Mercurial > emacs
annotate lisp/gnus/rfc2047.el @ 88919:acc369108294
(gnus-read-newsrc-el-file): Don't bind
coding-system-for-read.
(gnus-gnus-to-quick-newsrc-format): Insert coding cookie.
author | Dave Love <fx@gnu.org> |
---|---|
date | Tue, 30 Jul 2002 11:24:20 +0000 |
parents | 690e82320649 |
children | 4cda993e8f6e |
rev | line source |
---|---|
38413
a26d9b55abb6
Some fixes to follow coding conventions in files from Gnus.
Pavel Janík <Pavel@Janik.cz>
parents:
35985
diff
changeset
|
1 ;;; rfc2047.el --- functions for encoding and decoding rfc2047 messages |
31717 | 2 ;; Copyright (C) 1998, 1999, 2000 Free Software Foundation, Inc. |
3 | |
4 ;; Author: Lars Magne Ingebrigtsen <larsi@gnus.org> | |
5 ;; MORIOKA Tomohiko <morioka@jaist.ac.jp> | |
6 ;; This file is part of GNU Emacs. | |
7 | |
8 ;; GNU Emacs is free software; you can redistribute it and/or modify | |
9 ;; it under the terms of the GNU General Public License as published by | |
10 ;; the Free Software Foundation; either version 2, or (at your option) | |
11 ;; any later version. | |
12 | |
13 ;; GNU Emacs is distributed in the hope that it will be useful, | |
14 ;; but WITHOUT ANY WARRANTY; without even the implied warranty of | |
15 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
16 ;; GNU General Public License for more details. | |
17 | |
18 ;; You should have received a copy of the GNU General Public License | |
19 ;; along with GNU Emacs; see the file COPYING. If not, write to the | |
20 ;; Free Software Foundation, Inc., 59 Temple Place - Suite 330, | |
21 ;; Boston, MA 02111-1307, USA. | |
22 | |
23 ;;; Commentary: | |
24 | |
34674 | 25 ;; RFC 2047 is "MIME (Multipurpose Internet Mail Extensions) Part |
26 ;; Three: Message Header Extensions for Non-ASCII Text". | |
27 | |
31717 | 28 ;;; Code: |
29 | |
33304 | 30 (eval-when-compile (require 'cl)) |
31717 | 31 |
32 (require 'qp) | |
33 (require 'mm-util) | |
34 (require 'ietf-drums) | |
35 (require 'mail-prsvr) | |
33304 | 36 (require 'base64) |
37 ;; Fixme: Avoid this (for gnus-point-at-...) mm dependence on gnus. | |
38 (require 'gnus-util) | |
39 (autoload 'mm-body-7-or-8 "mm-bodies") | |
33127
eca95f9d7f05
(base64): Require unconditionally.
Dave Love <fx@gnu.org>
parents:
31764
diff
changeset
|
40 |
31717 | 41 (defvar rfc2047-header-encoding-alist |
42 '(("Newsgroups" . nil) | |
43 ("Message-ID" . nil) | |
44 (t . mime)) | |
45 "*Header/encoding method alist. | |
46 The list is traversed sequentially. The keys can either be | |
33304 | 47 header regexps or t. |
31717 | 48 |
49 The values can be: | |
50 | |
51 1) nil, in which case no encoding is done; | |
52 2) `mime', in which case the header will be encoded according to RFC2047; | |
53 3) a charset, in which case it will be encoded as that charset; | |
54 4) `default', in which case the field will be encoded as the rest | |
55 of the article.") | |
56 | |
57 (defvar rfc2047-charset-encoding-alist | |
58 '((us-ascii . nil) | |
59 (iso-8859-1 . Q) | |
60 (iso-8859-2 . Q) | |
61 (iso-8859-3 . Q) | |
62 (iso-8859-4 . Q) | |
63 (iso-8859-5 . B) | |
64 (koi8-r . B) | |
65 (iso-8859-7 . Q) | |
66 (iso-8859-8 . Q) | |
67 (iso-8859-9 . Q) | |
33304 | 68 (iso-8859-14 . Q) |
69 (iso-8859-15 . Q) | |
31717 | 70 (iso-2022-jp . B) |
