Mercurial > emacs
annotate lisp/international/utf-7.el @ 103928:fa669d6474cc
Refcard comments.
author | Glenn Morris <rgm@gnu.org> |
---|---|
date | Thu, 16 Jul 2009 03:30:31 +0000 |
parents | a9dc0e7c3f2b |
children | 1d1d5d9bd884 |
rev | line source |
---|---|
50897 | 1 ;;; utf-7.el --- utf-7 coding system |
2 | |
100908 | 3 ;; Copyright (C) 2003, 2004, 2005, 2006, 2007, 2008, 2009 |
94684
bfec2b1b2136
Switch to recommended form of GPLv3 permissions notice.
Glenn Morris <rgm@gnu.org>
parents:
93975
diff
changeset
|
4 ;; Free Software Foundation, Inc. |
50897 | 5 |
6 ;; Author: Dave Love <fx@gnu.org> | |
7 ;; Keywords: i18n, mail | |
8 | |
94684
bfec2b1b2136
Switch to recommended form of GPLv3 permissions notice.
Glenn Morris <rgm@gnu.org>
parents:
93975
diff
changeset
|
9 ;; This file is part of GNU Emacs. |
bfec2b1b2136
Switch to recommended form of GPLv3 permissions notice.
Glenn Morris <rgm@gnu.org>
parents:
93975
diff
changeset
|
10 |
bfec2b1b2136
Switch to recommended form of GPLv3 permissions notice.
Glenn Morris <rgm@gnu.org>
parents:
93975
diff
changeset
|
11 ;; GNU Emacs is free software: you can redistribute it and/or modify |
50897 | 12 ;; it under the terms of the GNU General Public License as published by |
94684
bfec2b1b2136
Switch to recommended form of GPLv3 permissions notice.
Glenn Morris <rgm@gnu.org>
parents:
93975
diff
changeset
|
13 ;; the Free Software Foundation, either version 3 of the License, or |
bfec2b1b2136
Switch to recommended form of GPLv3 permissions notice.
Glenn Morris <rgm@gnu.org>
parents:
93975
diff
changeset
|
14 ;; (at your option) any later version. |
50897 | 15 |
94684
bfec2b1b2136
Switch to recommended form of GPLv3 permissions notice.
Glenn Morris <rgm@gnu.org>
parents:
93975
diff
changeset
|
16 ;; GNU Emacs is distributed in the hope that it will be useful, |
50897 | 17 ;; but WITHOUT ANY WARRANTY; without even the implied warranty of |
18 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
19 ;; GNU General Public License for more details. | |
20 | |
21 ;; You should have received a copy of the GNU General Public License | |
94684
bfec2b1b2136
Switch to recommended form of GPLv3 permissions notice.
Glenn Morris <rgm@gnu.org>
parents:
93975
diff
changeset
|
22 ;; along with GNU Emacs. If not, see <http://www.gnu.org/licenses/>. |
50897 | 23 |
24 ;;; Commentary: | |
25 | |
26 ;; Defines a coding system for UTF-7, defined in RFC 2152. Non-ASCII | |
27 ;; segments are encoded as base64-encoded big endian UTF-16. Also | |
28 ;; defines a variation required for IMAP (RFC 2060). | |
29 | |
30 ;; The encoding and decoding was originally taken from Jon K Hellan's | |
31 ;; implementation in Gnus, but has been substantially re-done. | |
32 | |
33 ;; This probably needs more attention. In particular, it's not | |
96376
c3309dba6542
American English spelling fix.
Glenn Morris <rgm@gnu.org>
parents:
95395
diff
changeset
|
34 ;; completely consistent with iconv's behavior. It's arguable |
50897 | 35 ;; whether the IMAP version should be a coding system since it's |
36 ;; apparently only used for IMAP mailbox names, so it's commented out. | |
37 | |
38 ;;; Code: | |
39 | |
40 (defun utf-7-decode (len imap) | |
41 "Decode LEN bytes of UTF-7 at point. | |
42 IMAP non-nil means use the IMAP version." | |
43 (save-excursion | |
44 (save-restriction | |
45 (narrow-to-region (point) (+ (point) len)) | |
46 (let ((not-esc (if imap "^&" "^+")) | |
47 (skip-chars (if imap "A-Za-z0-9+," "A-Za-z0-9+/"))) | |
48 (while (not (eobp)) | |
49 (skip-chars-forward not-esc) | |
50 (unless (eobp) | |
51 (forward-char) | |
52 (let ((p (point)) | |
53 (run-length (skip-chars-forward skip-chars))) | |
54 (if (eq ?- (char-after)) | |
55 (delete-char 1)) | |
56 (unless (= run-length 0) ; encoded lone esc-char | |
57 (let ((pl (mod (- run-length) 4))) | |
58 (insert-char ?= pl) | |
59 (if imap | |
60 (subst-char-in-region p (point) ?, ?/)) | |
61 (base64-decode-region p (point))) | |
51630
90215b7be64d
Adjusted for the name change: xxx-utf-16-{le,be} -> xxx-utf-16{le,be}.
Kenichi Handa <handa@m17n.org>
parents:
50897
diff
changeset
|
62 (decode-coding-region p (point) 'utf-16be) |
50897 | 63 (save-excursion |
64 (goto-char p) | |
65 (delete-backward-char 1))))))) | |
66 (- (point-max) (point-min))))) | |
67 | |
90250
1c8f24770e68
Delete the definition of utf-7.
