Mercurial > emacs
annotate lisp/textmodes/po.el @ 103855:6479d8a2ebcf
(describe-character-set): Don't show width.
author | Kenichi Handa <handa@m17n.org> |
---|---|
date | Sat, 11 Jul 2009 06:36:14 +0000 |
parents | a9dc0e7c3f2b |
children | 1d1d5d9bd884 |
rev | line source |
---|---|
43926 | 1 ;;; po.el --- basic support of PO translation files -*- coding: latin-1; -*- |
2 | |
64751
5b1a238fcbb4
Update years in copyright notice; nfc.
Thien-Thi Nguyen <ttn@gnuvola.org>
parents:
64084
diff
changeset
|
3 ;; Copyright (C) 1995, 1996, 1997, 1998, 2000, 2001, 2002, 2003, |
100908 | 4 ;; 2004, 2005, 2006, 2007, 2008, 2009 Free Software Foundation, Inc. |
43926 | 5 |
6 ;; Authors: François Pinard <pinard@iro.umontreal.ca>, | |
7 ;; Greg McGary <gkm@magilla.cichlid.com>, | |
8 ;; Bruno Haible <bruno@clisp.org>. | |
9 ;; Keywords: i18n, files | |
10 | |
11 ;; This file is part of GNU Emacs. | |
12 | |
94670
f4a69fedbd46
Switch to recommended form of GPLv3 permissions notice.
Glenn Morris <rgm@gnu.org>
parents:
93975
diff
changeset
|
13 ;; GNU Emacs is free software: you can redistribute it and/or modify |
43926 | 14 ;; it under the terms of the GNU General Public License as published by |
94670
f4a69fedbd46
Switch to recommended form of GPLv3 permissions notice.
Glenn Morris <rgm@gnu.org>
parents:
93975
diff
changeset
|
15 ;; the Free Software Foundation, either version 3 of the License, or |
f4a69fedbd46
Switch to recommended form of GPLv3 permissions notice.
Glenn Morris <rgm@gnu.org>
parents:
93975
diff
changeset
|
16 ;; (at your option) any later version. |
43926 | 17 |
18 ;; GNU Emacs is distributed in the hope that it will be useful, | |
19 ;; but WITHOUT ANY WARRANTY; without even the implied warranty of | |
20 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
21 ;; GNU General Public License for more details. | |
22 | |
23 ;; You should have received a copy of the GNU General Public License | |
94670
f4a69fedbd46
Switch to recommended form of GPLv3 permissions notice.
Glenn Morris <rgm@gnu.org>
parents:
93975
diff
changeset
|
24 ;; along with GNU Emacs. If not, see <http://www.gnu.org/licenses/>. |
43926 | 25 |
26 ;;; Commentary: | |
27 | |
28 ;; This package makes sure visiting PO files decodes them correctly, | |
29 ;; according to the Charset= header in the PO file. For more support | |
30 ;; for editing PO files, see po-mode.el. | |
31 | |
32 ;;; Code: | |
33 | |
34 (defconst po-content-type-charset-alist | |
48092
10d4bf044393
(po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents:
47655
diff
changeset
|
35 '(("ASCII" . undecided) |
43958
2479ec7d435b
(po-content-type-charset-alist): Convert the
Eli Zaretskii <eliz@gnu.org>
parents:
43954
diff
changeset
|
36 ("ANSI_X3.4-1968" . undecided) |
48092
10d4bf044393
(po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents:
47655
diff
changeset
|
37 ("US-ASCII" . undecided)) |
10d4bf044393
(po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents:
47655
diff
changeset
|
38 "Alist of coding system versus GNU libc/libiconv canonical charset name. |
10d4bf044393
(po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents:
47655
diff
changeset
|
39 Contains canonical charset names that don't correspond to coding systems.") |
43926 | 40 |
41 (defun po-find-charset (filename) | |
70944
2e73c646b9df
(po-find-charset): Pay attention to the case
Kenichi Handa <handa@m17n.org>
parents:
68648
diff
changeset
|
42 "Return PO charset value for FILENAME. |
72053 | 43 If FILENAME is a cons cell, its CDR is a buffer that already contains |
70944
2e73c646b9df
(po-find-charset): Pay attention to the case
Kenichi Handa <handa@m17n.org>
parents:
68648
diff
changeset
|
44 the PO file (but not yet decoded)." |
43926 | 45 (let ((charset-regexp |
51890
9b78d0da1a28
(po-find-charset): White space at the start of the Content-Type field body is
Juanma Barranquero <lekktu@gmail.com>
parents:
48092
diff
changeset
|
46 "^\"Content-Type:[ \t]*text/plain;[ \t]*charset=\\(.*\\)\\\\n\"") |
70944
2e73c646b9df
(po-find-charset): Pay attention to the case
Kenichi Handa <handa@m17n.org>
parents:
68648
diff
changeset
|
47 (buf (and (consp filename) (cdr filename))) |
43926 | 48 (short-read nil)) |
70944
2e73c646b9df
(po-find-charset): Pay attention to the case
Kenichi Handa <handa@m17n.org>
parents:
68648
diff
changeset
|
49 (when buf |
2e73c646b9df
(po-find-charset): Pay attention to the case
Kenichi Handa <handa@m17n.org>
parents:
68648
diff
changeset
|
50 (set-buffer buf) |
2e73c646b9df
(po-find-charset): Pay attention to the case
Kenichi Handa <handa@m17n.org>
parents:
68648
diff
changeset
|
51 (goto-char (point-min))) |
43926 | 52 ;; Try the first 4096 bytes. In case we cannot find the charset value |
