Fri, 25 Apr 2014 22:07:19 +0200
updated CharDet to 2.2.1, updated changelog
3537
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
1 | ######################## BEGIN LICENSE BLOCK ######################## |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
2 | # The Original Code is Mozilla Universal charset detector code. |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
3 | # |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
4 | # The Initial Developer of the Original Code is |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
5 | # Netscape Communications Corporation. |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
6 | # Portions created by the Initial Developer are Copyright (C) 2001 |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
7 | # the Initial Developer. All Rights Reserved. |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
8 | # |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
9 | # Contributor(s): |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
10 | # Mark Pilgrim - port to Python |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
11 | # Shy Shalom - original C code |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
12 | # Proofpoint, Inc. |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
13 | # |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
14 | # This library is free software; you can redistribute it and/or |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
15 | # modify it under the terms of the GNU Lesser General Public |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
16 | # License as published by the Free Software Foundation; either |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
17 | # version 2.1 of the License, or (at your option) any later version. |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
18 | # |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
19 | # This library is distributed in the hope that it will be useful, |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
20 | # but WITHOUT ANY WARRANTY; without even the implied warranty of |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
21 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
22 | # Lesser General Public License for more details. |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
23 | # |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
24 | # You should have received a copy of the GNU Lesser General Public |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
25 | # License along with this library; if not, write to the Free Software |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
26 | # Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
27 | # 02110-1301 USA |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
28 | ######################### END LICENSE BLOCK ######################### |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
29 | |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
30 | from .charsetgroupprober import CharSetGroupProber |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
31 | from .utf8prober import UTF8Prober |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
32 | from .sjisprober import SJISProber |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
33 | from .eucjpprober import EUCJPProber |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
34 | from .gb2312prober import GB2312Prober |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
35 | from .euckrprober import EUCKRProber |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
36 | from .cp949prober import CP949Prober |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
37 | from .big5prober import Big5Prober |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
38 | from .euctwprober import EUCTWProber |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
39 | |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
40 | |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
41 | class MBCSGroupProber(CharSetGroupProber): |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
42 | def __init__(self): |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
43 | CharSetGroupProber.__init__(self) |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
44 | self._mProbers = [ |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
45 | UTF8Prober(), |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
46 | SJISProber(), |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
47 | EUCJPProber(), |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
48 | GB2312Prober(), |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
49 | EUCKRProber(), |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
50 | CP949Prober(), |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
51 | Big5Prober(), |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
52 | EUCTWProber() |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
53 | ] |
7662053c3906
updated CharDet to 2.2.1, updated changelog
T.Rzepka <Tobias.Rzepka@gmail.com>
parents:
12
diff
changeset
|
54 | self.reset() |