| blob | 0c93331de8bae2ff22f93b06a6c74bb1175d41c5 |
1 ;;; rfc2047.el --- functions for encoding and decoding rfc2047 messages -*- lexical-binding:t -*-
3 ;; Copyright (C) 1998-2017 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 <https://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:
34 ;; Fixme: Avoid this (used for mail-parse-charset) mm dependence on gnus.
40 "RFC2047 messages."
51 "Header/encoding method alist.
52 The list is traversed sequentially. The keys can either be
53 header regexps or t.
55 The values can be:
57 1) nil, in which case no encoding is done;
58 2) `mime', in which case the header will be encoded according to RFC2047;
59 3) `address-mime', like `mime', but takes account of the rules for address
60 fields (where quoted strings and comments must be treated separately);
61 4) a charset, in which case it will be encoded as that charset;
62 5) `default', in which case the field will be encoded as the rest
63 of the article."
67 coding-system
96 "Alist of MIME charsets to RFC2047 encodings.
97 Valid encodings are nil, `Q' and `B'. These indicate binary (no) encoding,
110 "Whether to decode irregular Q-encoded words."
119 "Regexp that matches encoded word."
120 ;; The patterns for the B encoding and the Q encoding, i.e. the ones
121 ;; beginning with "B" and "Q" respectively, are restricted into only
122 ;; the characters that those encodings may generally use.
123 )
129 "Regexp that matches encoded word allowing loose Q encoding."
130 ;; The pattern for the Q encoding, i.e. the one beginning with "Q",
131 ;; is similar to:
132 ;; "Q\\?\\(\\?+[^\n=?]\\)?\\([^\n?]+\\?+[^\n=?]\\)*[^\n?]*\\?*"
133 ;; <--------1-------><----------2,3----------><--4--><-5->
134 ;; They mean:
135 ;; 1. After "Q?", allow "?"s that follow a character other than "=".
136 ;; 2. Allow "=" after "Q?"; it isn't regarded as the terminator.
137 ;; 3. In the middle of an encoded word, allow "?"s that follow a
138 ;; character other than "=".
139 ;; 4. Allow any characters other than "?" in the middle of an
140 ;; encoded word.
141 ;; 5. At the end, allow "?"s.
142 ))
144 ;;;
145 ;;; Functions for encoding RFC2047 messages
146 ;;;
149 "Return the type with which to encode the buffer.
150 This is either `base64' or `quoted-printable'."
161 ;; Don't base64, say, a short line with a single
162 ;; non-ASCII char when splitting parts by charset.
164 'quoted-printable
168 "Narrow the buffer to the header on the current line."
180 "Return the value of the field at point."
188 encodable-regexp)
190 Quoting will not be done in a quoted string if it contains characters
191 matching ENCODABLE-REGEXP or it is within parentheses."
195 beg end)
237 "The type of encoding done by `rfc2047-encode-region'.
238 This should be dynamically bound around calls to
239 `rfc2047-encode-region' to either `mime' or `address-mime'. See
243 "Encode the message header according to `rfc2047-header-encoding-alist'.
244 Should be called narrowed to the head of the message."
253 alist rfc2047-header-encoding-alist
255 ;; M$ Outlook boycotts decoding of a header if it consists
256 ;; of two or more encoded words and those charsets differ;
257 ;; it seems to decode all words in a header from a charset
258 ;; found first in the header. So, we unify the charsets into
259 ;; a single one used for encoding the whole text in a header.
263 mm-coding-system-priorities)
264 mm-coding-system-priorities)))
279 ;; 8 bit must be decoded.
284 ;; We found something that may perhaps be encoded.
294 mail-parse-charset)
296 mail-parse-charset)))
297 ;; We get this when CC'ing messages to newsgroups with
298 ;; 8-bit names. The group name mail copy just got
299 ;; unconditionally encoded. Previously, it would ask
300 ;; whether to encode, which was quite confusing for the
301 ;; user. If the new behavior is wrong, tell me. I have
302 ;; left the old code commented out below.
303 ;; -- Per Abrahamsen <abraham@dina.kvl.dk> Date: 2001-10-07.
304 ;; Modified by Dave Love, with the commented-out code changed
305 ;; in accordance with changes elsewhere.
308 ;;; ((null method)
309 ;;; (if (or (message-options-get
310 ;;; 'rfc2047-encode-message-header-encode-any)
311 ;;; (message-options-set
312 ;;; 'rfc2047-encode-message-header-encode-any
313 ;;; (y-or-n-p
314 ;;; "Some texts are not encoded. Encode anyway?")))
