Coverage Report

Created: 2024-02-25 07:20

/src/libfshfs/libuna/libuna_codepage_koi8_r.c
Line
Count
Source (jump to first uncovered line)
1
/*
2
 * KOI8-R codepage (Russian Cyrillic) function
3
 *
4
 * Copyright (C) 2008-2024, Joachim Metz <joachim.metz@gmail.com>
5
 *
6
 * Refer to AUTHORS for acknowledgements.
7
 *
8
 * This program is free software: you can redistribute it and/or modify
9
 * it under the terms of the GNU Lesser General Public License as published by
10
 * the Free Software Foundation, either version 3 of the License, or
11
 * (at your option) any later version.
12
 *
13
 * This program is distributed in the hope that it will be useful,
14
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
15
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
16
 * GNU General Public License for more details.
17
 *
18
 * You should have received a copy of the GNU Lesser General Public License
19
 * along with this program.  If not, see <https://www.gnu.org/licenses/>.
20
 */
21
22
#include <common.h>
23
#include <types.h>
24
25
#include "libuna_codepage_koi8_r.h"
26
#include "libuna_libcerror.h"
27
#include "libuna_types.h"
28
29
/* Extended ASCII to Unicode character lookup table for the KOI8-R codepage
30
 * Unknown are filled with the Unicode replacement character 0xfffd
31
 */
32
const uint16_t libuna_codepage_koi8_r_byte_stream_to_unicode_base_0x80[ 128 ] = {
33
  0x2500, 0x2502, 0x250c, 0x2510, 0x2514, 0x2518, 0x251c, 0x2524,
34
  0x252c, 0x2534, 0x253c, 0x2580, 0x2584, 0x2588, 0x258c, 0x2590,
35
  0x2591, 0x2592, 0x2593, 0x2320, 0x25a0, 0x2219, 0x221a, 0x2248,
36
  0x2264, 0x2265, 0x00a0, 0x2321, 0x00b0, 0x00b2, 0x00b7, 0x00f7,
37
  0x2550, 0x2551, 0x2552, 0x0451, 0x2553, 0x2554, 0x2555, 0x2556,
38
  0x2557, 0x2558, 0x2559, 0x255a, 0x255b, 0x255c, 0x255d, 0x255e,
39
  0x255f, 0x2560, 0x2561, 0x0401, 0x2562, 0x2563, 0x2564, 0x2565,
40
  0x2566, 0x2567, 0x2568, 0x2569, 0x256a, 0x256b, 0x256c, 0x00a9,
41
  0x044e, 0x0430, 0x0431, 0x0446, 0x0434, 0x0435, 0x0444, 0x0433,
42
  0x0445, 0x0438, 0x0439, 0x043a, 0x043b, 0x043c, 0x043d, 0x043e,
43
  0x043f, 0x044f, 0x0440, 0x0441, 0x0442, 0x0443, 0x0436, 0x0432,
44
  0x044c, 0x044b, 0x0437, 0x0448, 0x044d, 0x0449, 0x0447, 0x044a,
45
  0x042e, 0x0410, 0x0411, 0x0426, 0x0414, 0x0415, 0x0424, 0x0413,
46
  0x0425, 0x0418, 0x0419, 0x041a, 0x041b, 0x041c, 0x041d, 0x041e,
47
  0x041f, 0x042f, 0x0420, 0x0421, 0x0422, 0x0423, 0x0416, 0x0412,
48
  0x042c, 0x042b, 0x0417, 0x0428, 0x042d, 0x0429, 0x0427, 0x042a
49
};
50
51
/* Unicode to ASCII character lookup tables for the KOI8-R codepage
52
 * Unknown are filled with the ASCII replacement character 0x1a
53
 */
54
const uint8_t libuna_codepage_koi8_r_unicode_to_byte_stream_base_0x0410[ 64 ] = {
55
  0xe1, 0xe2, 0xf7, 0xe7, 0xe4, 0xe5, 0xf6, 0xfa,
56
  0xe9, 0xea, 0xeb, 0xec, 0xed, 0xee, 0xef, 0xf0,
57
  0xf2, 0xf3, 0xf4, 0xf5, 0xe6, 0xe8, 0xe3, 0xfe,
58
  0xfb, 0xfd, 0xff, 0xf9, 0xf8, 0xfc, 0xe0, 0xf1,
59
  0xc1, 0xc2, 0xd7, 0xc7, 0xc4, 0xc5, 0xd6, 0xda,
60
