annotate lisp/gnus/utf7.el @ 101794:182f3f4f03c3

*** empty log message ***
author Jason Rumney <jasonr@gnu.org>
date Wed, 04 Feb 2009 09:13:18 +0000
parents a9dc0e7c3f2b
children 88f046f0db6f
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
56927
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
1 ;;; utf7.el --- UTF-7 encoding/decoding for Emacs -*-coding: iso-8859-1;-*-
64754
fafd692d1e40 Update years in copyright notice; nfc.
Thien-Thi Nguyen <ttn@gnuvola.org>
parents: 64085
diff changeset
2
74548
ce127a46b1ca Update copyright years.
Glenn Morris <rgm@gnu.org>
parents: 69135
diff changeset
3 ;; Copyright (C) 1999, 2000, 2001, 2002, 2003, 2004,
100908
a9dc0e7c3f2b Add 2009 to copyright years.
Glenn Morris <rgm@gnu.org>
parents: 94662
diff changeset
4 ;; 2005, 2006, 2007, 2008, 2009 Free Software Foundation, Inc.
31717
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
5
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
6 ;; Author: Jon K Hellan <hellan@acm.org>
56927
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
7 ;; Maintainer: bugs@gnus.org
31717
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
8 ;; Keywords: mail
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
9
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
10 ;; This file is part of GNU Emacs.
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
11
94662
f42ef85caf91 Switch to recommended form of GPLv3 permissions notice.
Glenn Morris <rgm@gnu.org>
parents: 93975
diff changeset
12 ;; GNU Emacs is free software: you can redistribute it and/or modify
31717
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
13 ;; it under the terms of the GNU General Public License as published by
94662
f42ef85caf91 Switch to recommended form of GPLv3 permissions notice.
Glenn Morris <rgm@gnu.org>
parents: 93975
diff changeset
14 ;; the Free Software Foundation, either version 3 of the License, or
f42ef85caf91 Switch to recommended form of GPLv3 permissions notice.
Glenn Morris <rgm@gnu.org>
parents: 93975
diff changeset
15 ;; (at your option) any later version.
31717
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
16
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
17 ;; GNU Emacs is distributed in the hope that it will be useful,
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
18 ;; but WITHOUT ANY WARRANTY; without even the implied warranty of
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
19 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
20 ;; GNU General Public License for more details.
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
21
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
22 ;; You should have received a copy of the GNU General Public License
94662
f42ef85caf91 Switch to recommended form of GPLv3 permissions notice.
Glenn Morris <rgm@gnu.org>
parents: 93975
diff changeset
23 ;; along with GNU Emacs. If not, see <http://www.gnu.org/licenses/>.
31717
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
24
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
25 ;;; Commentary:
56927
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
26
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
27 ;; UTF-7 - A Mail-Safe Transformation Format of Unicode - RFC 2152
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
28 ;; This is a transformation format of Unicode that contains only 7-bit
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
29 ;; ASCII octets and is intended to be readable by humans in the limiting
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
30 ;; case that the document consists of characters from the US-ASCII
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
31 ;; repertoire.
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
32 ;; In short, runs of characters outside US-ASCII are encoded as base64
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
33 ;; inside delimiters.
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
34 ;; A variation of UTF-7 is specified in IMAP 4rev1 (RFC 2060) as the way
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
35 ;; to represent characters outside US-ASCII in mailbox names in IMAP.
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
36 ;; This library supports both variants, but the IMAP variation was the
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
37 ;; reason I wrote it.
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
38 ;; The routines convert UTF-7 -> UTF-16 (16 bit encoding of Unicode)
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
39 ;; -> current character set, and vice versa.
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
40 ;; However, until Emacs supports Unicode, the only Emacs character set
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
41 ;; supported here is ISO-8859.1, which can trivially be converted to/from
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
42 ;; Unicode.
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
43 ;; When decoding results in a character outside the Emacs character set,
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
44 ;; an error is thrown. It is up to the application to recover.
