annotate lisp/textmodes/po.el @ 72307:725af3953ee0

*** empty log message ***
author Andreas Schwab <schwab@suse.de>
date Mon, 07 Aug 2006 16:39:36 +0000
parents 51964b589c64
children e3694f1cb928 858cb33ae39d
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
43926
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
1 ;;; po.el --- basic support of PO translation files -*- coding: latin-1; -*-
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
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,
68648
067115a6e738 Update years in copyright notice; nfc.
Thien-Thi Nguyen <ttn@gnuvola.org>
parents: 64751
diff changeset
4 ;; 2004, 2005, 2006 Free Software Foundation, Inc.
43926
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
5
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
6 ;; Authors: François Pinard <pinard@iro.umontreal.ca>,
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
7 ;; Greg McGary <gkm@magilla.cichlid.com>,
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
8 ;; Bruno Haible <bruno@clisp.org>.
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
9 ;; Keywords: i18n, files
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
10
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
11 ;; This file is part of GNU Emacs.
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
12
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
13 ;; GNU Emacs is free software; you can redistribute it and/or modify
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
14 ;; it under the terms of the GNU General Public License as published by
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
15 ;; the Free Software Foundation; either version 2, or (at your option)
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
16 ;; any later version.
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
17
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
18 ;; GNU Emacs is distributed in the hope that it will be useful,
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
19 ;; but WITHOUT ANY WARRANTY; without even the implied warranty of
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
20 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
21 ;; GNU General Public License for more details.
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
22
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
23 ;; You should have received a copy of the GNU General Public License
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
24 ;; along with GNU Emacs; see the file COPYING. If not, write to the
64084
a8fa7c632ee4 Update FSF's address.
Lute Kamstra <lute@gnu.org>
parents: 62120
diff changeset
25 ;; Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
a8fa7c632ee4 Update FSF's address.
Lute Kamstra <lute@gnu.org>
parents: 62120
diff changeset
26 ;; Boston, MA 02110-1301, USA.
43926
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
27
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
28 ;;; Commentary:
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
29
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
30 ;; This package makes sure visiting PO files decodes them correctly,
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
31 ;; according to the Charset= header in the PO file. For more support
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
32 ;; for editing PO files, see po-mode.el.
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
33
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
34 ;;; Code:
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
35
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
36 (defconst po-content-type-charset-alist
48092
10d4bf044393 (po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents: 47655
diff changeset
37 '(("ASCII" . undecided)
43958
2479ec7d435b (po-content-type-charset-alist): Convert the
Eli Zaretskii <eliz@gnu.org>
parents: 43954
diff changeset
38 ("ANSI_X3.4-1968" . undecided)
48092
10d4bf044393 (po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents: 47655
diff changeset
39 ("US-ASCII" . undecided))
10d4bf044393 (po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents: 47655
diff changeset
40 "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
41 Contains canonical charset names that don't correspond to coding systems.")
43926
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
42
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
43 (defun po-find-charset (filename)
70944
2e73c646b9df (po-find-charset): Pay attention to the case
Kenichi Handa <handa@m17n.org>
parents: 68648
diff changeset
44 "Return PO charset value for FILENAME.
72053
51964b589c64 (po-find-charset): Doc fix.
Eli Zaretskii <eliz@gnu.org>
parents: 70944
diff changeset
45 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
46 the PO file (but not yet decoded)."
43926
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
47 (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
48 "^\"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
49 (buf (and (consp filename) (cdr filename)))
43926
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
50 (short-read nil))
70944
2e73c646b9df (po-find-charset): Pay attention to the case
Kenichi Handa <handa@m17n.org>
parents: 68648
diff changeset
51 (when buf
2e73c646b9df (po-find-charset): Pay attention to the case
Kenichi Handa <handa@m17n.org>
parents: 68648
diff changeset
52 (set-buffer buf)
2e73c646b9df (po-find-charset): Pay attention to the case
Kenichi Handa <handa@m17n.org>
parents: 68648
diff changeset
53 (goto-char (point-min)))
43926
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
54 ;; Try the first 4096 bytes. In case we cannot find the charset value
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
55 ;; within the first 4096 bytes (the PO file might start with a long
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
56 ;; comment) try the next 4096 bytes repeatedly until we'll know for sure
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
57 ;; 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
58 (while (not (or short-read (re-search-forward "^msgid" nil t) buf))
43926
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
59 (save-excursion
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
60 (goto-char (point-max))
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
61 (let ((pair (insert-file-contents-literally filename nil
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
62 (1- (point))
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
63 (1- (+ (point) 4096)))))
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
64 (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
65 (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
66 ((or short-read buf) nil)
43926
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
67 ;; We've found the first msgid; maybe, only a part of the msgstr
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
68 ;; value was loaded. Load the next 1024 bytes; if charset still
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
69 ;; isn't available, give up.