  0xc9, 0xca, 0xcb, 0xcc, 0xcd, 0xce, 0xcf, 0xd0,
61
  0xd2, 0xd3, 0xd4, 0xd5, 0xc6, 0xc8, 0xc3, 0xde,
62
  0xdb, 0xdd, 0xdf, 0xd9, 0xd8, 0xdc, 0xc0, 0xd1
63
};
64
65
const uint8_t libuna_codepage_koi8_r_unicode_to_byte_stream_base_0x2550[ 32 ] = {
66
  0xa0, 0xa1, 0xa2, 0xa4, 0xa5, 0xa6, 0xa7, 0xa8,
67
  0xa9, 0xaa, 0xab, 0xac, 0xad, 0xae, 0xaf, 0xb0,
68
  0xb1, 0xb2, 0xb4, 0xb5, 0xb6, 0xb7, 0xb8, 0xb9,
69
  0xba, 0xbb, 0xbc, 0xbd, 0xbe, 0x1a, 0x1a, 0x1a
70
};
71
72
/* Copies an Unicode character from a KOI8-R encoded byte stream
73
 * Returns 1 if successful or -1 on error
74
 */
75
int libuna_codepage_koi8_r_copy_from_byte_stream(
76
     libuna_unicode_character_t *unicode_character,
77
     const uint8_t *byte_stream,
78
     size_t byte_stream_size,
79
     size_t *byte_stream_index,
80
     libcerror_error_t **error )
81
0
{
82
0
  static char *function                             = "libuna_codepage_koi8_r_copy_from_byte_stream";
83
0
  libuna_unicode_character_t safe_unicode_character = 0xfffd;
84
0
  size_t safe_byte_stream_index                     = 0;
85
0
  uint8_t byte_stream_character                     = 0;
86
87
0
  if( unicode_character == NULL )
88
0
  {
89
0
    libcerror_error_set(
90
0
     error,
91
0
     LIBCERROR_ERROR_DOMAIN_ARGUMENTS,
92
0
     LIBCERROR_ARGUMENT_ERROR_INVALID_VALUE,
93
0
     "%s: invalid Unicode character.",
94
0
     function );
95
96
0
    return( -1 );
97
0
  }
98
0
  if( byte_stream == NULL )
99
0
  {
100
0
    libcerror_error_set(
101
0
     error,
102
0
     LIBCERROR_ERROR_DOMAIN_ARGUMENTS,
103
0
     LIBCERROR_ARGUMENT_ERROR_INVALID_VALUE,
104
0
     "%s: invalid byte stream.",
105
0
     function );
106
107
0
    return( -1 );
108
0
  }
109
0
  if( byte_stream_size > (size_t) SSIZE_MAX )
110
0
  {
111
0
    libcerror_error_set(
112
0
     error,
113
0
     LIBCERROR_ERROR_DOMAIN_ARGUMENTS,
114
0
     LIBCERROR_ARGUMENT_ERROR_VALUE_EXCEEDS_MAXIMUM,
115
0
     "%s: invalid byte stream size value exceeds maximum.",
116
0
     function );
117
118
0
    return( -1 );
119
0
  }
120
0
  if( byte_stream_index == NULL )
121
0
  {
122
0
    libcerror_error_set(
123
0
     error,
124
0
     LIBCERROR_ERROR_DOMAIN_ARGUMENTS,
125
0
     LIBCERROR_ARGUMENT_ERROR_INVALID_VALUE,
126
0
     "%s: invalid byte stream index.",
127
0
     function );
128
129
0
    return( -1 );
130
0
  }
131
0
  safe_byte_stream_index = *byte_stream_index;
132
133
0
  if( safe_byte_stream_index >= byte_stream_size )
134
0
  {
135
0
    libcerror_error_set(
136
0
     error,
137
0
     LIBCERROR_ERROR_DOMAIN_ARGUMENTS,
138
0
     LIBCERROR_ARGUMENT_ERROR_VALUE_TOO_SMALL,
139
0
     "%s: byte stream too small.",
140
0
     function );
141
142
0
    return( -1 );
143
0
  }
144
0
  byte_stream_character = byte_stream[ safe_byte_stream_index++ ];
145
146
0
  if( byte_stream_character < 0x80 )
147
0
  {
148
0
    safe_unicode_character = byte_stream_character;
149
0
  }
150
0
  else
151
0
  {
152
0
    byte_stream_character -= 0x80;
153
154
0