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
45
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
46 ;; UTF-7 should be done by providing a coding system. Mule-UCS does
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
47 ;; already, but I don't know if it does the IMAP version and it's not
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
48 ;; clear whether that should really be a coding system. The UTF-16
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
49 ;; part of the conversion can be done with coding systems available
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
50 ;; with Mule-UCS or some versions of Emacs. Unfortunately these were
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
51 ;; done wrongly (regarding handling of byte-order marks and how the
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
52 ;; variants were named), so we don't have a consistent name for the
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
53 ;; necessary coding system. The code below doesn't seem to DTRT
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
54 ;; generally. E.g.:
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
55 ;;
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
56 ;; (utf7-encode "a+£")
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
57 ;; => "a+ACsAow-"
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
58 ;;
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
59 ;; $ echo "a+£"|iconv -f iso-8859-1 -t utf-7
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
60 ;; a+-+AKM
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
61 ;;
59996
aac0a33f5772 Change release version from 21.4 to 22.1 throughout.
Kim F. Storm <storm@cua.dk>
parents: 56927
diff changeset
62 ;; -- fx
56927
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
63
31717
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
64
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
65 ;;; Code:
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
66
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
67 (require 'base64)
33119
e08be2ed6301 Require cl when compiling.
Dave Love <fx@gnu.org>
parents: 31717
diff changeset
68 (eval-when-compile (require 'cl))
56927
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
69 (require 'mm-util)
31717
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
70
56927
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
71 (defconst utf7-direct-encoding-chars " -%'-*,-[]-}"
31717
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
72 "Character ranges which do not need escaping in UTF-7.")
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
73
56927
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
74 (defconst utf7-imap-direct-encoding-chars
31717
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
75 (concat utf7-direct-encoding-chars "+\\~")
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
76 "Character ranges which do not need escaping in the IMAP variant of UTF-7.")
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
77
56927
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
78 (defconst utf7-utf-16-coding-system
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
79 (cond ((mm-coding-system-p 'utf-16-be-no-signature) ; Mule-UCS
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
80 'utf-16-be-no-signature)
69135
12d289b6e5de Revision: emacs@sv.gnu.org/emacs--devo--0--patch-118
Miles Bader <miles@gnu.org>
parents: 68633
diff changeset
81 ((and (mm-coding-system-p 'utf-16-be) ; Emacs 21.3, Emacs 22
56927
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
82 ;; Avoid versions with BOM.
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
83 (= 2 (length (encode-coding-string "a" 'utf-16-be))))
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
84 'utf-16-be)
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
85 ((mm-coding-system-p 'utf-16-be-nosig) ; ?
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
86 'utf-16-be-nosig))
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
87 "Coding system which encodes big endian UTF-16 without a BOM signature.")
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
88
31717
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
89 (defsubst utf7-imap-get-pad-length (len modulus)
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
90 "Return required length of padding for IMAP modified base64 fragment."
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
91 (mod (- len) modulus))
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
92
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
93 (defun utf7-encode-internal (&optional for-imap)
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
94 "Encode text in (temporary) buffer as UTF-7.
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
95 Use IMAP modification if FOR-IMAP is non-nil."
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
96 (let ((start (point-min))
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
97 (end (point-max)))
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
98 (narrow-to-region start end)
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
99 (goto-char start)
56927
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
100 (let* ((esc-char (if for-imap ?& ?+))
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
101 (direct-encoding-chars
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
102 (if for-imap utf7-imap-direct-encoding-chars
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
103 utf7-direct-encoding-chars))
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
104 (not-direct-encoding-chars (concat "^" direct-encoding-chars)))
31717
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
105 (while (not (eobp))
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
106 (skip-chars-forward direct-encoding-chars)
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
107 (unless (eobp)
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
108 (insert esc-char)
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
109 (let ((p (point))
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
110 (fc (following-char))
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
111 (run-length
56927
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
112 (skip-chars-forward not-direct-encoding-chars)))
31717
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
113 (if (and (= fc esc-char)
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
114 (= run-length 1)) ; Lone esc-char?
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
115 (delete-backward-char 1) ; Now there's one too many
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
116 (utf7-fragment-encode p (point) for-imap))
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
117 (insert "-")))))))
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
118
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
119 (defun utf7-fragment-encode (start end &optional for-imap)
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
120 "Encode text from START to END in buffer as UTF-7 escape fragment.
