comparison Resources/EncodingTests.py @ 1091:a66224eec125

encoding tests
author Sebastien Jodogne <s.jodogne@gmail.com>
date Tue, 05 Aug 2014 12:28:55 +0200
parents e494ceb8d763
children 60cc0ee61edb
comparison
equal deleted inserted replaced
1090:e494ceb8d763 1091:a66224eec125
1 #!/usr/bin/python 1 #!/usr/bin/python
2 # -*- coding: utf-8 -*- 2 # -*- coding: utf-8 -*-
3 3
4 source = u'éäöòДΘĝדصķћ๛ネİ' 4 source = u'TestéäöòДΘĝדصķћ๛ネİ'
5 5
6 encodings = { 6 encodings = {
7 'UTF-8' : 'Utf8', 7 'UTF-8' : 'Utf8',
8 'ASCII' : 'Ascii', 8 'ASCII' : 'Ascii',
9 'ISO-8859-1' : 'Latin1', 9 'ISO-8859-1' : 'Latin1',
15 'ISO-8859-6' : 'Arabic', 15 'ISO-8859-6' : 'Arabic',
16 'ISO-8859-7' : 'Greek', 16 'ISO-8859-7' : 'Greek',
17 'ISO-8859-8' : 'Hebrew', 17 'ISO-8859-8' : 'Hebrew',
18 'TIS-620' : 'Thai', 18 'TIS-620' : 'Thai',
19 'SHIFT-JIS' : 'Japanese', 19 'SHIFT-JIS' : 'Japanese',
20 #'GB18030' : 'Chinese', 20 #'GB18030' : 'Chinese', # Done manually below (*)
21 } 21 }
22 22
23 #from encodings.aliases import aliases 23 #from encodings.aliases import aliases
24 #for a, b in aliases.iteritems(): 24 #for a, b in aliases.iteritems():
25 # print '%s : %s' % (a, b) 25 # print '%s : %s' % (a, b)
42 s = source.encode(encoding, 'ignore') 42 s = source.encode(encoding, 'ignore')
43 encoded.append(ToArray(s)) 43 encoded.append(ToArray(s))
44 expected.append(ToArray(s.decode(encoding).encode('utf-8'))) 44 expected.append(ToArray(s.decode(encoding).encode('utf-8')))
45 45
46 46
47 # https://en.wikipedia.org/wiki/GB_18030#Technical_details 47 # https://en.wikipedia.org/wiki/GB_18030#Technical_details (*)
48 l.append('::Orthanc::Encoding_Chinese') 48 l.append('::Orthanc::Encoding_Chinese')
49 expected.append(ToArray('Þßàáâã')) 49 expected.append(ToArray('Þßàáâã'))
50 encoded.append('"\\x81\\x30\\x89\\x37\\x81\\x30\\x89\\x38\\xA8\\xA4\\xA8\\xA2\\x81\\x30\\x89\\x39\\x81\\x30\\x8A\\x30"') 50 encoded.append('"\\x81\\x30\\x89\\x37\\x81\\x30\\x89\\x38\\xA8\\xA4\\xA8\\xA2\\x81\\x30\\x89\\x39\\x81\\x30\\x8A\\x30"')
51 51
52 52