| blob | 840b02a26fdf072b12c00723f86f1c4900eff8c4 |
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 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/>.
23 ;;; Commentary:
25 ;; RFC 2047 is "MIME (Multipurpose Internet Mail Extensions) Part
26 ;; Three: Message Header Extensions for Non-ASCII Text".
28 ;;; Code:
36 ;; Fixme: Avoid this (used for mail-parse-charset) mm dependence on gnus.
48 "*Header/encoding method alist.
49 The list is traversed sequentially. The keys can either be
50 header regexps or t.
52 The values can be:
54 1) nil, in which case no encoding is done;
55 2) `mime', in which case the header will be encoded according to RFC2047;
56 3) `address-mime', like `mime', but takes account of the rules for address
57 fields (where quoted strings and comments must be treated separately);
58 4) a charset, in which case it will be encoded as that charset;
59 5) `default', in which case the field will be encoded as the rest
88 "Alist of MIME charsets to RFC2047 encodings.
89 Valid encodings are nil, `Q' and `B'. These indicate binary (no) encoding,
110 "Regexp that matches encoded word."
111 ;; The patterns for the B encoding and the Q encoding, i.e. the ones
112 ;; beginning with "B" and "Q" respectively, are restricted into only
113 ;; the characters that those encodings may generally use.
114 )
120 "Regexp that matches encoded word allowing loose Q encoding."
121 ;; The pattern for the Q encoding, i.e. the one beginning with "Q",
122 ;; is similar to:
123 ;; "Q\\?\\(\\?+[^\n=?]\\)?\\([^\n?]+\\?+[^\n=?]\\)*[^\n?]*\\?*"
124 ;; <--------1-------><----------2,3----------><--4--><-5->
125 ;; They mean:
126 ;; 1. After "Q?", allow "?"s that follow a character other than "=".
127 ;; 2. Allow "=" after "Q?"; it isn't regarded as the terminator.
128 ;; 3. In the middle of an encoded word, allow "?"s that follow a
129 ;; character other than "=".
130 ;; 4. Allow any characters other than "?" in the middle of an
131 ;; encoded word.
132 ;; 5. At the end, allow "?"s.
133 ))
135 ;;;
136 ;;; Functions for encoding RFC2047 messages
137 ;;;
140 "Return the type with which to encode the buffer.
141 This is either `base64' or `quoted-printable'."
152 ;; Don't base64, say, a short line with a single
153 ;; non-ASCII char when splitting parts by charset.
155 'quoted-printable
159 "Narrow the buffer to the header on the current line."
171 "Return the value of the field at point."
179 encodable-regexp)
181 Quoting will not be done in a quoted string if it contains characters
182 matching ENCODABLE-REGEXP or it is within parentheses."
186 beg end)
228 "The type of encoding done by `rfc2047-encode-region'.
229 This should be dynamically bound around calls to
230 `rfc2047-encode-region' to either `mime' or `address-mime'. See
234 "Encode the message header according to `rfc2047-header-encoding-alist'.
235 Should be called narrowed to the head of the message."
244 alist rfc2047-header-encoding-alist)
259 ;; 8 bit must be decoded.
264 ;; No encoding necessary, but folding is nice
274 ;; We found something that may perhaps be encoded.
286 mail-parse-charset)
288 mail-parse-charset)))
289 ;; We get this when CC'ing messsages to newsgroups with
290 ;; 8-bit names. The group name mail copy just got
291 ;; unconditionally encoded. Previously, it would ask
292 ;; whether to encode, which was quite confusing for the
293 ;; user. If the new behavior is wrong, tell me. I have
294 ;; left the old code commented out below.
295 ;; -- Per Abrahamsen <abraham@dina.kvl.dk> Date: 2001-10-07.
296 ;; Modified by Dave Love, with the commented-out code changed
297 ;; in accordance with changes elsewhere.
300 ;;; ((null method)
301 ;;; (if (or (message-options-get
302 ;;; 'rfc2047-encode-message-header-encode-any)
303 ;;; (message-options-set
304 ;;; 'rfc2047-encode-message-header-encode-any
305 ;;; (y-or-n-p
306 ;;; "Some texts are not encoded. Encode anyway?")))
307 ;;; (rfc2047-encode-region (point-min) (point-max))
308 ;;; (error "Cannot send unencoded text")))
315 ;; Hm.
319 ;; Fixme: This, and the require below may not be the Right Thing, but
320 ;; should be safe just before release. -- fx 2001-02-08
323 "Return non-nil if any characters in current buffer need encoding in headers.
324 The buffer may be narrowed."
336 ;; Use this syntax table when parsing into regions that may need
337 ;; encoding. Double quotes are string delimiters, backslash is
338 ;; character quoting, and all other RFC 2822 special characters are
339 ;; treated as punctuation so we can use forward-sexp/forward-word to
340 ;; skip to the end of regions appropriately. Nb. ietf-drums does
341 ;; things differently.
343 ;; (make-char-table 'syntax-table '(2)) only works in Emacs.
345 ;; The following is done to work for setting all elements of the table
346 ;; in Emacs 21-23 and XEmacs; it appears to be the cleanest way.
347 ;; Play safe and don't assume the form of the word syntax entry --
348 ;; copy it from ?a.
355 table))))
368 table))
371 "Encode words in region B to E that need encoding.
372 By default, the region is treated as containing RFC2822 addresses.
373 Dynamically bind `rfc2047-encoding-type' to change that."
379 start ; start of current token
380 end begin csyntax
381 ;; Whether there's an encoded word before the current token,
382 ;; either immediately or separated by space.