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
121 Use IMAP modification if FOR-IMAP is non-nil."
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
122 (save-restriction
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
123 (narrow-to-region start end)
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
124 (funcall (utf7-get-u16char-converter 'to-utf-16))
56927
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
125 (mm-with-unibyte-current-buffer
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
126 (base64-encode-region start (point-max)))
31717
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
127 (goto-char start)
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
128 (let ((pm (point-max)))
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
129 (when for-imap
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
130 (while (search-forward "/" nil t)
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
131 (replace-match ",")))
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
132 (skip-chars-forward "^= \t\n" pm)
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
133 (delete-region (point) pm))))
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
134
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
135 (defun utf7-decode-internal (&optional for-imap)
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
136 "Decode UTF-7 text in (temporary) buffer.
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
137 Use IMAP modification if FOR-IMAP is non-nil."
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
138 (let ((start (point-min))
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
139 (end (point-max)))
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
140 (goto-char start)
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
141 (let* ((esc-pattern (concat "^" (char-to-string (if for-imap ?& ?+))))
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
142 (base64-chars (concat "A-Za-z0-9+"
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
143 (char-to-string (if for-imap ?, ?/)))))
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
144 (while (not (eobp))
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
145 (skip-chars-forward esc-pattern)
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
146 (unless (eobp)
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
147 (forward-char)
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
148 (let ((p (point))
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
149 (run-length (skip-chars-forward base64-chars)))
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
150 (when (and (not (eobp)) (= (following-char) ?-))
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
151 (delete-char 1))
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
152 (unless (= run-length 0) ; Encoded lone esc-char?
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
153 (save-excursion
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
154 (utf7-fragment-decode p (point) for-imap)
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
155 (goto-char p)
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
156 (delete-backward-char 1)))))))))
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
157
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
158 (defun utf7-fragment-decode (start end &optional for-imap)
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
159 "Decode base64 encoded fragment from START to END of UTF-7 text in buffer.
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
160 Use IMAP modification if FOR-IMAP is non-nil."
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
161 (save-restriction
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
162 (narrow-to-region start end)
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
163 (when for-imap
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
164 (goto-char start)
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
165 (while (search-forward "," nil 'move-to-end) (replace-match "/")))
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
166 (let ((pl (utf7-imap-get-pad-length (- end start) 4)))
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
167 (insert (make-string pl ?=))
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
168 (base64-decode-region start (+ end pl)))
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
169 (funcall (utf7-get-u16char-converter 'from-utf-16))))
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
170
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
171 (defun utf7-get-u16char-converter (which-way)
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
172 "Return a function to convert between UTF-16 and current character set."
56927
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
173 (if utf7-utf-16-coding-system
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
174 (if (eq which-way 'to-utf-16)
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
175 (lambda ()
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
176 (encode-coding-region (point-min) (point-max)
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
177 utf7-utf-16-coding-system))
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
178 (lambda ()
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
179 (decode-coding-region (point-min) (point-max)
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
180 utf7-utf-16-coding-system)))
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
181 ;; Add test to check if we are really Latin-1.
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
182 (if (eq which-way 'to-utf-16)
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
183 'utf7-latin1-u16-char-converter
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
184 'utf7-u16-latin1-char-converter)))
31717
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
185
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
186 (defun utf7-latin1-u16-char-converter ()
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
187 "Convert latin 1 (ISO-8859.1) characters to 16 bit Unicode.
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
188 Characters are converted to raw byte pairs in narrowed buffer."
56927
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
189 (mm-encode-coding-region (point-min) (point-max) 'iso-8859-1)
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
190 (mm-disable-multibyte)
31717
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
191 (goto-char (point-min))
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
192 (while (not (eobp))
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
193 (insert 0)
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
194 (forward-char)))
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
195
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
196 (defun utf7-u16-latin1-char-converter ()
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
197 "Convert 16 bit Unicode characters to latin 1 (ISO-8859.1).
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
198 Characters are in raw byte pairs in narrowed buffer."
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
199 (goto-char (point-min))
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
200 (while (not (eobp))
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
201 (if (= 0 (following-char))
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
202 (delete-char 1)
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
203 (error "Unable to convert from Unicode"))
56927
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
204 (forward-char))
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
205 (mm-decode-coding-region (point-min) (point-max) 'iso-8859-1)
55fd4f77387a Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-523
Miles Bader <miles@gnu.org>
parents: 52401
diff changeset
206 (mm-enable-multibyte))
31717
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
207
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
208 (defun utf7-encode (string &optional for-imap)
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
209 "Encode UTF-7 STRING. Use IMAP modification if FOR-IMAP is non-nil."