    safe_unicode_character = libuna_codepage_koi8_r_byte_stream_to_unicode_base_0x80[ byte_stream_character ];
155
0
  }
156
0
  *unicode_character = safe_unicode_character;
157
0
  *byte_stream_index = safe_byte_stream_index;
158
159
0
  return( 1 );
160
0
}
161
162
/* Copies an Unicode character to a KOI8-R encoded byte stream
163
 * Returns 1 if successful or -1 on error
164
 */
165
int libuna_codepage_koi8_r_copy_to_byte_stream(
166
     libuna_unicode_character_t unicode_character,
167
     uint8_t *byte_stream,
168
     size_t byte_stream_size,
169
     size_t *byte_stream_index,
170
     libcerror_error_t **error )
171
0
{
172
0
  static char *function         = "libuna_codepage_koi8_r_copy_to_byte_stream";
173
0
  size_t safe_byte_stream_index = 0;
174
0
  uint16_t byte_stream_value    = 0x001a;
175
176
0
  if( byte_stream == NULL )
177
0
  {
178
0
    libcerror_error_set(
179
0
     error,
180
0
     LIBCERROR_ERROR_DOMAIN_ARGUMENTS,
181
0
     LIBCERROR_ARGUMENT_ERROR_INVALID_VALUE,
182
0
     "%s: invalid byte stream.",
183
0
     function );
184
185
0
    return( -1 );
186
0
  }
187
0
  if( byte_stream_size > (size_t) SSIZE_MAX )
188
0
  {
189
0
    libcerror_error_set(
190
0
     error,
191
0
     LIBCERROR_ERROR_DOMAIN_ARGUMENTS,
192
0
     LIBCERROR_ARGUMENT_ERROR_VALUE_EXCEEDS_MAXIMUM,
193
0
     "%s: invalid byte stream size value exceeds maximum.",
194
0
     function );
195
196
0
    return( -1 );
197
0
  }
198
0
  if( byte_stream_index == NULL )
199
0
  {
200
0
    libcerror_error_set(
201
0
     error,
202
0
     LIBCERROR_ERROR_DOMAIN_ARGUMENTS,
203
0
     LIBCERROR_ARGUMENT_ERROR_INVALID_VALUE,
204
0
     "%s: invalid byte stream index.",
205
0
     function );
206
207
0
    return( -1 );
208
0
  }
209
0
  safe_byte_stream_index = *byte_stream_index;
210
211
0
  if( safe_byte_stream_index >= byte_stream_size )
212
0
  {
213
0
    libcerror_error_set(
214
0
     error,
215
0
     LIBCERROR_ERROR_DOMAIN_ARGUMENTS,
216
0
     LIBCERROR_ARGUMENT_ERROR_VALUE_TOO_SMALL,
217
0
     "%s: byte stream too small.",
218
0
     function );
219
220
0
    return( -1 );
221
0
  }
222
0
  if( unicode_character < 0x0080 )
223
0
  {
224
0
    byte_stream_value = (uint16_t) unicode_character;
225
0
  }
226
0
  else if( ( unicode_character >= 0x0410 )
227
0
        && ( unicode_character < 0x0450 ) )
228
0
  {
229
0
    unicode_character -= 0x0410;
230
231
0
    byte_stream_value = libuna_codepage_koi8_r_unicode_to_byte_stream_base_0x0410[ unicode_character ];
232
0
  }
233
0
  else if( ( unicode_character >= 0x2550 )
234
0
        && ( unicode_character < 0x2570 ) )
235
0
  {
236
0
    unicode_character -= 0x2550;
237
238
0
    byte_stream_value = libuna_codepage_koi8_r_unicode_to_byte_stream_base_0x2550[ unicode_character ];
239
0
  }
240
0
  else switch( unicode_character )
241
0
  {
242
0
    case 0x00a0:
243
0
      byte_stream_value = 0x9a;
244
0
      break;
245
246
0
    case 0x00a9:
247
0
      byte_stream_value = 0xbf;
248
0
      break;
249
250
0
    case 0x00b0:
251
