| Guido van Rossum | 0229bf6 | 2000-03-10 23:17:24 +0000 | [diff] [blame] | 1 | """ Python Character Mapping Codec generated from 'CP1252.TXT'. |
| 2 | |
| 3 | |
| 4 | Written by Marc-Andre Lemburg (mal@lemburg.com). |
| 5 | |
| 6 | (c) Copyright CNRI, All Rights Reserved. NO WARRANTY. |
| 7 | |
| 8 | """#" |
| 9 | |
| 10 | import codecs |
| 11 | |
| 12 | ### Codec APIs |
| 13 | |
| 14 | class Codec(codecs.Codec): |
| 15 | |
| 16 | def encode(self,input,errors='strict'): |
| 17 | |
| 18 | return codecs.charmap_encode(input,errors,encoding_map) |
| 19 | |
| 20 | def decode(self,input,errors='strict'): |
| 21 | |
| 22 | return codecs.charmap_decode(input,errors,decoding_map) |
| 23 | |
| 24 | class StreamWriter(Codec,codecs.StreamWriter): |
| 25 | pass |
| 26 | |
| 27 | class StreamReader(Codec,codecs.StreamReader): |
| 28 | pass |
| 29 | |
| 30 | ### encodings module API |
| 31 | |
| 32 | def getregentry(): |
| 33 | |
| 34 | return (Codec().encode,Codec().decode,StreamReader,StreamWriter) |
| 35 | |
| 36 | ### Decoding Map |
| 37 | |
| 38 | decoding_map = { |
| 39 | |
| 40 | 0x0080: 0x20ac, # EURO SIGN |
| 41 | 0x0081: None, # UNDEFINED |
| 42 | 0x0082: 0x201a, # SINGLE LOW-9 QUOTATION MARK |
| 43 | 0x0083: 0x0192, # LATIN SMALL LETTER F WITH HOOK |
| 44 | 0x0084: 0x201e, # DOUBLE LOW-9 QUOTATION MARK |
| 45 | 0x0085: 0x2026, # HORIZONTAL ELLIPSIS |
| 46 | 0x0086: 0x2020, # DAGGER |
| 47 | 0x0087: 0x2021, # DOUBLE DAGGER |
| 48 | 0x0088: 0x02c6, # MODIFIER LETTER CIRCUMFLEX ACCENT |
| 49 | 0x0089: 0x2030, # PER MILLE SIGN |
| 50 | 0x008a: 0x0160, # LATIN CAPITAL LETTER S WITH CARON |
| 51 | 0x008b: 0x2039, # SINGLE LEFT-POINTING ANGLE QUOTATION MARK |
| 52 | 0x008c: 0x0152, # LATIN CAPITAL LIGATURE OE |
| 53 | 0x008d: None, # UNDEFINED |
| 54 | 0x008e: 0x017d, # LATIN CAPITAL LETTER Z WITH CARON |
| 55 | 0x008f: None, # UNDEFINED |
| 56 | 0x0090: None, # UNDEFINED |
| 57 | 0x0091: 0x2018, # LEFT SINGLE QUOTATION MARK |
| 58 | 0x0092: 0x2019, # RIGHT SINGLE QUOTATION MARK |
| 59 | 0x0093: 0x201c, # LEFT DOUBLE QUOTATION MARK |
| 60 | 0x0094: 0x201d, # RIGHT DOUBLE QUOTATION MARK |
| 61 | 0x0095: 0x2022, # BULLET |
| 62 | 0x0096: 0x2013, # EN DASH |
| 63 | 0x0097: 0x2014, # EM DASH |
| 64 | 0x0098: 0x02dc, # SMALL TILDE |
| 65 | 0x0099: 0x2122, # TRADE MARK SIGN |
| 66 | 0x009a: 0x0161, # LATIN SMALL LETTER S WITH CARON |
| 67 | 0x009b: 0x203a, # SINGLE RIGHT-POINTING ANGLE QUOTATION MARK |
| 68 | 0x009c: 0x0153, # LATIN SMALL LIGATURE OE |
| 69 | 0x009d: None, # UNDEFINED |
| 70 | 0x009e: 0x017e, # LATIN SMALL LETTER Z WITH CARON |
| 71 | 0x009f: 0x0178, # LATIN CAPITAL LETTER Y WITH DIAERESIS |
| 72 | } |
| 73 | |
| 74 | ### Encoding Map |
| 75 | |
| 76 | encoding_map = {} |
| 77 | for k,v in decoding_map.items(): |
| 78 | encoding_map[v] = k |