86255
50442c77e242 (utf7-encode, utf7-decode): Use coding system
Reiner Steib <Reiner.Steib@gmx.de>
parents: 78224
diff changeset
210 (if (and (coding-system-p 'utf-7) (coding-system-p 'utf-7-imap))
50442c77e242 (utf7-encode, utf7-decode): Use coding system
Reiner Steib <Reiner.Steib@gmx.de>
parents: 78224
diff changeset
211 ;; Emacs 23 with proper support for IMAP
50442c77e242 (utf7-encode, utf7-decode): Use coding system
Reiner Steib <Reiner.Steib@gmx.de>
parents: 78224
diff changeset
212 (encode-coding-string string (if for-imap 'utf-7-imap 'utf-7))
50442c77e242 (utf7-encode, utf7-decode): Use coding system
Reiner Steib <Reiner.Steib@gmx.de>
parents: 78224
diff changeset
213 (let ((default-enable-multibyte-characters t))
50442c77e242 (utf7-encode, utf7-decode): Use coding system
Reiner Steib <Reiner.Steib@gmx.de>
parents: 78224
diff changeset
214 (with-temp-buffer
50442c77e242 (utf7-encode, utf7-decode): Use coding system
Reiner Steib <Reiner.Steib@gmx.de>
parents: 78224
diff changeset
215 (insert string)
50442c77e242 (utf7-encode, utf7-decode): Use coding system
Reiner Steib <Reiner.Steib@gmx.de>
parents: 78224
diff changeset
216 (utf7-encode-internal for-imap)
50442c77e242 (utf7-encode, utf7-decode): Use coding system
Reiner Steib <Reiner.Steib@gmx.de>
parents: 78224
diff changeset
217 (buffer-string)))))
31717
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
218
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
219 (defun utf7-decode (string &optional for-imap)
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
220 "Decode UTF-7 STRING. Use IMAP modification if FOR-IMAP is non-nil."
86255
50442c77e242 (utf7-encode, utf7-decode): Use coding system
Reiner Steib <Reiner.Steib@gmx.de>
parents: 78224
diff changeset
221 (if (and (coding-system-p 'utf-7) (coding-system-p 'utf-7-imap))
50442c77e242 (utf7-encode, utf7-decode): Use coding system
Reiner Steib <Reiner.Steib@gmx.de>
parents: 78224
diff changeset
222 ;; Emacs 23 with proper support for IMAP
50442c77e242 (utf7-encode, utf7-decode): Use coding system
Reiner Steib <Reiner.Steib@gmx.de>
parents: 78224
diff changeset
223 (decode-coding-string string (if for-imap 'utf-7-imap 'utf-7))
50442c77e242 (utf7-encode, utf7-decode): Use coding system
Reiner Steib <Reiner.Steib@gmx.de>
parents: 78224
diff changeset
224 (let ((default-enable-multibyte-characters nil))
50442c77e242 (utf7-encode, utf7-decode): Use coding system
Reiner Steib <Reiner.Steib@gmx.de>
parents: 78224
diff changeset
225 (with-temp-buffer
50442c77e242 (utf7-encode, utf7-decode): Use coding system
Reiner Steib <Reiner.Steib@gmx.de>
parents: 78224
diff changeset
226 (insert string)
50442c77e242 (utf7-encode, utf7-decode): Use coding system
Reiner Steib <Reiner.Steib@gmx.de>
parents: 78224
diff changeset
227 (utf7-decode-internal for-imap)
50442c77e242 (utf7-encode, utf7-decode): Use coding system
Reiner Steib <Reiner.Steib@gmx.de>
parents: 78224
diff changeset
228 (mm-enable-multibyte)
50442c77e242 (utf7-encode, utf7-decode): Use coding system
Reiner Steib <Reiner.Steib@gmx.de>
parents: 78224
diff changeset
229 (buffer-string)))))
31717
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
230
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
231 (provide 'utf7)
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
232
93975
1e3a407766b9 Fix up comment convention on the arch-tag lines.
Stefan Monnier <monnier@iro.umontreal.ca>
parents: 87649
diff changeset
233 ;; arch-tag: 96078b55-85c7-4161-aed2-932c24b282c7
31717
6b20b7e85e3c *** empty log message ***
Gerd Moellmann <gerd@gnu.org>
parents:
diff changeset
234 ;;; utf7.el ends here