71 (iso-2022-kr . B) | |
72 (gb2312 . B) | |
35838
53eebdb81828
2001-02-01 ShengHuo ZHU <zsh@cs.rochester.edu>
ShengHuo ZHU <zsh@cs.rochester.edu>
parents:
35453
diff
changeset
|
73 (big5 . B) |
53eebdb81828
2001-02-01 ShengHuo ZHU <zsh@cs.rochester.edu>
ShengHuo ZHU <zsh@cs.rochester.edu>
parents:
35453
diff
changeset
|
74 (cn-big5 . B) |
31717 | 75 (cn-gb . B) |
76 (cn-gb-2312 . B) | |
77 (euc-kr . B) | |
78 (iso-2022-jp-2 . B) | |
79 (iso-2022-int-1 . B)) | |
80 "Alist of MIME charsets to RFC2047 encodings. | |
81 Valid encodings are nil, `Q' and `B'.") | |
82 | |
83 (defvar rfc2047-encoding-function-alist | |
84 '((Q . rfc2047-q-encode-region) | |
85 (B . rfc2047-b-encode-region) | |
86 (nil . ignore)) | |
87 "Alist of RFC2047 encodings to encoding functions.") | |
88 | |
89 (defvar rfc2047-q-encoding-alist | |
33304 | 90 '(("\\(From\\|Cc\\|To\\|Bcc\||Reply-To\\):" . "-A-Za-z0-9!*+/") |
31764 | 91 ;; = (\075), _ (\137), ? (\077) are used in the encoded word. |
35985
b9c371244b90
(rfc2047-fold-region): Don't forward-char at EOB.
Dave Love <fx@gnu.org>
parents:
35838
diff
changeset
|
92 ;; Avoid using 8bit characters. |
31764 | 93 ;; Equivalent to "^\000-\007\011\013\015-\037\200-\377=_?" |
94 ("." . "\010\012\014\040-\074\076\100-\136\140-\177")) | |
31717 | 95 "Alist of header regexps and valid Q characters.") |
96 | |
97 ;;; | |
98 ;;; Functions for encoding RFC2047 messages | |
99 ;;; | |
100 | |
101 (defun rfc2047-narrow-to-field () | |
102 "Narrow the buffer to the header on the current line." | |
103 (beginning-of-line) | |
104 (narrow-to-region | |
105 (point) | |
106 (progn | |
107 (forward-line 1) | |
108 (if (re-search-forward "^[^ \n\t]" nil t) | |
109 (progn | |
110 (beginning-of-line) | |
111 (point)) | |
112 (point-max)))) | |
113 (goto-char (point-min))) | |
114 | |
115 (defun rfc2047-encode-message-header () | |
116 "Encode the message header according to `rfc2047-header-encoding-alist'. | |
117 Should be called narrowed to the head of the message." | |
118 (interactive "*") | |
119 (save-excursion | |
120 (goto-char (point-min)) | |
121 (let (alist elem method) | |
122 (while (not (eobp)) | |
123 (save-restriction | |
124 (rfc2047-narrow-to-field) | |
125 (if (not (rfc2047-encodable-p)) | |
126 (if (and (eq (mm-body-7-or-8) '8bit) | |
127 (mm-multibyte-p) | |
128 (mm-coding-system-p | |
129 (car message-posting-charset))) | |
130 ;; 8 bit must be decoded. | |
131 ;; Is message-posting-charset a coding system? | |
33304 | 132 (mm-encode-coding-region |
133 (point-min) (point-max) | |
31717 | 134 (car message-posting-charset))) |
135 ;; We found something that may perhaps be encoded. | |
136 (setq method nil | |
137 alist rfc2047-header-encoding-alist) | |
138 (while (setq elem (pop alist)) | |
139 (when (or (and (stringp (car elem)) | |
140 (looking-at (car elem))) | |
141 (eq (car elem) t)) | |
142 (setq alist nil | |
143 method (cdr elem)))) | |
144 (cond | |
145 ((eq method 'mime) | |
33304 | 146 (rfc2047-encode-region (point-min) (point-max))) |
31717 | 147 ((eq method 'default) |
148 (if (and (featurep 'mule) | |
33815
61c7f3065929
(rfc2047-encode-message-header): Don't encode if
Dave Love <fx@gnu.org>
parents:
33304
diff
changeset
|
149 (if (boundp 'default-enable-multibyte-characters) |
