| blob | b2520b26af97c3dc3a619df4ac89f00c418e7030 |
1 ;;; rfc2047.el --- functions for encoding and decoding rfc2047 messages
3 ;; Copyright (C) 1998-2015 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
245 ;; M$ Outlook boycotts decoding of a header if it consists
246 ;; of two or more encoded words and those charsets differ;
247 ;; it seems to decode all words in a header from a charset
248 ;; found first in the header. So, we unify the charsets into
249 ;; a single one used for encoding the whole text in a header.
253 mm-coding-system-priorities)
254 mm-coding-system-priorities)))
269 ;; 8 bit must be decoded.
274 ;; No encoding necessary, but folding is nice
284 ;; We found something that may perhaps be encoded.
296 mail-parse-charset)
298 mail-parse-charset)))
299 ;; We get this when CC'ing messages to newsgroups with
300 ;; 8-bit names. The group name mail copy just got
301 ;; unconditionally encoded. Previously, it would ask
302 ;; whether to encode, which was quite confusing for the
303 ;; user. If the new behavior is wrong, tell me. I have
304 ;; left the old code commented out below.
305 ;; -- Per Abrahamsen <abraham@dina.kvl.dk> Date: 2001-10-07.
306 ;; Modified by Dave Love, with the commented-out code changed
307 ;; in accordance with changes elsewhere.
310 ;;; ((null method)
311 ;;; (if (or (message-options-get
312 ;;; 'rfc2047-encode-message-header-encode-any)
313 ;;; (message-options-set
314 ;;; 'rfc2047-encode-message-header-encode-any
315 ;;; (y-or-n-p
316 ;;; "Some texts are not encoded. Encode anyway?")))
317 ;;; (rfc2047-encode-region (point-min) (point-max))
318 ;;; (error "Cannot send unencoded text")))
325 ;; Hm.
329 ;; Fixme: This, and the require below may not be the Right Thing, but
330 ;; should be safe just before release. -- fx 2001-02-08
333 "Return non-nil if any characters in current buffer need encoding in headers.
334 The buffer may be narrowed."
346 ;; Use this syntax table when parsing into regions that may need
347 ;; encoding. Double quotes are string delimiters, backslash is
348 ;; character quoting, and all other RFC 2822 special characters are
349 ;; treated as punctuation so we can use forward-sexp/forward-word to
350 ;; skip to the end of regions appropriately. Nb. ietf-drums does
351 ;; things differently.
353 ;; (make-char-table 'syntax-table '(2)) only works in Emacs.
355 ;; The following is done to work for setting all elements of the table;
356 ;; it appears to be the cleanest way.
357 ;; Play safe and don't assume the form of the word syntax entry --
358 ;; copy it from ?a.
374 table))
377 "Encode words in region B to E that need encoding.
378 By default, the region is treated as containing RFC2822 addresses.
379 Dynamically bind `rfc2047-encoding-type' to change that."
385 start ; start of current token
386 end begin csyntax
387 ;; Whether there's an encoded word before the current token,
388 ;; either immediately or separated by space.
389 last-encoded
392 ;; Simple case. Continuous words in which all those contain
393 ;; non-ASCII characters are encoded collectively. Encoding
394 ;; ASCII words, including `Re:' used in Subject headers, is
395 ;; avoided for interoperability with non-MIME clients and
396 ;; for making it easy to find keywords.
410 ;; `address-mime' case -- take care of quoted words, comments.
415 ;; Look for rfc2822-style: sequences of atoms, quoted
416 ;; strings, specials, whitespace. (Specials mustn't be
417 ;; encoded.)
419 ;; Skip whitespace.
424 ;; else token start
426 ;; Quoted word.
429 ;; Does it need encoding?
432 ;; It needs encoding. Strip the quotes first,
433 ;; since encoded words can't occur in quotes.
440 ;; There was a preceding quoted word. We need
441 ;; to include any separating whitespace in this
442 ;; word to avoid it getting lost.
444 ;; A space is needed between the encoded words.
