Mercurial > repos > rliterman > csp2

comparison CSP2/CSP2_env/env-d9b9114564458d9d-741b3de822f2aaca6c6caa4325c4afce/lib/python3.8/encodings/mac_cyrillic.py @ 69:33d812a61356

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload commit 2e9511a184a1ca667c7be0c6321a36dc4e3d116d
author jpayne
date Tue, 18 Mar 2025 17:55:14 -0400
parents
children
comparison
equal deleted inserted replaced
67:0e9998148a16 69:33d812a61356
1 """ Python Character Mapping Codec mac_cyrillic generated from 'MAPPINGS/VENDORS/APPLE/CYRILLIC.TXT' with gencodec.py.
2
3 """#"
4
5 import codecs
6
7 ### Codec APIs
8
9 class Codec(codecs.Codec):
10
11 def encode(self,input,errors='strict'):
12 return codecs.charmap_encode(input,errors,encoding_table)
13
14 def decode(self,input,errors='strict'):
15 return codecs.charmap_decode(input,errors,decoding_table)
16
17 class IncrementalEncoder(codecs.IncrementalEncoder):
18 def encode(self, input, final=False):
19 return codecs.charmap_encode(input,self.errors,encoding_table)[0]
20
21 class IncrementalDecoder(codecs.IncrementalDecoder):
22 def decode(self, input, final=False):
23 return codecs.charmap_decode(input,self.errors,decoding_table)[0]
24
25 class StreamWriter(Codec,codecs.StreamWriter):
26 pass
27
28 class StreamReader(Codec,codecs.StreamReader):
29 pass
30
31 ### encodings module API
32
33 def getregentry():
34 return codecs.CodecInfo(
35 name='mac-cyrillic',
36 encode=Codec().encode,
37 decode=Codec().decode,
38 incrementalencoder=IncrementalEncoder,
39 incrementaldecoder=IncrementalDecoder,
40 streamreader=StreamReader,
41 streamwriter=StreamWriter,
42 )
43
44
45 ### Decoding Table
46
47 decoding_table = (
48 '\x00' # 0x00 -> CONTROL CHARACTER
49 '\x01' # 0x01 -> CONTROL CHARACTER
50 '\x02' # 0x02 -> CONTROL CHARACTER
51 '\x03' # 0x03 -> CONTROL CHARACTER
52 '\x04' # 0x04 -> CONTROL CHARACTER
53 '\x05' # 0x05 -> CONTROL CHARACTER
54 '\x06' # 0x06 -> CONTROL CHARACTER
55 '\x07' # 0x07 -> CONTROL CHARACTER
56 '\x08' # 0x08 -> CONTROL CHARACTER
57 '\t' # 0x09 -> CONTROL CHARACTER
58 '\n' # 0x0A -> CONTROL CHARACTER
59 '\x0b' # 0x0B -> CONTROL CHARACTER
60 '\x0c' # 0x0C -> CONTROL CHARACTER
61 '\r' # 0x0D -> CONTROL CHARACTER
62 '\x0e' # 0x0E -> CONTROL CHARACTER
63 '\x0f' # 0x0F -> CONTROL CHARACTER
64 '\x10' # 0x10 -> CONTROL CHARACTER
65 '\x11' # 0x11 -> CONTROL CHARACTER
66 '\x12' # 0x12 -> CONTROL CHARACTER
67 '\x13' # 0x13 -> CONTROL CHARACTER
68 '\x14' # 0x14 -> CONTROL CHARACTER
69 '\x15' # 0x15 -> CONTROL CHARACTER