0
      byte_stream_value = 0x9c;
252
0
      break;
253
254
0
    case 0x00b2:
255
0
      byte_stream_value = 0x9d;
256
0
      break;
257
258
0
    case 0x00b7:
259
0
      byte_stream_value = 0x9e;
260
0
      break;
261
262
0
    case 0x00f7:
263
0
      byte_stream_value = 0x9f;
264
0
      break;
265
266
0
    case 0x0401:
267
0
      byte_stream_value = 0xb3;
268
0
      break;
269
270
0
    case 0x0451:
271
0
      byte_stream_value = 0xa3;
272
0
      break;
273
274
0
    case 0x2219:
275
0
      byte_stream_value = 0x95;
276
0
      break;
277
278
0
    case 0x221a:
279
0
      byte_stream_value = 0x96;
280
0
      break;
281
282
0
    case 0x2248:
283
0
      byte_stream_value = 0x97;
284
0
      break;
285
286
0
    case 0x2264:
287
0
      byte_stream_value = 0x98;
288
0
      break;
289
290
0
    case 0x2265:
291
0
      byte_stream_value = 0x99;
292
0
      break;
293
294
0
    case 0x2320:
295
0
      byte_stream_value = 0x93;
296
0
      break;
297
298
0
    case 0x2321:
299
0
      byte_stream_value = 0x9b;
300
0
      break;
301
302
0
    case 0x2500:
303
0
      byte_stream_value = 0x80;
304
0
      break;
305
306
0
    case 0x2502:
307
0
      byte_stream_value = 0x81;
308
0
      break;
309
310
0
    case 0x250c:
311
0
      byte_stream_value = 0x82;
312
0
      break;
313
314
0
    case 0x2510:
315
0
      byte_stream_value = 0x83;
316
0
      break;
317
318
0
    case 0x2514:
319
0
      byte_stream_value = 0x84;
320
0
      break;
321
322
0
    case 0x2518:
323
0
      byte_stream_value = 0x85;
324
0
      break;
325
326
0
    case 0x251c:
327
0
      byte_stream_value = 0x86;
328
0
      break;
329
330
0
    case 0x2524:
331
0
      byte_stream_value = 0x87;
332
0
      break;
333
334
0
    case 0x252c:
335
0
      byte_stream_value = 0x88;
336
0
      break;
337
338
0
    case 0x2534:
339
0
      byte_stream_value = 0x89;
340
0
      break;
341
342
0
    case 0x253c:
343
0
      byte_stream_value = 0x8a;
344
0
      break;
345
346
0
    case 0x2580:
347
0
      byte_stream_value = 0x8b;
348
0
      break;
349
350
0
    case 0x2584:
351
0
      byte_stream_value = 0x8c;
352
0
      break;
353
354
0
    case 0x2588:
355
0
      byte_stream_value = 0x8d;
356
0
      break;
357
358
0
    case 0x258c:
359
0
      byte_stream_value = 0x8e;
360
0
      break;
361
362
0
    case 0x2590:
363
0
      byte_stream_value = 0x8f;
364
0
      break;
365
366
0
    case 0x2591:
367
0
      byte_stream_value = 0x90;
368
0
      break;
369
370
0
    case 0x2592:
371
0
      byte_stream_value = 0x91;
372
0
      break;
373
374
0
    case 0x2593:
375
0
      byte_stream_value = 0x92;
376
0
      break;
377
378
0
    case 0x25a0:
379
0
      byte_stream_value = 0x94;
380
0
      break;
381
382
0
    default:
383
0
      byte_stream_value = 0x1a;
384
0
      break;
385
0
  }
386
0
  byte_stream[ safe_byte_stream_index++ ] = (uint8_t) ( byte_stream_value & 0x00ff );
387
388
0
  *byte_stream_index = safe_byte_stream_index;
389
390
0
  return( 1 );
391
0
}
392