383 last-encoded
386 ;; Simple case. Continuous words in which all those contain
387 ;; non-ASCII characters are encoded collectively. Encoding
388 ;; ASCII words, including `Re:' used in Subject headers, is
389 ;; avoided for interoperability with non-MIME clients and
390 ;; for making it easy to find keywords.
404 ;; `address-mime' case -- take care of quoted words, comments.
409 ;; Look for rfc2822-style: sequences of atoms, quoted
410 ;; strings, specials, whitespace. (Specials mustn't be
411 ;; encoded.)
413 ;; Skip whitespace.
418 ;; else token start
420 ;; Quoted word.
423 ;; Does it need encoding?
426 ;; It needs encoding. Strip the quotes first,
427 ;; since encoded words can't occur in quotes.
434 ;; There was a preceding quoted word. We need
435 ;; to include any separating whitespace in this
436 ;; word to avoid it getting lost.
438 ;; A space is needed between the encoded words.
442 ;; Adjust the end position for the deleted quotes.
447 ;; Skip other delimiters, but record that they've
448 ;; potentially separated quoted words.
454 ;; Look for the end of parentheses.
456 ;; Encode text as an unstructured field.
461 ;; Find the end.
462 ;; Skip one ASCII word, or encode continuous words
463 ;; in which all those contain non-ASCII characters.
474 ;; Skip backslash-quoted characters.
476 end)
485 ;; Where the value nil of `end' means there may be
486 ;; text to have to be encoded following the point.
487 ;; Otherwise, the point reached to the end of ASCII
488 ;; words separated by whitespace or a special char.
496 encodable-regexp end t))
502 nil)))
507 ;; There are special characters better
508 ;; to be encoded so that MTAs may parse
509 ;; them safely.
512 encodable-regexp
518 ;; Separate encodable text and
519 ;; delimiter.
539 ;; Also encode bogus delimiters.
541 ;; Separate encodable text and delimiter.
558 "Encode words in STRING.
559 By default, the string is treated as containing addresses (see
560 `rfc2047-encoding-type')."
566 ;; From RFC 2047:
567 ;; 2. Syntax of encoded-words
568 ;; [...]
569 ;; While there is no limit to the length of a multiple-line header
570 ;; field, each line of a header field that contains one or more
571 ;; 'encoded-word's is limited to 76 characters.
572 ;;
573 ;; In `rfc2047-encode-parameter' it is bound to nil, so don't defconst it.
575 "Maximum characters of each header line that contain encoded-words.
576 According to RFC 2047, it is 76. If it is nil, encoded-words
577 will not be folded. Too small value may cause an error. You
582 "Subroutine used by `rfc2047-encode'."
589 string))
594 ;; Reomove a superfluous empty line.
597 ;; Break the line before the open parenthesis.
606 next len)
614 cs)
620 prev nil)
624 rfc2047-encode-max-chars))
628 ;; Break the line after the close parenthesis.
632 prev next
635 prev nil)))))
647 ;; Reomove a superfluous empty line.
650 cs encoder start " " tail
654 "Encode the word(s) in the region B to E.
655 Point moves to the end of the region."
657 cs encoding tail crest eword)
671 rfc2047-charset-encoding-alist))
672 ;; For the charsets that don't have a preferred
673 ;; encoding, choose the one that's shorter.
675 'B
680 ;; `tail' may contain a close parenthesis.
689 ;; `crest' may contain whitespace and an open parenthesis.
696 cs
698 rfc2047-encode-function-alist))
703 tail))
706 ;; The line was folded before encoding.
711 ;; `eword' contains `crest' and `tail'.
723 "Fold the current header field."
730 "Fold long lines in region B to E."
745 qword-break nil)
751 ;; Don't break before the first non-LWSP characters.
759 break nil
760 qword-break nil)
775 ;; Don't break at the start of the field.
786 qword-break nil)
792 ;; Don't break before the first non-LWSP characters.
798 "Fold the current line."
805 "Unfold lines in region B to E."
826 "Base64-encode the header contained in STRING."
832 "Quoted-printable-encode the header in STRING."
837 ;; = (\075), _ (\137), ? (\077) are used in the encoded word.
838 ;; Avoid using 8bit characters.
839 ;; This list excludes `especials' (see the RFC2047 syntax),
840 ;; meaning that some characters in non-structured fields will
841 ;; get encoded when they con't need to be. The following is
842 ;; what it used to be.
843 ;;; ;; Equivalent to "^\000-\007\011\013\015-\037\200-\377=_?"
844 ;;; "\010\012\014\040-\074\076\100-\136\140-\177")
850 "Return and PARAM=VALUE string encoded in the RFC2047-like style.
851 This is a replacement for the `rfc2231-encode-string' function.
853 When attaching files as MIME parts, we should use the RFC2231 encoding
854 to specify the file names containing non-ASCII characters. However,
855 many mail softwares don't support it in practice and recipients won't
856 be able to extract files with correct names. Instead, the RFC2047-like
857 encoding is acceptable generally. This function provides the very
858 RFC2047-like encoding, resigning to such a regrettable trend. To use
859 it, put the following line in your ~/.gnus.el file:
861 \(defalias 'mail-header-encode-parameter 'rfc2047-encode-parameter)
862 "
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 ;; arch-tag: a07fe3d4-22b5-4c4a-bd89-b1f82d5d36f6
1176 ;;; rfc2047.el ends here