfc2047-header-encoding-alist)
258 ;; 8 bit must be decoded.
263 ;; No encoding necessary, but folding is nice
273 ;; We found something that may perhaps be encoded.
285 mail-parse-charset)
287 mail-parse-charset)))
288 ;; We get this when CC'ing messages 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.
299 ;;; ((null method)
300 ;;; (if (or (message-options-get
301 ;;; 'rfc2047-encode-message-header-encode-any)
302 ;;; (message-options-set
303 ;;; 'rfc2047-encode-message-header-encode-any
304 ;;; (y-or-n-p
305 ;;; "Some texts are not encoded. Encode anyway?")))
306 ;;; (rfc2047-encode-region (point-min) (point-max))
307 ;;; (error "Cannot send unencoded text")))
314 ;; Hm.
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
322 "Return non-nil if any characters in current buffer need encoding in headers.
323 The buffer may be narrowed."
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.
342 ;; (make-char-table 'syntax-table '(2)) only works in Emacs.
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 --
347 ;; copy it from ?a.
363 table))
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."
374 start ; start of current token
375 end begin csyntax
376 ;; Whether there's an encoded word before the current token,
377 ;; either immediately or separated by space.
378 last-encoded
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.
399 ;; `address-mime' case -- take care of quoted words, comments.
404 ;; Look for rfc2822-style: sequences of atoms, quoted
405 ;; strings, specials, whitespace. (Specials mustn't be
406 ;; encoded.)
408 ;; Skip whitespace.
413 ;; else token start
415 ;; Quoted word.
418 ;; Does it need encoding?
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.
433 ;; A space is needed between the encoded words.
437 ;; Adjust the end position for the deleted quotes.
442 ;; Skip other delimiters, but record that they've
443 ;; potentially separated quoted words.
449 ;; Look for the end of parentheses.
451 ;; Encode text as an unstructured field.
456 ;; Find the end.
457 ;; Skip one ASCII word, or encode continuous words
458 ;; in which all those contain non-ASCII characters.
469 ;; Skip backslash-quoted characters.
471 end)
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.
491 encodable-regexp end t))
497 nil)))
502 ;; There are special characters better
503 ;; to be encoded so that MTAs may parse
504 ;; them safely.
507 encodable-regexp
513 ;; Separate encodable text and
514 ;; delimiter.
534 ;; Also encode bogus delimiters.
536 ;; Separate encodable text and delimiter.
554 "Encode words in STRING.
555 By default, the string is treated as containing addresses (see
556 `rfc2047-encoding-type')."
562 ;; From RFC 2047:
563 ;; 2. Syntax of encoded-words
564 ;; [...]
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.
568 ;;
569 ;; In `rfc2047-encode-parameter' it is bound to nil, so don't defconst it.
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
578 "Subroutine used by `rfc2047-encode'."
585 string))
590 ;; Remove a superfluous empty line.
593 ;; Break the line before the open parenthesis.
602 next len)
610 cs)
616 prev nil)
620 rfc2047-encode-max-chars))
624 ;; Break the line after the close parenthesis.
628 prev next
631 prev nil)))))
643 ;; Remove a superfluous empty line.
646 cs encoder start " " tail
650 "Encode the word(s) in the region B to E.
651 Point moves to the end of the region."
653 cs encoding tail crest eword)
654 ;; Use utf-8 as a last resort if determining charset of text fails.
670 rfc2047-charset-encoding-alist))
671 ;; For the charsets that don't have a preferred
672 ;; encoding, choose the one that's shorter.
674 'B
679 ;; `tail' may contain a close parenthesis.
688 ;; `crest' may contain whitespace and an open parenthesis.
695 cs
697 rfc2047-encode-function-alist))
702 tail))
705 ;; The line was folded before encoding.
710 ;; `eword' contains `crest' and `tail'.
722 "Fold the current header field."
729 "Fold long lines in region B to E."
744 qword-break nil)
750 ;; Don't break before the first non-LWSP characters.
758 break nil
759 qword-break nil)
774 ;; Don't break at the start of the field.
785 qword-break nil)
791 ;; Don't break before the first non-LWSP characters.
797 "Fold the current line."
804 "Unfold lines in region B to E."
825 "Base64-encode the header contained in STRING."
831 "Quoted-printable-encode the header in STRING."
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")
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."
856 ;;;
857 ;;; Functions for decoding RFC2047 messages
858 ;;;
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
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."
919 'raw-text
920 cs)))
925 "Decode successive encoded-words in WORDS and return a decoded string.
926 Each element of WORDS looks like (CHARSET ENCODING ENCODED-TEXT
927 ENCODED-WORD)."
943 nil)))
946 ;; Concatenate text of which the charset is the same.
949 ;; Don't decode encoded-word.
958 nil)))
964 words)
966 words))
968 ;; Fixme: This should decode in place, not cons intermediate strings.
969 ;; Also check whether it needs to worry about delimiting fields like
970 ;; encoding.
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?=@'.
979 "Decode MIME-encoded words in region between START and END.
980 If ADDRESS-MIME is non-nil, strip backslashes which precede characters
990 b e match words)
997 ;; Look for the encoded-words.
1001 words nil)
1007 words)
1008 ;; Look for the subsequent encoded-words.
1011 ;; Replace the encoded-words with the decoded one.
1017 ;; Remove newlines between decoded words, though such
1018 ;; things essentially must not be there.
1022 ;; Quote decoded words if there are special characters
1023 ;; which might violate RFC2822.
1026 'address-mime
1027 rfc2047-header-encoding-alist))))
1032 ;; Don't quote words if already quoted.
1068 end t)))
1076 mail-parse-charset
1082 mail-parse-charset
1088 "Decode MIME-encoded words in region between START and END.
1090 strings are stripped."
1094 "Decode MIME-encoded STRING and return the result.
1095 If ADDRESS-MIME is non-nil, strip backslashes which precede characters
1097 ;; (let ((m (mm-multibyte-p)))
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.
1120 ;; Fixme: As above, `m' here is inappropriate.
1122 mail-parse-charset
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
1129 ;; trivial.
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.
1135 ;; string is purely ASCII
1137 string
1142 "Decode MIME-encoded STRING and return the result.
1144 strings are stripped."
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.
1153 string
1164 ;;; rfc2047.el ends here