Kenichi Handa <handa@m17n.org>
parents:
90200
diff
changeset
|
68 ;;;###autoload |
50897 | 69 (defun utf-7-post-read-conversion (len) |
70 (utf-7-decode len nil)) | |
71 | |
91117
57ad20761f87
(utf-7-imap-post-read-conversion)
Kenichi Handa <handa@m17n.org>
parents:
90996
diff
changeset
|
72 ;;;###autoload |
85926
025d69230a9b
(utf-7-imap): New coding system.
Kenichi Handa <handa@m17n.org>
parents:
78274
diff
changeset
|
73 (defun utf-7-imap-post-read-conversion (len) |
025d69230a9b
(utf-7-imap): New coding system.
Kenichi Handa <handa@m17n.org>
parents:
78274
diff
changeset
|
74 (utf-7-decode len t)) |
50897 | 75 |
76 (defun utf-7-encode (from to imap) | |
77 "Encode bytes between FROM and TO to UTF-7. | |
78 ESC and SKIP-CHARS are adjusted for the normal and IMAP versions." | |
79 (let* ((old-buf (current-buffer)) | |
80 (esc (if imap ?& ?+)) | |
81 ;; These are characters which can be encoded asis. | |
82 (skip-chars (if imap | |
83 "\t\n\r\x20-\x25\x27-\x7e" ; rfc2060 | |
84 ;; This includes the rfc2152 optional set. | |
85 ;; Perhaps it shouldn't (like iconv). | |
86 "\t\n\r -*,-[]-}")) | |
87 (not-skip-chars (format "^%s%c" skip-chars esc))) | |
88 (set-buffer (generate-new-buffer " *temp*")) | |
89 (if (stringp from) | |
90 (insert from) | |
91 (insert-buffer-substring old-buf from to)) | |
92 (goto-char (point-min)) | |
93 (while (not (eobp)) | |
94 (skip-chars-forward skip-chars) | |
95395
a374273bf2a5
(utf-7-encode): Use the right escape char depending on imap/nonimap encoding.
Stefan Monnier <monnier@iro.umontreal.ca>
parents:
94684
diff
changeset
|
95 (if (eq esc (char-after)) |
50897 | 96 (progn (forward-char) |
97 (insert ?-)) | |
98 (unless (eobp) | |
99 (insert esc) | |
100 (let ((p (point))) | |
101 (skip-chars-forward not-skip-chars) | |
102 (save-restriction | |
103 ;; encode-coding-region doesn't preserve point | |
104 (narrow-to-region p (point)) | |
51630
90215b7be64d
Adjusted for the name change: xxx-utf-16-{le,be} -> xxx-utf-16{le,be}.
Kenichi Handa <handa@m17n.org>
parents:
50897
diff
changeset
|
105 (encode-coding-region p (point-max) 'utf-16be) |
50897 | 106 (base64-encode-region p (point-max)) |
107 (if imap | |
108 (subst-char-in-region p (point-max) ?/ ?,)) | |
109 (goto-char p) | |
110 ;; As I read the RFC, this isn't correct, but it's | |
111 ;; consistent with iconv, at least regarding `='. | |
112 (skip-chars-forward "^= \t\n") | |
113 (delete-region (point) (point-max)))) | |
80235
58e18c83bf75
(utf-7-encode): Never skip the trailing - for the `imap' variant.
Stefan Monnier <monnier@iro.umontreal.ca>
parents:
79709
diff
changeset
|
114 ;; RFC2060 stipulates that all names MUST end in US-ASCII (i.e. |
58e18c83bf75
(utf-7-encode): Never skip the trailing - for the `imap' variant.
Stefan Monnier <monnier@iro.umontreal.ca>
parents:
79709
diff
changeset
|
115 ;; a name that ends with a Unicode octet MUST end with a "-"). |
58e18c83bf75
(utf-7-encode): Never skip the trailing - for the `imap' variant.
Stefan Monnier <monnier@iro.umontreal.ca>
parents:
79709
diff
changeset
|
116 (if (or imap (not (eobp))) |
50897 | 117 (insert ?-))))) |
118 nil)) | |
119 | |
90250
1c8f24770e68
Delete the definition of utf-7.
Kenichi Handa <handa@m17n.org>
parents:
90200
diff
changeset
|
120 ;;;###autoload |
50897 | 121 (defun utf-7-pre-write-conversion (from to) |
122 (utf-7-encode from to nil)) | |
123 | |
91117
57ad20761f87
(utf-7-imap-post-read-conversion)
Kenichi Handa <handa@m17n.org>
parents:
90996
diff
changeset
|
124 ;;;###autoload |
85926
025d69230a9b
(utf-7-imap): New coding system.
Kenichi Handa <handa@m17n.org>
parents:
78274
diff
changeset
|
125 (defun utf-7-imap-pre-write-conversion (from to) |
025d69230a9b
(utf-7-imap): New coding system.
Kenichi Handa <handa@m17n.org>
parents:
78274
diff
changeset
|
126 (utf-7-encode from to t)) |
50897 | 127 |
128 (provide 'utf-7) | |
52401 | 129 |
93975
1e3a407766b9
Fix up comment convention on the arch-tag lines.
Stefan Monnier <monnier@iro.umontreal.ca>
parents:
92360
diff
changeset
|
130 ;; arch-tag: 975ee403-90a4-4286-97d2-4ed1323f4ef9 |
50897 | 131 ;;; utf-7.el ends here |