61c7f3065929
(rfc2047-encode-message-header): Don't encode if
Dave Love <fx@gnu.org>
parents:
33304
diff
changeset
|
150 default-enable-multibyte-characters) |
31717 | 151 mail-parse-charset) |
33304 | 152 (mm-encode-coding-region (point-min) (point-max) |
31717 | 153 mail-parse-charset))) |
154 ((mm-coding-system-p method) | |
33815
61c7f3065929
(rfc2047-encode-message-header): Don't encode if
Dave Love <fx@gnu.org>
parents:
33304
diff
changeset
|
155 (if (and (featurep 'mule) |
61c7f3065929
(rfc2047-encode-message-header): Don't encode if
Dave Love <fx@gnu.org>
parents:
33304
diff
changeset
|
156 (if (boundp 'default-enable-multibyte-characters) |
61c7f3065929
(rfc2047-encode-message-header): Don't encode if
Dave Love <fx@gnu.org>
parents:
33304
diff
changeset
|
157 default-enable-multibyte-characters)) |
31717 | 158 (mm-encode-coding-region (point-min) (point-max) method))) |
159 ;; Hm. | |
160 (t))) | |
161 (goto-char (point-max))))))) | |
162 | |
35985
b9c371244b90
(rfc2047-fold-region): Don't forward-char at EOB.
Dave Love <fx@gnu.org>
parents:
35838
diff
changeset
|
163 ;; Fixme: This, and the require below may not be the Right Thing, but |
b9c371244b90
(rfc2047-fold-region): Don't forward-char at EOB.
Dave Love <fx@gnu.org>
parents:
35838
diff
changeset
|
164 ;; should be safe just before release. -- fx 2001-02-08 |
b9c371244b90
(rfc2047-fold-region): Don't forward-char at EOB.
Dave Love <fx@gnu.org>
parents:
35838
diff
changeset
|
165 (eval-when-compile (defvar message-posting-charset)) |
b9c371244b90
(rfc2047-fold-region): Don't forward-char at EOB.
Dave Love <fx@gnu.org>
parents:
35838
diff
changeset
|
166 |
33304 | 167 (defun rfc2047-encodable-p () |
168 "Return non-nil if any characters in current buffer need encoding in headers. | |
169 The buffer may be narrowed." | |
35985
b9c371244b90
(rfc2047-fold-region): Don't forward-char at EOB.
Dave Love <fx@gnu.org>
parents:
35838
diff
changeset
|
170 (require 'message) ; for message-posting-charset |
31717 | 171 (let ((charsets |
88792
690e82320649
(rfc2047-encodable-p): Avoid mm-find-charset-region.
Dave Love <fx@gnu.org>
parents:
38413
diff
changeset
|
172 (mm-find-mime-charset-region (point-min) (point-max)))) |
690e82320649
(rfc2047-encodable-p): Avoid mm-find-charset-region.
Dave Love <fx@gnu.org>
parents:
38413
diff
changeset
|
173 (and charsets (not (equal charsets (list message-posting-charset)))))) |
31717 | 174 |
175 (defun rfc2047-dissect-region (b e) | |
176 "Dissect the region between B and E into words." | |
33304 | 177 (let ((word-chars "-A-Za-z0-9!*+/") |
178 ;; Not using ietf-drums-specials-token makes life simple. | |
179 mail-parse-mule-charset | |
88792
690e82320649
(rfc2047-encodable-p): Avoid mm-find-charset-region.
Dave Love <fx@gnu.org>
parents:
38413
diff
changeset
|
180 words point nonascii |
33304 | 181 result word) |
31717 | 182 (save-restriction |
183 (narrow-to-region b e) | |
184 (goto-char (point-min)) | |
33304 | 185 (skip-chars-forward "\000-\177") |
88792
690e82320649
(rfc2047-encodable-p): Avoid mm-find-charset-region.
Dave Love <fx@gnu.org>
parents:
38413
diff
changeset
|
186 ;; Fixme: This loop used to check charsets when it found |
690e82320649
(rfc2047-encodable-p): Avoid mm-find-charset-region.