53 ;; within the first 4096 bytes (the PO file might start with a long | |
54 ;; comment) try the next 4096 bytes repeatedly until we'll know for sure | |
55 ;; we've checked the empty header entry entirely. | |
70944
2e73c646b9df
(po-find-charset): Pay attention to the case
Kenichi Handa <handa@m17n.org>
parents:
68648
diff
changeset
|
56 (while (not (or short-read (re-search-forward "^msgid" nil t) buf)) |
43926 | 57 (save-excursion |
58 (goto-char (point-max)) | |
59 (let ((pair (insert-file-contents-literally filename nil | |
60 (1- (point)) | |
61 (1- (+ (point) 4096))))) | |
62 (setq short-read (< (nth 1 pair) 4096))))) | |
43952
bd36495e6ade
(po-find-charset): Search for Charset= header even if we've read less than
Eli Zaretskii <eliz@gnu.org>
parents:
43937
diff
changeset
|
63 (cond ((re-search-forward charset-regexp nil t) (match-string 1)) |
70944
2e73c646b9df
(po-find-charset): Pay attention to the case
Kenichi Handa <handa@m17n.org>
parents:
68648
diff
changeset
|
64 ((or short-read buf) nil) |
43926 | 65 ;; We've found the first msgid; maybe, only a part of the msgstr |
66 ;; value was loaded. Load the next 1024 bytes; if charset still | |
67 ;; isn't available, give up. | |
68 (t (save-excursion | |
69 (goto-char (point-max)) | |
70 (insert-file-contents-literally filename nil | |
71 (1- (point)) | |
72 (1- (+ (point) 1024)))) | |
73 (if (re-search-forward charset-regexp nil t) | |
74 (match-string 1)))))) | |
75 | |
76 (defun po-find-file-coding-system-guts (operation filename) | |
48092
10d4bf044393
(po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents:
47655
diff
changeset
|
77 "Return a (DECODING . ENCODING) pair for OPERATION on PO file FILENAME. |
70944
2e73c646b9df
(po-find-charset): Pay attention to the case
Kenichi Handa <handa@m17n.org>
parents:
68648
diff
changeset
|
78 Do so according to FILENAME's declared charset. |
2e73c646b9df
(po-find-charset): Pay attention to the case
Kenichi Handa <handa@m17n.org>
parents:
68648
diff
changeset
|
79 FILENAME may be a cons (NAME . BUFFER). In that case, detect charset |
2e73c646b9df
(po-find-charset): Pay attention to the case
Kenichi Handa <handa@m17n.org>
parents:
68648
diff
changeset
|
80 in BUFFER." |
48092
10d4bf044393
(po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents:
47655
diff
changeset
|
81 (and |
10d4bf044393
(po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents:
47655
diff
changeset
|
82 (eq operation 'insert-file-contents) |
70944
2e73c646b9df
(po-find-charset): Pay attention to the case
Kenichi Handa <handa@m17n.org>
parents:
68648
diff
changeset
|
83 (or (if (consp filename) (buffer-live-p (cdr filename))) |
2e73c646b9df
(po-find-charset): Pay attention to the case
Kenichi Handa <handa@m17n.org>
parents:
68648
diff
changeset
|
84 (file-exists-p filename)) |
48092
10d4bf044393
(po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents:
47655
diff
changeset
|
85 (with-temp-buffer |
10d4bf044393
(po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents:
47655
diff
changeset
|
86 (let* ((coding-system-for-read 'no-conversion) |
10d4bf044393
(po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents:
47655
diff
changeset
|
87 (charset (or (po-find-charset filename) "ascii")) |
10d4bf044393
(po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents:
47655
diff
changeset
|
88 assoc) |
10d4bf044393
(po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents:
47655
diff
changeset
|
89 (list (cond |
10d4bf044393
(po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents:
47655
diff
changeset
|
90 ((setq assoc |
62120
912c2cf79005
(po-find-file-coding-system-guts): Replace `assoc-ignore-case' by
Juanma Barranquero <lekktu@gmail.com>
parents:
52401
diff
changeset
|
91 (assoc-string charset |
912c2cf79005
(po-find-file-coding-system-guts): Replace `assoc-ignore-case' by
Juanma Barranquero <lekktu@gmail.com>
parents:
52401
diff
changeset
|
92 po-content-type-charset-alist |
912c2cf79005
(po-find-file-coding-system-guts): Replace `assoc-ignore-case' by
Juanma Barranquero <lekktu@gmail.com>
parents:
52401
diff
changeset
|
93 t)) |
48092
10d4bf044393
(po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents:
47655
diff
changeset
|
94 (cdr assoc)) |
62120
912c2cf79005
(po-find-file-coding-system-guts): Replace `assoc-ignore-case' by
Juanma Barranquero <lekktu@gmail.com>
parents:
52401
diff
changeset
|
95 ((or (setq assoc (assoc-string charset coding-system-alist t)) |
48092
10d4bf044393
(po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents:
47655
diff
changeset
|
96 (setq assoc |
62120
912c2cf79005
(po-find-file-coding-system-guts): Replace `assoc-ignore-case' by
Juanma Barranquero <lekktu@gmail.com>
parents:
52401