315 ;;; (rfc2047-encode-region (point-min) (point-max))
316 ;;; (error "Cannot send unencoded text")))
320 ;; Hm.
324 ;; Fixme: This, and the require below may not be the Right Thing, but
325 ;; should be safe just before release. -- fx 2001-02-08
328 "Return non-nil if any characters in current buffer need encoding in headers.
329 The buffer may be narrowed."
341 ;; Use this syntax table when parsing into regions that may need
342 ;; encoding. Double quotes are string delimiters, backslash is
343 ;; character quoting, and all other RFC 2822 special characters are
344 ;; treated as punctuation so we can use forward-sexp/forward-word to
345 ;; skip to the end of regions appropriately. Nb. ietf-drums does
346 ;; things differently.
348 ;; (make-char-table 'syntax-table '(2)) only works in Emacs.
350 ;; The following is done to work for setting all elements of the table;
351 ;; it appears to be the cleanest way.
352 ;; Play safe and don't assume the form of the word syntax entry --
353 ;; copy it from ?a.
367 table))
370 "Encode words in region B to E that need encoding.
371 By default, the region is treated as containing RFC2822 addresses.
372 Dynamically bind `rfc2047-encoding-type' to change that."
378 start ; start of current token
379 end begin csyntax
380 ;; Whether there's an encoded word before the current token,
381 ;; either immediately or separated by space.
382 last-encoded
385 ;; Simple case. Continuous words in which all those contain
386 ;; non-ASCII characters are encoded collectively. Encoding
387 ;; ASCII words, including `Re:' used in Subject headers, is
388 ;; avoided for interoperability with non-MIME clients and
389 ;; for making it easy to find keywords.
403 ;; `address-mime' case -- take care of quoted words, comments.
408 ;; Look for rfc2822-style: sequences of atoms, quoted
409 ;; strings, specials, whitespace. (Specials mustn't be
410 ;; encoded.)
412 ;; Skip whitespace.
417 ;; else token start
419 ;; Quoted word.
422 ;; Does it need encoding?
425 ;; It needs encoding. Strip the quotes first,
426 ;; since encoded words can't occur in quotes.
433 ;; There was a preceding quoted word. We need
434 ;; to include any separating whitespace in this
435 ;; word to avoid it getting lost.
437 ;; A space is needed between the encoded words.
441 ;; Adjust the end position for the deleted quotes.
446 ;; Skip other delimiters, but record that they've
447 ;; potentially separated quoted words.
453 ;; Look for the end of parentheses.
455 ;; Encode text as an unstructured field.
460 ;; Find the end.
461 ;; Skip one ASCII word, or encode continuous words
462 ;; in which all those contain non-ASCII characters.
473 ;; Skip backslash-quoted characters.
475 end)
484 ;; Where the value nil of `end' means there may be
485 ;; text to have to be encoded following the point.
486 ;; Otherwise, the point reached to the end of ASCII
487 ;; words separated by whitespace or a special char.
495 encodable-regexp end t))
501 nil)))
506 ;; There are special characters better
507 ;; to be encoded so that MTAs may parse
508 ;; them safely.
511 encodable-regexp
517 ;; Separate encodable text and
518 ;; delimiter.
538 ;; Also encode bogus delimiters.
540 ;; 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 ;; Remove 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 ;; Remove 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)
658 ;; Use utf-8 as a last resort if determining charset of text fails.
674 rfc2047-charset-encoding-alist))
675 ;; For the charsets that don't have a preferred
676 ;; encoding, choose the one that's shorter.
678 'B
683 ;; `tail' may contain a close parenthesis.
692 ;; `crest' may contain whitespace and an open parenthesis.
699 cs
701 rfc2047-encode-function-alist))
706 tail))
709 ;; The line was folded before encoding.
714 ;; `eword' contains `crest' and `tail'.
726 "Fold the current header field."
733 "Fold long lines in region B to E."
748 qword-break nil)
754 ;; Don't break before the first non-LWSP characters.
762 break nil
763 qword-break nil)
778 ;; Don't break at the start of the field.
789 qword-break nil)
795 ;; Don't break before the first non-LWSP characters.
801 "Fold the current line."
808 "Unfold lines in region B to E."
829 "Base64-encode the header contained in STRING."
835 "Quoted-printable-encode the header in STRING."
840 ;; = (\075), _ (\137), ? (\077) are used in the encoded word.
841 ;; Avoid using 8bit characters.
842 ;; This list excludes `especials' (see the RFC2047 syntax),
843 ;; meaning that some characters in non-structured fields will
844 ;; get encoded when they con't need to be. The following is
845 ;; what it used to be.