Dave Love <fx@gnu.org>
parents:
38413
diff
changeset
|
187 ;; non-ASCII characters. That's removed, since it doesn't make |
690e82320649
(rfc2047-encodable-p): Avoid mm-find-charset-region.
Dave Love <fx@gnu.org>
parents:
38413
diff
changeset
|
188 ;; much sense in Emacs 22 and doesn't seem necessary in Emacs |
690e82320649
(rfc2047-encodable-p): Avoid mm-find-charset-region.
Dave Love <fx@gnu.org>
parents:
38413
diff
changeset
|
189 ;; 21, even. I'm not sure exactly what it should be doing, and |
690e82320649
(rfc2047-encodable-p): Avoid mm-find-charset-region.
Dave Love <fx@gnu.org>
parents:
38413
diff
changeset
|
190 ;; it needs another look, especially for efficiency's sake. -- fx |
31717 | 191 (while (not (eobp)) |
88792
690e82320649
(rfc2047-encodable-p): Avoid mm-find-charset-region.
Dave Love <fx@gnu.org>
parents:
38413
diff
changeset
|
192 (setq point (point) |
690e82320649
(rfc2047-encodable-p): Avoid mm-find-charset-region.
Dave Love <fx@gnu.org>
parents:
38413
diff
changeset
|
193 nonascii nil) |
33304 | 194 (skip-chars-backward word-chars b) |
195 (unless (eq b (point)) | |
196 (push (cons (buffer-substring b (point)) nil) words)) | |
88792
690e82320649
(rfc2047-encodable-p): Avoid mm-find-charset-region.
Dave Love <fx@gnu.org>
parents:
38413
diff
changeset
|
197 (setq b (point) |
690e82320649
(rfc2047-encodable-p): Avoid mm-find-charset-region.
Dave Love <fx@gnu.org>
parents:
38413
diff
changeset
|
198 nonascii t) |
33304 | 199 (goto-char point) |
200 (forward-char 1) | |
201 (skip-chars-forward word-chars) | |
88792
690e82320649
(rfc2047-encodable-p): Avoid mm-find-charset-region.
Dave Love <fx@gnu.org>
parents:
38413
diff
changeset
|
202 (while (not (eobp)) |
33304 | 203 (forward-char 1) |
204 (skip-chars-forward word-chars)) | |
205 (unless (eq b (point)) | |
88792
690e82320649
(rfc2047-encodable-p): Avoid mm-find-charset-region.
Dave Love <fx@gnu.org>
parents:
38413
diff
changeset
|
206 (push (cons (buffer-substring b (point)) nonascii) words)) |
33304 | 207 (setq b (point)) |
208 (skip-chars-forward "\000-\177")) | |
209 (unless (eq b (point)) | |
210 (push (cons (buffer-substring b (point)) nil) words))) | |
211 ;; merge adjacent words | |
212 (setq word (pop words)) | |
213 (while word | |
214 (if (and (cdr word) | |
215 (caar words) | |
216 (not (cdar words)) | |
217 (not (string-match "[^ \t]" (caar words)))) | |
218 (if (eq (cdr (nth 1 words)) (cdr word)) | |
219 (progn | |
220 (setq word (cons (concat | |
221 (car (nth 1 words)) (caar words) | |
222 (car word)) | |
223 (cdr word))) | |
224 (pop words) | |
225 (pop words)) | |
226 (push (cons (concat (caar words) (car word)) (cdr word)) | |
227 result) | |
228 (pop words) | |
229 (setq word (pop words))) | |
230 (push word result) | |
231 (setq word (pop words)))) | |
232 result)) | |
31717 | 233 |
234 (defun rfc2047-encode-region (b e) | |
35985
b9c371244b90
(rfc2047-fold-region): Don't forward-char at EOB.
Dave Love <fx@gnu.org>
parents:
35838
diff
changeset
|
235 "Encode all encodable words in region B to E." |
33304 | 236 (let ((words (rfc2047-dissect-region b e)) word) |
237 (save-restriction | |
238 (narrow-to-region b e) | |
239 (delete-region (point-min) (point-max)) | |
240 (while (setq word (pop words)) | |
241 (if (not (cdr word)) | |
242 (insert (car word)) | |
243 (rfc2047-fold-region (gnus-point-at-bol) (point)) | |
244 (goto-char (point-max)) | |
245 (if (> (- (point) (save-restriction | |
246 (widen) | |
247 (gnus-point-at-bol))) 76) | |
248 (insert "\n ")) | |
249 ;; Insert blank between encoded words | |
250 (if (eq (char-before) ?=) (insert " ")) | |
251 (rfc2047-encode (point) | |
88792
690e82320649
(rfc2047-encodable-p): Avoid mm-find-charset-region.