70 '\x16' # 0x16 -> CONTROL CHARACTER
71 '\x17' # 0x17 -> CONTROL CHARACTER
72 '\x18' # 0x18 -> CONTROL CHARACTER
73 '\x19' # 0x19 -> CONTROL CHARACTER
74 '\x1a' # 0x1A -> CONTROL CHARACTER
75 '\x1b' # 0x1B -> CONTROL CHARACTER
76 '\x1c' # 0x1C -> CONTROL CHARACTER
77 '\x1d' # 0x1D -> CONTROL CHARACTER
78 '\x1e' # 0x1E -> CONTROL CHARACTER
79 '\x1f' # 0x1F -> CONTROL CHARACTER
80 ' ' # 0x20 -> SPACE
81 '!' # 0x21 -> EXCLAMATION MARK
82 '"' # 0x22 -> QUOTATION MARK
83 '#' # 0x23 -> NUMBER SIGN
84 '$' # 0x24 -> DOLLAR SIGN
85 '%' # 0x25 -> PERCENT SIGN
86 '&' # 0x26 -> AMPERSAND
87 "'" # 0x27 -> APOSTROPHE
88 '(' # 0x28 -> LEFT PARENTHESIS
89 ')' # 0x29 -> RIGHT PARENTHESIS
90 '*' # 0x2A -> ASTERISK
91 '+' # 0x2B -> PLUS SIGN
92 ',' # 0x2C -> COMMA
93 '-' # 0x2D -> HYPHEN-MINUS
94 '.' # 0x2E -> FULL STOP
95 '/' # 0x2F -> SOLIDUS
96 '0' # 0x30 -> DIGIT ZERO
97 '1' # 0x31 -> DIGIT ONE
98 '2' # 0x32 -> DIGIT TWO
99 '3' # 0x33 -> DIGIT THREE
100 '4' # 0x34 -> DIGIT FOUR
101 '5' # 0x35 -> DIGIT FIVE
102 '6' # 0x36 -> DIGIT SIX
103 '7' # 0x37 -> DIGIT SEVEN
104 '8' # 0x38 -> DIGIT EIGHT
105 '9' # 0x39 -> DIGIT NINE
106 ':' # 0x3A -> COLON
107 ';' # 0x3B -> SEMICOLON
108 '<' # 0x3C -> LESS-THAN SIGN
109 '=' # 0x3D -> EQUALS SIGN
110 '>' # 0x3E -> GREATER-THAN SIGN
111 '?' # 0x3F -> QUESTION MARK
112 '@' # 0x40 -> COMMERCIAL AT
113 'A' # 0x41 -> LATIN CAPITAL LETTER A
114 'B' # 0x42 -> LATIN CAPITAL LETTER B
115 'C' # 0x43 -> LATIN CAPITAL LETTER C
116 'D' # 0x44 -> LATIN CAPITAL LETTER D
117 'E' # 0x45 -> LATIN CAPITAL LETTER E
118 'F' # 0x46 -> LATIN CAPITAL LETTER F
119 'G' # 0x47 -> LATIN CAPITAL LETTER G
120 'H' # 0x48 -> LATIN CAPITAL LETTER H
121 'I' # 0x49 -> LATIN CAPITAL LETTER I
122 'J' # 0x4A -> LATIN CAPITAL LETTER J
123 'K' # 0x4B -> LATIN CAPITAL LETTER K
124 'L' # 0x4C -> LATIN CAPITAL LETTER L
125 'M' # 0x4D -> LATIN CAPITAL LETTER M
126 'N' # 0x4E -> LATIN CAPITAL LETTER N
127 'O' # 0x4F -> LATIN CAPITAL LETTER O
128 'P' # 0x50 -> LATIN CAPITAL LETTER P
129 'Q' # 0x51 -> LATIN CAPITAL LETTER Q
130 'R' # 0x52 -> LATIN CAPITAL LETTER R
131 'S' # 0x53 -> LATIN CAPITAL LETTER S
132 'T' # 0x54 -> LATIN CAPITAL LETTER T
133 'U' # 0x55 -> LATIN CAPITAL LETTER U
134 'V' # 0x56 -> LATIN CAPITAL LETTER V
135 'W' # 0x57 -> LATIN CAPITAL LETTER W
136 'X' # 0x58 -> LATIN CAPITAL LETTER X
137 'Y' # 0x59 -> LATIN CAPITAL LETTER Y
138 'Z' # 0x5A -> LATIN CAPITAL LETTER Z
139 '[' # 0x5B -> LEFT SQUARE BRACKET
140 '\\' # 0x5C -> REVERSE SOLIDUS
141 ']' # 0x5D -> RIGHT SQUARE BRACKET