448 ;; Adjust the end position for the deleted quotes.
453 ;; Skip other delimiters, but record that they've
454 ;; potentially separated quoted words.
460 ;; Look for the end of parentheses.
462 ;; Encode text as an unstructured field.
467 ;; Find the end.
468 ;; Skip one ASCII word, or encode continuous words
469 ;; in which all those contain non-ASCII characters.
480 ;; Skip backslash-quoted characters.
482 end)
491 ;; Where the value nil of `end' means there may be
492 ;; text to have to be encoded following the point.
493 ;; Otherwise, the point reached to the end of ASCII
494 ;; words separated by whitespace or a special char.
502 encodable-regexp end t))
508 nil)))
513 ;; There are special characters better
514 ;; to be encoded so that MTAs may parse
515 ;; them safely.
518 encodable-regexp
524 ;; Separate encodable text and
525 ;; delimiter.
545 ;; Also encode bogus delimiters.
547 ;; Separate encodable text and delimiter.
565 "Encode words in STRING.
566 By default, the string is treated as containing addresses (see
567 `rfc2047-encoding-type')."
573 ;; From RFC 2047:
574 ;; 2. Syntax of encoded-words
575 ;; [...]
576 ;; While there is no limit to the length of a multiple-line header
577 ;; field, each line of a header field that contains one or more
578 ;; 'encoded-word's is limited to 76 characters.
579 ;;
580 ;; In `rfc2047-encode-parameter' it is bound to nil, so don't defconst it.
582 "Maximum characters of each header line that contain encoded-words.
583 According to RFC 2047, it is 76. If it is nil, encoded-words
584 will not be folded. Too small value may cause an error. You
589 "Subroutine used by `rfc2047-encode'."
596 string))
601 ;; Remove a superfluous empty line.
604 ;; Break the line before the open parenthesis.
613 next len)
621 cs)
627 prev nil)
631 rfc2047-encode-max-chars))
635 ;; Break the line after the close parenthesis.
639 prev next
642 prev nil)))))
654 ;; Remove a superfluous empty line.
657 cs encoder start " " tail
661 "Encode the word(s) in the region B to E.
662 Point moves to the end of the region."
664 cs encoding tail crest eword)
665 ;; Use utf-8 as a last resort if determining charset of text fails.
681 rfc2047-charset-encoding-alist))
682 ;; For the charsets that don't have a preferred
683 ;; encoding, choose the one that's shorter.
685 'B
690 ;; `tail' may contain a close parenthesis.
699 ;; `crest' may contain whitespace and an open parenthesis.
706 cs
708 rfc2047-encode-function-alist))
713 tail))
716 ;; The line was folded before encoding.
721 ;; `eword' contains `crest' and `tail'.
733 "Fold the current header field."
740 "Fold long lines in region B to E."
755 qword-break nil)
761 ;; Don't break before the first non-LWSP characters.
769 break nil
770 qword-break nil)
785 ;; Don't break at the start of the field.
796 qword-break nil)
802 ;; Don't break before the first non-LWSP characters.
808 "Fold the current line."
815 "Unfold lines in region B to E."
836 "Base64-encode the header contained in STRING."
842 "Quoted-printable-encode the header in STRING."
847 ;; = (\075), _ (\137), ? (\077) are used in the encoded word.
848 ;; Avoid using 8bit characters.
849 ;; This list excludes `especials' (see the RFC2047 syntax),
850 ;; meaning that some characters in non-structured fields will
851 ;; get encoded when they con't need to be. The following is
852 ;; what it used to be.
853 ;;; ;; Equivalent to "^\000-\007\011\013\015-\037\200-\377=_?"
854 ;;; "\010\012\014\040-\074\076\100-\136\140-\177")
860 "Return and PARAM=VALUE string encoded in the RFC2047-like style.
861 This is a substitution for the `rfc2231-encode-string' function, that
862 is the standard but many mailers don't support it."
867 ;;;
868 ;;; Functions for decoding RFC2047 messages
869 ;;;
875 "*Non-nil means allow incomplete encoded-text in successive encoded-words.