Dave Love <fx@gnu.org>
parents:
38413
diff
changeset
|
252 (progn (insert (car word)) (point))))) |
33304 | 253 (rfc2047-fold-region (point-min) (point-max))))) |
31717 | 254 |
255 (defun rfc2047-encode-string (string) | |
256 "Encode words in STRING." | |
257 (with-temp-buffer | |
258 (insert string) | |
259 (rfc2047-encode-region (point-min) (point-max)) | |
260 (buffer-string))) | |
261 | |
88792
690e82320649
(rfc2047-encodable-p): Avoid mm-find-charset-region.
Dave Love <fx@gnu.org>
parents:
38413
diff
changeset
|
262 (defun rfc2047-encode (b e) |
690e82320649
(rfc2047-encodable-p): Avoid mm-find-charset-region.
Dave Love <fx@gnu.org>
parents:
38413
diff
changeset
|
263 "Encode the word in the region B to E." |
690e82320649
(rfc2047-encodable-p): Avoid mm-find-charset-region.
Dave Love <fx@gnu.org>
parents:
38413
diff
changeset
|
264 (let* ((buff (current-buffer)) |
690e82320649
(rfc2047-encodable-p): Avoid mm-find-charset-region.
Dave Love <fx@gnu.org>
parents:
38413
diff
changeset
|
265 (mime-charset (with-temp-buffer |
690e82320649
(rfc2047-encodable-p): Avoid mm-find-charset-region.
Dave Love <fx@gnu.org>
parents:
38413
diff
changeset
|
266 (insert-buffer-substring buff b e) |
690e82320649
(rfc2047-encodable-p): Avoid mm-find-charset-region.
Dave Love <fx@gnu.org>
parents:
38413
diff
changeset
|
267 (mm-find-mime-charset-region b e))) |
690e82320649
(rfc2047-encodable-p): Avoid mm-find-charset-region.
Dave Love <fx@gnu.org>
parents:
38413
diff
changeset
|
268 (cs (if (> (length mime-charset) 1) |
690e82320649
(rfc2047-encodable-p): Avoid mm-find-charset-region.
Dave Love <fx@gnu.org>
parents:
38413
diff
changeset
|
269 (mm-charset-to-coding-system mime-charset) |
690e82320649
(rfc2047-encodable-p): Avoid mm-find-charset-region.
Dave Love <fx@gnu.org>
parents:
38413
diff
changeset
|
270 (error "Can't encode word: %s" (buffer-substring b e)))) |
31717 | 271 (encoding (or (cdr (assq mime-charset |
272 rfc2047-charset-encoding-alist)) | |
273 'B)) | |
274 (start (concat | |
275 "=?" (downcase (symbol-name mime-charset)) "?" | |
276 (downcase (symbol-name encoding)) "?")) | |
277 (first t)) | |
278 (save-restriction | |
279 (narrow-to-region b e) | |
280 (when (eq encoding 'B) | |
281 ;; break into lines before encoding | |
282 (goto-char (point-min)) | |
283 (while (not (eobp)) | |
284 (goto-char (min (point-max) (+ 15 (point)))) | |
285 (unless (eobp) | |
286 (insert "\n")))) | |
287 (if (and (mm-multibyte-p) | |
35838
53eebdb81828
2001-02-01 ShengHuo ZHU <zsh@cs.rochester.edu>
ShengHuo ZHU <zsh@cs.rochester.edu>
parents:
35453
diff
changeset
|
288 (mm-coding-system-p cs)) |
53eebdb81828
2001-02-01 ShengHuo ZHU <zsh@cs.rochester.edu>
ShengHuo ZHU <zsh@cs.rochester.edu>
parents:
35453
diff
changeset
|
289 (mm-encode-coding-region (point-min) (point-max) cs)) |
31717 | 290 (funcall (cdr (assq encoding rfc2047-encoding-function-alist)) |
291 (point-min) (point-max)) | |
292 (goto-char (point-min)) | |
293 (while (not (eobp)) | |
294 (unless first | |
295 (insert " ")) | |
296 (setq first nil) | |
297 (insert start) | |
298 (end-of-line) | |
299 (insert "?=") | |
300 (forward-line 1))))) | |
301 | |
302 (defun rfc2047-fold-region (b e) | |
35985
b9c371244b90
(rfc2047-fold-region): Don't forward-char at EOB.