142 '^' # 0x5E -> CIRCUMFLEX ACCENT
143 '_' # 0x5F -> LOW LINE
144 '`' # 0x60 -> GRAVE ACCENT
145 'a' # 0x61 -> LATIN SMALL LETTER A
146 'b' # 0x62 -> LATIN SMALL LETTER B
147 'c' # 0x63 -> LATIN SMALL LETTER C
148 'd' # 0x64 -> LATIN SMALL LETTER D
149 'e' # 0x65 -> LATIN SMALL LETTER E
150 'f' # 0x66 -> LATIN SMALL LETTER F
151 'g' # 0x67 -> LATIN SMALL LETTER G
152 'h' # 0x68 -> LATIN SMALL LETTER H
153 'i' # 0x69 -> LATIN SMALL LETTER I
154 'j' # 0x6A -> LATIN SMALL LETTER J
155 'k' # 0x6B -> LATIN SMALL LETTER K
156 'l' # 0x6C -> LATIN SMALL LETTER L
157 'm' # 0x6D -> LATIN SMALL LETTER M
158 'n' # 0x6E -> LATIN SMALL LETTER N
159 'o' # 0x6F -> LATIN SMALL LETTER O
160 'p' # 0x70 -> LATIN SMALL LETTER P
161 'q' # 0x71 -> LATIN SMALL LETTER Q
162 'r' # 0x72 -> LATIN SMALL LETTER R
163 's' # 0x73 -> LATIN SMALL LETTER S
164 't' # 0x74 -> LATIN SMALL LETTER T
165 'u' # 0x75 -> LATIN SMALL LETTER U
166 'v' # 0x76 -> LATIN SMALL LETTER V
167 'w' # 0x77 -> LATIN SMALL LETTER W
168 'x' # 0x78 -> LATIN SMALL LETTER X
169 'y' # 0x79 -> LATIN SMALL LETTER Y
170 'z' # 0x7A -> LATIN SMALL LETTER Z
171 '{' # 0x7B -> LEFT CURLY BRACKET
172 '|' # 0x7C -> VERTICAL LINE
173 '}' # 0x7D -> RIGHT CURLY BRACKET
174 '~' # 0x7E -> TILDE
175 '\x7f' # 0x7F -> CONTROL CHARACTER
176 '\u0410' # 0x80 -> CYRILLIC CAPITAL LETTER A
177 '\u0411' # 0x81 -> CYRILLIC CAPITAL LETTER BE
178 '\u0412' # 0x82 -> CYRILLIC CAPITAL LETTER VE
179 '\u0413' # 0x83 -> CYRILLIC CAPITAL LETTER GHE
180 '\u0414' # 0x84 -> CYRILLIC CAPITAL LETTER DE
181 '\u0415' # 0x85 -> CYRILLIC CAPITAL LETTER IE
182 '\u0416' # 0x86 -> CYRILLIC CAPITAL LETTER ZHE
183 '\u0417' # 0x87 -> CYRILLIC CAPITAL LETTER ZE
184 '\u0418' # 0x88 -> CYRILLIC CAPITAL LETTER I
185 '\u0419' # 0x89 -> CYRILLIC CAPITAL LETTER SHORT I
186 '\u041a' # 0x8A -> CYRILLIC CAPITAL LETTER KA
187 '\u041b' # 0x8B -> CYRILLIC CAPITAL LETTER EL
188 '\u041c' # 0x8C -> CYRILLIC CAPITAL LETTER EM
189 '\u041d' # 0x8D -> CYRILLIC CAPITAL LETTER EN
190 '\u041e' # 0x8E -> CYRILLIC CAPITAL LETTER O
191 '\u041f' # 0x8F -> CYRILLIC CAPITAL LETTER PE
192 '\u0420' # 0x90 -> CYRILLIC CAPITAL LETTER ER
193 '\u0421' # 0x91 -> CYRILLIC CAPITAL LETTER ES
194 '\u0422' # 0x92 -> CYRILLIC CAPITAL LETTER TE
195 '\u0423' # 0x93 -> CYRILLIC CAPITAL LETTER U
196 '\u0424' # 0x94 -> CYRILLIC CAPITAL LETTER EF
197 '\u0425' # 0x95 -> CYRILLIC CAPITAL LETTER HA
198 '\u0426' # 0x96 -> CYRILLIC CAPITAL LETTER TSE
199 '\u0427' # 0x97 -> CYRILLIC CAPITAL LETTER CHE