diff
changeset
|
97 (assoc-string (subst-char-in-string ?_ ?- |
912c2cf79005
(po-find-file-coding-system-guts): Replace `assoc-ignore-case' by
Juanma Barranquero <lekktu@gmail.com>
parents:
52401
diff
changeset
|
98 charset) |
912c2cf79005
(po-find-file-coding-system-guts): Replace `assoc-ignore-case' by
Juanma Barranquero <lekktu@gmail.com>
parents:
52401
diff
changeset
|
99 coding-system-alist t))) |
48092
10d4bf044393
(po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents:
47655
diff
changeset
|
100 (intern (car assoc))) |
10d4bf044393
(po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents:
47655
diff
changeset
|
101 ;; In principle we should also check the `mime-charset' |
10d4bf044393
(po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents:
47655
diff
changeset
|
102 ;; property of everything in the base coding system |
10d4bf044393
(po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents:
47655
diff
changeset
|
103 ;; list, but there should always be a coding system |
10d4bf044393
(po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents:
47655
diff
changeset
|
104 ;; corresponding to the MIME name. |
10d4bf044393
(po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents:
47655
diff
changeset
|
105 ((featurep 'code-pages) |
10d4bf044393
(po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents:
47655
diff
changeset
|
106 ;; Give up. |
10d4bf044393
(po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents:
47655
diff
changeset
|
107 'raw-text) |
10d4bf044393
(po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents:
47655
diff
changeset
|
108 (t |
10d4bf044393
(po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents:
47655
diff
changeset
|
109 ;; Try again with code-pages loaded. Maybe it's best |
10d4bf044393
(po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents:
47655
diff
changeset
|
110 ;; to require it initially? |
10d4bf044393
(po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents:
47655
diff
changeset
|
111 (require 'code-pages nil t) |
10d4bf044393
(po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents:
47655
diff
changeset
|
112 (if (or |
62120
912c2cf79005
(po-find-file-coding-system-guts): Replace `assoc-ignore-case' by
Juanma Barranquero <lekktu@gmail.com>
parents:
52401
diff
changeset
|
113 (setq assoc (assoc-string charset coding-system-alist t)) |
912c2cf79005
(po-find-file-coding-system-guts): Replace `assoc-ignore-case' by
Juanma Barranquero <lekktu@gmail.com>
parents:
52401
diff
changeset
|
114 (setq assoc (assoc-string (subst-char-in-string |
912c2cf79005
(po-find-file-coding-system-guts): Replace `assoc-ignore-case' by
Juanma Barranquero <lekktu@gmail.com>
parents:
52401
diff
changeset
|
115 ?_ ?- charset) |
912c2cf79005
(po-find-file-coding-system-guts): Replace `assoc-ignore-case' by
Juanma Barranquero <lekktu@gmail.com>
parents:
52401
diff
changeset
|
116 coding-system-alist t))) |
48092
10d4bf044393
(po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents:
47655
diff
changeset
|
117 (intern (car assoc)) |
10d4bf044393
(po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents:
47655
diff
changeset
|
118 'raw-text)))))))) |
43926 | 119 |
120 ;;;###autoload | |
121 (defun po-find-file-coding-system (arg-list) | |
48092
10d4bf044393
(po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents:
47655
diff
changeset
|
122 "Return a (DECODING . ENCODING) pair, according to PO file's charset. |
10d4bf044393
(po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents:
47655
diff
changeset
|
123 Called through `file-coding-system-alist', before the file is visited for real." |
43926 | 124 (po-find-file-coding-system-guts (car arg-list) (car (cdr arg-list)))) |
125 ;; This is for XEmacs. | |
126 ;(defun po-find-file-coding-system (operation filename) | |
127 ; "\ | |
128 ;Return a Mule (DECODING . ENCODING) pair, according to PO file charset. | |
129 ;Called through file-coding-system-alist, before the file is visited for real." | |
130 ; (po-find-file-coding-system-guts operation filename)) | |
48092
10d4bf044393
(po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents:
47655
diff
changeset
|
131 |
10d4bf044393
(po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents:
47655
diff
changeset
|
132 (provide 'po) |
10d4bf044393
(po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents:
47655
diff
changeset
|
133 |
93975
1e3a407766b9
Fix up comment convention on the arch-tag lines.
Stefan Monnier <monnier@iro.umontreal.ca>
parents:
79719
diff
changeset
|
134 ;; arch-tag: 56748a57-d64c-4200-8f6b-c3a70496eb8c |
48092
10d4bf044393
(po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents:
47655
diff
changeset
|
135 ;;; po.el ends here |