Dave Love <fx@gnu.org>
parents:
35838
diff
changeset
|
303 "Fold long lines in region B to E." |
31717 | 304 (save-restriction |
305 (narrow-to-region b e) | |
306 (goto-char (point-min)) | |
33304 | 307 (let ((break nil) |
308 (qword-break nil) | |
309 (bol (save-restriction | |
310 (widen) | |
311 (gnus-point-at-bol)))) | |
31717 | 312 (while (not (eobp)) |
33304 | 313 (when (and (or break qword-break) (> (- (point) bol) 76)) |
314 (goto-char (or break qword-break)) | |
315 (setq break nil | |
316 qword-break nil) | |
35453
26726eff41ca
2001-01-21 ShengHuo ZHU <zsh@cs.rochester.edu>
ShengHuo ZHU <zsh@cs.rochester.edu>
parents:
34674
diff
changeset
|
317 (if (looking-at " \t") |
26726eff41ca
2001-01-21 ShengHuo ZHU <zsh@cs.rochester.edu>
ShengHuo ZHU <zsh@cs.rochester.edu>
parents:
34674
diff
changeset
|
318 (insert "\n") |
26726eff41ca
2001-01-21 ShengHuo ZHU <zsh@cs.rochester.edu>
ShengHuo ZHU <zsh@cs.rochester.edu>
parents:
34674
diff
changeset
|
319 (insert "\n ")) |
33304 | 320 (setq bol (1- (point))) |
321 ;; Don't break before the first non-LWSP characters. | |
322 (skip-chars-forward " \t") | |
35985
b9c371244b90
(rfc2047-fold-region): Don't forward-char at EOB.
Dave Love <fx@gnu.org>
parents:
35838
diff
changeset
|
323 (unless (eobp) (forward-char 1))) |
31717 | 324 (cond |
33304 | 325 ((eq (char-after) ?\n) |
326 (forward-char 1) | |
327 (setq bol (point) | |
328 break nil | |
329 qword-break nil) | |
330 (skip-chars-forward " \t") | |
331 (unless (or (eobp) (eq (char-after) ?\n)) | |
332 (forward-char 1))) | |
333 ((eq (char-after) ?\r) | |
334 (forward-char 1)) | |
31717 | 335 ((memq (char-after) '(? ?\t)) |
33304 | 336 (skip-chars-forward " \t") |
337 (setq break (1- (point)))) | |
338 ((not break) | |
339 (if (not (looking-at "=\\?[^=]")) | |
340 (if (eq (char-after) ?=) | |
341 (forward-char 1) | |
342 (skip-chars-forward "^ \t\n\r=")) | |
343 (setq qword-break (point)) | |
344 (skip-chars-forward "^ \t\n\r"))) | |
345 (t | |
346 (skip-chars-forward "^ \t\n\r")))) | |
347 (when (and (or break qword-break) (> (- (point) bol) 76)) | |
348 (goto-char (or break qword-break)) | |
349 (setq break nil | |
350 qword-break nil) | |
35453
26726eff41ca
2001-01-21 ShengHuo ZHU <zsh@cs.rochester.edu>
ShengHuo ZHU <zsh@cs.rochester.edu>
parents:
34674
diff
changeset
|
351 (if (looking-at " \t") |
26726eff41ca
2001-01-21 ShengHuo ZHU <zsh@cs.rochester.edu>
ShengHuo ZHU <zsh@cs.rochester.edu>
parents:
34674
diff
changeset
|
352 (insert "\n") |
26726eff41ca
2001-01-21 ShengHuo ZHU <zsh@cs.rochester.edu>
ShengHuo ZHU <zsh@cs.rochester.edu>
parents:
34674
diff
changeset
|
353 (insert "\n ")) |
33304 | 354 (setq bol (1- (point))) |
355 ;; Don't break before the first non-LWSP characters. | |
356 (skip-chars-forward " \t") | |
35985
b9c371244b90
(rfc2047-fold-region): Don't forward-char at EOB.