846 ;;; ;; Equivalent to "^\000-\007\011\013\015-\037\200-\377=_?"
847 ;;; "\010\012\014\040-\074\076\100-\136\140-\177")
853 "Return a PARAM=VALUE string encoded in the RFC2047-like style.
854 This is a substitution for the `rfc2231-encode-string' function, that
855 is the standard but many mailers don't support it."
860 ;;;
861 ;;; Functions for decoding RFC2047 messages
862 ;;;
868 "Non-nil means allow incomplete encoded-text in successive encoded-words.
869 Dividing of encoded-text in the place other than character boundaries
870 violates RFC2047 section 5, while we have a capability to decode it.
871 If it is non-nil, the decoder will decode B- or Q-encoding in each
872 encoded-word, concatenate them, and decode it by charset. Otherwise,
873 the decoder will fully decode each encoded-word before concatenating
874 them."
901 "Return coding-system corresponding to MIME CHARSET.
902 If your Emacs implementation can't decode CHARSET, return nil.
904 If allow-override is given, use `mm-charset-override-alist' to
905 map undesired charset names to their replacement. This should
906 only be used for decoding, not for encoding."
924 'raw-text
925 cs)))
930 "Decode successive encoded-words in WORDS and return a decoded string.
931 Each element of WORDS looks like (CHARSET ENCODING ENCODED-TEXT
932 ENCODED-WORD)."
947 nil)))
950 ;; Concatenate text of which the charset is the same.
953 ;; Don't decode encoded-word.
957 ;; FIXME: This looks O(N²). Can we make it more efficient
958 ;; with something like mapconcat?
965 nil)))
971 words)
973 words))
975 ;; Fixme: This should decode in place, not cons intermediate strings.
976 ;; Also check whether it needs to worry about delimiting fields like
977 ;; encoding.
979 ;; In fact it's reported that (invalid) encoding of mailboxes in
980 ;; addr-specs is in use, so delimiting fields might help. Probably
981 ;; not decoding a word which isn't properly delimited is good enough
982 ;; and worthwhile (is it more correct or not?), e.g. something like
983 ;; `=?iso-8859-1?q?foo?=@'.
986 "Decode MIME-encoded words in region between START and END.
987 If ADDRESS-MIME is non-nil, strip backslashes which precede characters
997 b e match words)
1004 ;; Look for the encoded-words.
1008 words nil)
1014 words)
1015 ;; Look for the subsequent encoded-words.
1018 ;; Replace the encoded-words with the decoded one.
1024 ;; Remove newlines between decoded words, though such
1025 ;; things essentially must not be there.
1029 ;; Quote decoded words if there are special characters
1030 ;; which might violate RFC2822.
1033 'address-mime
1034 rfc2047-header-encoding-alist))))
1039 ;; Don't quote words if already quoted.
1075 end t)))
1083 mail-parse-charset
1089 mail-parse-charset
1095 "Decode MIME-encoded words in region between START and END.
1097 strings are stripped."
1101 "Decode MIME-encoded STRING and return the result.
1102 If ADDRESS-MIME is non-nil, strip backslashes which precede characters
1106 ;; We used to only call mm-enable-multibyte if `m' is non-nil,
1107 ;; but this can't be the right criterion. Don't just revert this
1108 ;; change if it encounters a bug. Please help me fix it
1109 ;; right instead. --Stef
1110 ;; The string returned should always be multibyte in a multibyte
1111 ;; session, i.e. the buffer should be multibyte before
1112 ;; `buffer-string' is called.
1126 ;; Fixme: As above, `m' here is inappropriate.
1128 mail-parse-charset
1131 ;; `decode-coding-string' in Emacs offers a third optional
1132 ;; arg NOCOPY to avoid consing a new string if the decoding
1133 ;; is "trivial". Unfortunately it currently doesn't
1134 ;; consider anything else than a nil coding system
1135 ;; trivial.
1136 ;; `rfc2047-decode-string' is called multiple times for each
1137 ;; article during summary buffer generation, and we really
1138 ;; want to avoid unnecessary consing. So we bypass
1139 ;; `decode-coding-string' if the string is purely ASCII.
1141 ;; string is purely ASCII
1142 string
1145 string
1149 "Decode MIME-encoded STRING and return the result.
1151 strings are stripped."
1155 "Pad STRING to quartets."
1156 ;; Be more liberal to accept buggy base64 strings. If
1157 ;; base64-decode-string accepts buggy strings, this function could
1158 ;; be aliased to identity.
1160 string
1171 ;;; rfc2047.el ends here