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
70 (t (save-excursion
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
71 (goto-char (point-max))
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
72 (insert-file-contents-literally filename nil
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
73 (1- (point))
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
74 (1- (+ (point) 1024))))
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
75 (if (re-search-forward charset-regexp nil t)
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
76 (match-string 1))))))
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
77
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
78 (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
79 "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
80 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
81 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
82 in BUFFER."
48092
10d4bf044393 (po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents: 47655
diff changeset
83 (and
10d4bf044393 (po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents: 47655
diff changeset
84 (eq operation 'insert-file-contents)
70944
2e73c646b9df (po-find-charset): Pay attention to the case
Kenichi Handa <handa@m17n.org>
parents: 68648
diff changeset
85 (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
86 (file-exists-p filename))
48092
10d4bf044393 (po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents: 47655
diff changeset
87 (with-temp-buffer
10d4bf044393 (po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents: 47655
diff changeset
88 (let* ((coding-system-for-read 'no-conversion)
10d4bf044393 (po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents: 47655
diff changeset
89 (charset (or (po-find-charset filename) "ascii"))
10d4bf044393 (po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents: 47655
diff changeset
90 assoc)
10d4bf044393 (po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents: 47655
diff changeset
91 (list (cond
10d4bf044393 (po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents: 47655
diff changeset
92 ((setq assoc
62120
912c2cf79005 (po-find-file-coding-system-guts): Replace `assoc-ignore-case' by
Juanma Barranquero <lekktu@gmail.com>
parents: 52401
diff changeset
93 (assoc-string charset
912c2cf79005 (po-find-file-coding-system-guts): Replace `assoc-ignore-case' by
Juanma Barranquero <lekktu@gmail.com>
parents: 52401
diff changeset
94 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
95 t))
48092
10d4bf044393 (po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents: 47655
diff changeset
96 (cdr assoc))
62120
912c2cf79005 (po-find-file-coding-system-guts): Replace `assoc-ignore-case' by
Juanma Barranquero <lekktu@gmail.com>
parents: 52401
diff changeset
97 ((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
98 (setq assoc
62120
912c2cf79005 (po-find-file-coding-system-guts): Replace `assoc-ignore-case' by
Juanma Barranquero <lekktu@gmail.com>
parents: 52401
diff changeset
99 (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
100 charset)
912c2cf79005 (po-find-file-coding-system-guts): Replace `assoc-ignore-case' by
Juanma Barranquero <lekktu@gmail.com>
parents: 52401
diff changeset
101 coding-system-alist t)))
48092
10d4bf044393 (po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents: 47655
diff changeset
102 (intern (car assoc)))
10d4bf044393 (po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents: 47655
diff changeset
103 ;; 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
104 ;; 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
105 ;; 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
106 ;; corresponding to the MIME name.
10d4bf044393 (po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents: 47655
diff changeset
107 ((featurep 'code-pages)
10d4bf044393 (po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents: 47655
diff changeset
108 ;; Give up.
10d4bf044393 (po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents: 47655
diff changeset
109 'raw-text)
10d4bf044393 (po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents: 47655
diff changeset
110 (t
10d4bf044393 (po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents: 47655
diff changeset
111 ;; 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
112 ;; to require it initially?
10d4bf044393 (po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents: 47655
diff changeset
113 (require 'code-pages nil t)
10d4bf044393 (po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents: 47655
diff changeset
114 (if (or
62120
912c2cf79005 (po-find-file-coding-system-guts): Replace `assoc-ignore-case' by
Juanma Barranquero <lekktu@gmail.com>
parents: 52401
diff changeset
115 (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
116 (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
117 ?_ ?- charset)
912c2cf79005 (po-find-file-coding-system-guts): Replace `assoc-ignore-case' by
Juanma Barranquero <lekktu@gmail.com>
parents: 52401
diff changeset
118 coding-system-alist t)))
48092
10d4bf044393 (po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents: 47655
diff changeset
119 (intern (car assoc))
10d4bf044393 (po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents: 47655
diff changeset
120 'raw-text))))))))
43926
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
121
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
122 ;;;###autoload
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
123 (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
124 "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
125 Called through `file-coding-system-alist', before the file is visited for real."
43926
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
126 (po-find-file-coding-system-guts (car arg-list) (car (cdr arg-list))))
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
127 ;; This is for XEmacs.
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
128 ;(defun po-find-file-coding-system (operation filename)
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
129 ; "\
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
130 ;Return a Mule (DECODING . ENCODING) pair, according to PO file charset.
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
131 ;Called through file-coding-system-alist, before the file is visited for real."
11fdbf69b362 New file.
Eli Zaretskii <eliz@gnu.org>
parents:
diff changeset
132 ; (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
133
10d4bf044393 (po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents: 47655
diff changeset
134 (provide 'po)
10d4bf044393 (po-content-type-charset-alist): Delete most
Dave Love <fx@gnu.org>
parents: 47655
diff changeset
135
52401
695cf19ef79e Add arch taglines
Miles Bader <miles@gnu.org>
parents: 51890
diff changeset
136 ;;; 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
137 ;;; po.el ends here