Dave Love <fx@gnu.org>
parents:
35838
diff
changeset
|
357 (unless (eobp) (forward-char 1)))))) |
33304 | 358 |
359 (defun rfc2047-unfold-region (b e) | |
35985
b9c371244b90
(rfc2047-fold-region): Don't forward-char at EOB.
Dave Love <fx@gnu.org>
parents:
35838
diff
changeset
|
360 "Unfold lines in region B to E." |
33304 | 361 (save-restriction |
362 (narrow-to-region b e) | |
363 (goto-char (point-min)) | |
364 (let ((bol (save-restriction | |
365 (widen) | |
366 (gnus-point-at-bol))) | |
367 (eol (gnus-point-at-eol)) | |
368 leading) | |
369 (forward-line 1) | |
370 (while (not (eobp)) | |
371 (looking-at "[ \t]*") | |
372 (setq leading (- (match-end 0) (match-beginning 0))) | |
373 (if (< (- (gnus-point-at-eol) bol leading) 76) | |
374 (progn | |
375 (goto-char eol) | |
376 (delete-region eol (progn | |
377 (skip-chars-forward "[ \t\n\r]+") | |
378 (1- (point))))) | |
379 (setq bol (gnus-point-at-bol))) | |
380 (setq eol (gnus-point-at-eol)) | |
381 (forward-line 1))))) | |
31717 | 382 |
383 (defun rfc2047-b-encode-region (b e) | |
33304 | 384 "Base64-encode the header contained in region B to E." |
31717 | 385 (save-restriction |
386 (narrow-to-region (goto-char b) e) | |
387 (while (not (eobp)) | |
388 (base64-encode-region (point) (progn (end-of-line) (point)) t) | |
389 (if (and (bolp) (eolp)) | |
390 (delete-backward-char 1)) | |
391 (forward-line)))) | |
392 | |
393 (defun rfc2047-q-encode-region (b e) | |
33304 | 394 "Quoted-printable-encode the header in region B to E." |
31717 | 395 (save-excursion |
396 (save-restriction | |
397 (narrow-to-region (goto-char b) e) | |
33304 | 398 (let ((alist rfc2047-q-encoding-alist) |
399 (bol (save-restriction | |
400 (widen) | |
401 (gnus-point-at-bol)))) | |
31717 | 402 (while alist |
403 (when (looking-at (caar alist)) | |
404 (quoted-printable-encode-region b e nil (cdar alist)) | |
405 (subst-char-in-region (point-min) (point-max) ? ?_) | |
406 (setq alist nil)) | |
407 (pop alist)) | |
33304 | 408 ;; The size of QP encapsulation is about 20, so set limit to |
409 ;; 56=76-20. | |
410 (unless (< (- (point-max) (point-min)) 56) | |
411 ;; Don't break if it could fit in one line. | |
412 ;; Let rfc2047-encode-region break it later. | |
413 (goto-char (1+ (point-min))) | |
414 (while (and (not (bobp)) (not (eobp))) | |
415 (goto-char (min (point-max) (+ 56 bol))) | |
416 (search-backward "=" (- (point) 2) t) | |
417 (unless (or (bobp) (eobp)) | |
418 (insert "\n") | |
419 (setq bol (point))))))))) | |
31717 | 420 |
421 ;;; | |
422 ;;; Functions for decoding RFC2047 messages | |
423 ;;; | |
424 | |
425 (defvar rfc2047-encoded-word-regexp | |
426 "=\\?\\([^][\000-\040()<>@,\;:\\\"/?.=]+\\)\\?\\(B\\|Q\\)\\?\\([!->@-~ +]+\\)\\?=") | |
427 | |
428 (defun rfc2047-decode-region (start end) | |
429 "Decode MIME-encoded words in region between START and END." | |
430 (interactive "r") | |
431 (let ((case-fold-search t) | |
432 b e) | |
433 (save-excursion | |
434 (save-restriction | |
435 (narrow-to-region start end) | |
436 (goto-char (point-min)) | |
437 ;; Remove whitespace between encoded words. | |
438 (while (re-search-forward | |
439 (concat "\\(" rfc2047-encoded-word-regexp "\\)" | |
440 "\\(\n?[ \t]\\)+" | |