200 '\u0428' # 0x98 -> CYRILLIC CAPITAL LETTER SHA
201 '\u0429' # 0x99 -> CYRILLIC CAPITAL LETTER SHCHA
202 '\u042a' # 0x9A -> CYRILLIC CAPITAL LETTER HARD SIGN
203 '\u042b' # 0x9B -> CYRILLIC CAPITAL LETTER YERU
204 '\u042c' # 0x9C -> CYRILLIC CAPITAL LETTER SOFT SIGN
205 '\u042d' # 0x9D -> CYRILLIC CAPITAL LETTER E
206 '\u042e' # 0x9E -> CYRILLIC CAPITAL LETTER YU
207 '\u042f' # 0x9F -> CYRILLIC CAPITAL LETTER YA
208 '\u2020' # 0xA0 -> DAGGER
209 '\xb0' # 0xA1 -> DEGREE SIGN
210 '\u0490' # 0xA2 -> CYRILLIC CAPITAL LETTER GHE WITH UPTURN
211 '\xa3' # 0xA3 -> POUND SIGN
212 '\xa7' # 0xA4 -> SECTION SIGN
213 '\u2022' # 0xA5 -> BULLET
214 '\xb6' # 0xA6 -> PILCROW SIGN
215 '\u0406' # 0xA7 -> CYRILLIC CAPITAL LETTER BYELORUSSIAN-UKRAINIAN I
216 '\xae' # 0xA8 -> REGISTERED SIGN
217 '\xa9' # 0xA9 -> COPYRIGHT SIGN
218 '\u2122' # 0xAA -> TRADE MARK SIGN
219 '\u0402' # 0xAB -> CYRILLIC CAPITAL LETTER DJE
220 '\u0452' # 0xAC -> CYRILLIC SMALL LETTER DJE
221 '\u2260' # 0xAD -> NOT EQUAL TO
222 '\u0403' # 0xAE -> CYRILLIC CAPITAL LETTER GJE
223 '\u0453' # 0xAF -> CYRILLIC SMALL LETTER GJE
224 '\u221e' # 0xB0 -> INFINITY
225 '\xb1' # 0xB1 -> PLUS-MINUS SIGN
226 '\u2264' # 0xB2 -> LESS-THAN OR EQUAL TO
227 '\u2265' # 0xB3 -> GREATER-THAN OR EQUAL TO
228 '\u0456' # 0xB4 -> CYRILLIC SMALL LETTER BYELORUSSIAN-UKRAINIAN I
229 '\xb5' # 0xB5 -> MICRO SIGN
230 '\u0491' # 0xB6 -> CYRILLIC SMALL LETTER GHE WITH UPTURN
231 '\u0408' # 0xB7 -> CYRILLIC CAPITAL LETTER JE
232 '\u0404' # 0xB8 -> CYRILLIC CAPITAL LETTER UKRAINIAN IE
233 '\u0454' # 0xB9 -> CYRILLIC SMALL LETTER UKRAINIAN IE
234 '\u0407' # 0xBA -> CYRILLIC CAPITAL LETTER YI
235 '\u0457' # 0xBB -> CYRILLIC SMALL LETTER YI
236 '\u0409' # 0xBC -> CYRILLIC CAPITAL LETTER LJE
237 '\u0459' # 0xBD -> CYRILLIC SMALL LETTER LJE
238 '\u040a' # 0xBE -> CYRILLIC CAPITAL LETTER NJE
239 '\u045a' # 0xBF -> CYRILLIC SMALL LETTER NJE
240 '\u0458' # 0xC0 -> CYRILLIC SMALL LETTER JE
241 '\u0405' # 0xC1 -> CYRILLIC CAPITAL LETTER DZE
242 '\xac' # 0xC2 -> NOT SIGN
243 '\u221a' # 0xC3 -> SQUARE ROOT
244 '\u0192' # 0xC4 -> LATIN SMALL LETTER F WITH HOOK
245 '\u2248' # 0xC5 -> ALMOST EQUAL TO
246 '\u2206' # 0xC6 -> INCREMENT
247 '\xab' # 0xC7 -> LEFT-POINTING DOUBLE ANGLE QUOTATION MARK
248 '\xbb' # 0xC8 -> RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK
249 '\u2026' # 0xC9 -> HORIZONTAL ELLIPSIS
250 '\xa0' # 0xCA -> NO-BREAK SPACE
251 '\u040b' # 0xCB -> CYRILLIC CAPITAL LETTER TSHE
252 '\u045b' # 0xCC -> CYRILLIC SMALL LETTER TSHE