876 Dividing of encoded-text in the place other than character boundaries
877 violates RFC2047 section 5, while we have a capability to decode it.
878 If it is non-nil, the decoder will decode B- or Q-encoding in each
879 encoded-word, concatenate them, and decode it by charset. Otherwise,
880 the decoder will fully decode each encoded-word before concatenating
907 "Return coding-system corresponding to MIME CHARSET.
908 If your Emacs implementation can't decode CHARSET, return nil.
910 If allow-override is given, use `mm-charset-override-alist' to
911 map undesired charset names to their replacement. This should
912 only be used for decoding, not for encoding."
930 'raw-text
931 cs)))
936 "Decode successive encoded-words in WORDS and return a decoded string.
937 Each element of WORDS looks like (CHARSET ENCODING ENCODED-TEXT
938 ENCODED-WORD)."
954 nil)))
957 ;; Concatenate text of which the charset is the same.
960 ;; Don't decode encoded-word.
969 nil)))
975 words)
977 words))
979 ;; Fixme: This should decode in place, not cons intermediate strings.
980 ;; Also check whether it needs to worry about delimiting fields like
981 ;; encoding.
983 ;; In fact it's reported that (invalid) encoding of mailboxes in
984 ;; addr-specs is in use, so delimiting fields might help. Probably
985 ;; not decoding a word which isn't properly delimited is good enough
986 ;; and worthwhile (is it more correct or not?), e.g. something like
987 ;; `=?iso-8859-1?q?foo?=@'.
990 "Decode MIME-encoded words in region between START and END.
991 If ADDRESS-MIME is non-nil, strip backslashes which precede characters
1001 b e match words)
1008 ;; Look for the encoded-words.
1012 words nil)
1018 words)
1019 ;; Look for the subsequent encoded-words.
1022 ;; Replace the encoded-words with the decoded one.
1028 ;; Remove newlines between decoded words, though such
1029 ;; things essentially must not be there.
1033 ;; Quote decoded words if there are special characters
1034 ;; which might violate RFC2822.
1037 'address-mime
1038 rfc2047-header-encoding-alist))))
1043 ;; Don't quote words if already quoted.
1079 end t)))
1087 mail-parse-charset
1093 mail-parse-charset
1099 "Decode MIME-encoded words in region between START and END.
1101 strings are stripped."
1105 "Decode MIME-encoded STRING and return the result.
1106 If ADDRESS-MIME is non-nil, strip backslashes which precede characters
1108 ;; (let ((m (mm-multibyte-p)))
1111 ;; We used to only call mm-enable-multibyte if `m' is non-nil,
1112 ;; but this can't be the right criterion. Don't just revert this
1113 ;; change if it encounters a bug. Please help me fix it
1114 ;; right instead. --Stef
1115 ;; The string returned should always be multibyte in a multibyte
1116 ;; session, i.e. the buffer should be multibyte before
1117 ;; `buffer-string' is called.
1131 ;; Fixme: As above, `m' here is inappropriate.
1133 mail-parse-charset
1136 ;; `decode-coding-string' in Emacs offers a third optional
1137 ;; arg NOCOPY to avoid consing a new string if the decoding
1138 ;; is "trivial". Unfortunately it currently doesn't
1139 ;; consider anything else than a `nil' coding system
1140 ;; trivial.
1141 ;; `rfc2047-decode-string' is called multiple times for each
1142 ;; article during summary buffer generation, and we really
1143 ;; want to avoid unnecessary consing. So we bypass
1144 ;; `decode-coding-string' if the string is purely ASCII.
1146 ;; string is purely ASCII
1148 string
1153 "Decode MIME-encoded STRING and return the result.
1155 strings are stripped."
1159 "Pad STRING to quartets."
1160 ;; Be more liberal to accept buggy base64 strings. If
1161 ;; base64-decode-string accepts buggy strings, this function could
1162 ;; be aliased to identity.
1164 string
1175 ;;; rfc2047.el ends here