441 "\\(" rfc2047-encoded-word-regexp "\\)") | |
442 nil t) | |
443 (delete-region (goto-char (match-end 1)) (match-beginning 6))) | |
444 ;; Decode the encoded words. | |
445 (setq b (goto-char (point-min))) | |
446 (while (re-search-forward rfc2047-encoded-word-regexp nil t) | |
447 (setq e (match-beginning 0)) | |
448 (insert (rfc2047-parse-and-decode | |
449 (prog1 | |
450 (match-string 0) | |
451 (delete-region (match-beginning 0) (match-end 0))))) | |
452 (when (and (mm-multibyte-p) | |
453 mail-parse-charset | |
454 (not (eq mail-parse-charset 'gnus-decoded))) | |
455 (mm-decode-coding-region b e mail-parse-charset)) | |
456 (setq b (point))) | |
457 (when (and (mm-multibyte-p) | |
458 mail-parse-charset | |
459 (not (eq mail-parse-charset 'us-ascii)) | |
460 (not (eq mail-parse-charset 'gnus-decoded))) | |
33304 | 461 (mm-decode-coding-region b (point-max) mail-parse-charset)) |
462 (rfc2047-unfold-region (point-min) (point-max)))))) | |
31717 | 463 |
464 (defun rfc2047-decode-string (string) | |
465 "Decode the quoted-printable-encoded STRING and return the results." | |
466 (let ((m (mm-multibyte-p))) | |
467 (with-temp-buffer | |
468 (when m | |
469 (mm-enable-multibyte)) | |
470 (insert string) | |
471 (inline | |
472 (rfc2047-decode-region (point-min) (point-max))) | |
473 (buffer-string)))) | |
474 | |
475 (defun rfc2047-parse-and-decode (word) | |
476 "Decode WORD and return it if it is an encoded word. | |
477 Return WORD if not." | |
478 (if (not (string-match rfc2047-encoded-word-regexp word)) | |
479 word | |
480 (or | |
481 (condition-case nil | |
482 (rfc2047-decode | |
483 (match-string 1 word) | |
484 (upcase (match-string 2 word)) | |
485 (match-string 3 word)) | |
486 (error word)) | |
487 word))) | |
488 | |
489 (defun rfc2047-decode (charset encoding string) | |
33304 | 490 "Decode STRING from the given MIME CHARSET in the given ENCODING. |
31717 | 491 Valid ENCODINGs are \"B\" and \"Q\". |
33304 | 492 If your Emacs implementation can't decode CHARSET, return nil." |
31717 | 493 (if (stringp charset) |
494 (setq charset (intern (downcase charset)))) | |
33304 | 495 (if (or (not charset) |
31717 | 496 (eq 'gnus-all mail-parse-ignored-charsets) |
497 (memq 'gnus-all mail-parse-ignored-charsets) | |
498 (memq charset mail-parse-ignored-charsets)) | |
499 (setq charset mail-parse-charset)) | |
500 (let ((cs (mm-charset-to-coding-system charset))) | |
33304 | 501 (if (and (not cs) charset |
31717 | 502 (listp mail-parse-ignored-charsets) |
503 (memq 'gnus-unknown mail-parse-ignored-charsets)) | |
504 (setq cs (mm-charset-to-coding-system mail-parse-charset))) | |
505 (when cs | |
506 (when (and (eq cs 'ascii) | |
507 mail-parse-charset) | |
508 (setq cs mail-parse-charset)) | |
33304 | 509 ;; Ensure unibyte result in Emacs 20. |
510 (let (default-enable-multibyte-characters) | |
511 (with-temp-buffer | |
512 (mm-decode-coding-string | |
513 (cond | |
514 ((equal "B" encoding) | |
515 (base64-decode-string string)) | |
516 ((equal "Q" encoding) | |
517 (quoted-printable-decode-string | |
518 (mm-replace-chars-in-string string ?_ ? ))) | |
519 (t (error "Invalid encoding: %s" encoding))) | |
520 cs)))))) | |
31717 | 521 |
522 (provide 'rfc2047) | |
523 | |
524 ;;; rfc2047.el ends here |