253 '\u040c' # 0xCD -> CYRILLIC CAPITAL LETTER KJE
254 '\u045c' # 0xCE -> CYRILLIC SMALL LETTER KJE
255 '\u0455' # 0xCF -> CYRILLIC SMALL LETTER DZE
256 '\u2013' # 0xD0 -> EN DASH
257 '\u2014' # 0xD1 -> EM DASH
258 '\u201c' # 0xD2 -> LEFT DOUBLE QUOTATION MARK
259 '\u201d' # 0xD3 -> RIGHT DOUBLE QUOTATION MARK
260 '\u2018' # 0xD4 -> LEFT SINGLE QUOTATION MARK
261 '\u2019' # 0xD5 -> RIGHT SINGLE QUOTATION MARK
262 '\xf7' # 0xD6 -> DIVISION SIGN
263 '\u201e' # 0xD7 -> DOUBLE LOW-9 QUOTATION MARK
264 '\u040e' # 0xD8 -> CYRILLIC CAPITAL LETTER SHORT U
265 '\u045e' # 0xD9 -> CYRILLIC SMALL LETTER SHORT U
266 '\u040f' # 0xDA -> CYRILLIC CAPITAL LETTER DZHE
267 '\u045f' # 0xDB -> CYRILLIC SMALL LETTER DZHE
268 '\u2116' # 0xDC -> NUMERO SIGN
269 '\u0401' # 0xDD -> CYRILLIC CAPITAL LETTER IO
270 '\u0451' # 0xDE -> CYRILLIC SMALL LETTER IO
271 '\u044f' # 0xDF -> CYRILLIC SMALL LETTER YA
272 '\u0430' # 0xE0 -> CYRILLIC SMALL LETTER A
273 '\u0431' # 0xE1 -> CYRILLIC SMALL LETTER BE
274 '\u0432' # 0xE2 -> CYRILLIC SMALL LETTER VE
275 '\u0433' # 0xE3 -> CYRILLIC SMALL LETTER GHE
276 '\u0434' # 0xE4 -> CYRILLIC SMALL LETTER DE
277 '\u0435' # 0xE5 -> CYRILLIC SMALL LETTER IE
278 '\u0436' # 0xE6 -> CYRILLIC SMALL LETTER ZHE
279 '\u0437' # 0xE7 -> CYRILLIC SMALL LETTER ZE
280 '\u0438' # 0xE8 -> CYRILLIC SMALL LETTER I
281 '\u0439' # 0xE9 -> CYRILLIC SMALL LETTER SHORT I
282 '\u043a' # 0xEA -> CYRILLIC SMALL LETTER KA
283 '\u043b' # 0xEB -> CYRILLIC SMALL LETTER EL
284 '\u043c' # 0xEC -> CYRILLIC SMALL LETTER EM
285 '\u043d' # 0xED -> CYRILLIC SMALL LETTER EN
286 '\u043e' # 0xEE -> CYRILLIC SMALL LETTER O
287 '\u043f' # 0xEF -> CYRILLIC SMALL LETTER PE
288 '\u0440' # 0xF0 -> CYRILLIC SMALL LETTER ER
289 '\u0441' # 0xF1 -> CYRILLIC SMALL LETTER ES
290 '\u0442' # 0xF2 -> CYRILLIC SMALL LETTER TE
291 '\u0443' # 0xF3 -> CYRILLIC SMALL LETTER U
292 '\u0444' # 0xF4 -> CYRILLIC SMALL LETTER EF
293 '\u0445' # 0xF5 -> CYRILLIC SMALL LETTER HA
294 '\u0446' # 0xF6 -> CYRILLIC SMALL LETTER TSE
295 '\u0447' # 0xF7 -> CYRILLIC SMALL LETTER CHE
296 '\u0448' # 0xF8 -> CYRILLIC SMALL LETTER SHA
297 '\u0449' # 0xF9 -> CYRILLIC SMALL LETTER SHCHA
298 '\u044a' # 0xFA -> CYRILLIC SMALL LETTER HARD SIGN
299 '\u044b' # 0xFB -> CYRILLIC SMALL LETTER YERU
300 '\u044c' # 0xFC -> CYRILLIC SMALL LETTER SOFT SIGN
301 '\u044d' # 0xFD -> CYRILLIC SMALL LETTER E
302 '\u044e' # 0xFE -> CYRILLIC SMALL LETTER YU
303 '\u20ac' # 0xFF -> EURO SIGN
304 )
305
306 ### Encoding table
307 encoding_table=codecs.charmap_build(decoding_table)