Coverage Report

Created: 2024-02-25 07:20

/src/libftxf/libuna/libuna_codepage_mac_thai.c
Line
Count
Source (jump to first uncovered line)
1
/*
2
 * MacThai codepage functions
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_mac_thai.h"
26
#include "libuna_libcerror.h"
27
#include "libuna_types.h"
28
29
/* Extended ASCII to Unicode character lookup table for the MacThai codepage
30
 * Unknown are filled with the Unicode replacement character 0xfffd
31
 */
32
const uint16_t libuna_codepage_mac_thai_byte_stream_to_unicode_base_0x80[ 128 ] = {
33
  0x00ab, 0x00bb, 0x2026, 0x0e48, 0x0e49, 0x0e4a, 0x0e4b, 0x0e4c,
34
  0x0e48, 0x0e49, 0x0e4a, 0x0e4b, 0x0e4c, 0x201c, 0x201d, 0x0e4d,
35
  0xfffd, 0x2022, 0x0e31, 0x0e47, 0x0e34, 0x0e35, 0x0e36, 0x0e37,
36
  0x0e48, 0x0e49, 0x0e4a, 0x0e4b, 0x0e4c, 0x2018, 0x2019, 0xfffd,
37
  0x00a0, 0x0e01, 0x0e02, 0x0e03, 0x0e04, 0x0e05, 0x0e06, 0x0e07,
38
  0x0e08, 0x0e09, 0x0e0a, 0x0e0b, 0x0e0c, 0x0e0d, 0x0e0e, 0x0e0f,
39
  0x0e10, 0x0e11, 0x0e12, 0x0e13, 0x0e14, 0x0e15, 0x0e16, 0x0e17,
40
  0x0e18, 0x0e19, 0x0e1a, 0x0e1b, 0x0e1c, 0x0e1d, 0x0e1e, 0x0e1f,
41
  0x0e20, 0x0e21, 0x0e22, 0x0e23, 0x0e24, 0x0e25, 0x0e26, 0x0e27,
42
  0x0e28, 0x0e29, 0x0e2a, 0x0e2b, 0x0e2c, 0x0e2d, 0x0e2e, 0x0e2f,
43
  0x0e30, 0x0e31, 0x0e32, 0x0e33, 0x0e34, 0x0e35, 0x0e36, 0x0e37,
44
  0x0e38, 0x0e39, 0x0e3a, 0x2060, 0x200b, 0x2013, 0x2014, 0x0e3f,
45
  0x0e40, 0x0e41, 0x0e42, 0x0e43, 0x0e44, 0x0e45, 0x0e46, 0x0e47,
46
  0x0e48, 0x0e49, 0x0e4a, 0x0e4b, 0x0e4c, 0x0e4d, 0x2122, 0x0e4f,
47
  0x0e50, 0x0e51, 0x0e52, 0x0e53, 0x0e54, 0x0e55, 0x0e56, 0x0e57,
48
  0x0e58, 0x0e59, 0x00ae, 0x00a9, 0xfffd, 0xfffd, 0xfffd, 0xfffd
49
};
50
51
/* Unicode to ASCII character lookup tables for the MacThai codepage
52
 * Unknown are filled with the ASCII replacement character 0x1a
53
 */
54
const uint8_t libuna_codepage_mac_thai_unicode_to_byte_stream_base_0x0e00[ 96 ] = {
55
  0x1a, 0xa1, 0xa2, 0xa3, 0xa4, 0xa5, 0xa6, 0xa7,
56
  0xa8, 0xa9, 0xaa, 0xab, 0xac, 0xad, 0xae, 0xaf,
57
  0xb0, 0xb1, 0xb2, 0xb3, 0xb4, 0xb5, 0xb6, 0xb7,
58
  0xb8, 0xb9, 0xba, 0xbb, 0xbc, 0xbd, 0xbe, 0xbf,
59
  0xc0, 0xc1, 0xc2, 0xc3, 0xc4, 0xc5, 0xc6, 0xc7,
60
  0xc8, 0xc9, 0xca, 0xcb, 0xcc, 0xcd, 0xce, 0xcf,
61
  0xd0, 0xd1, 0xd2, 0xd3, 0xd4, 0xd5, 0xd6, 0xd7,
62
  0xd8, 0xd9, 0xda, 0x1a, 0x1a, 0x1a, 0x1a, 0xdf,
63
  0xe0, 0xe1, 0xe2, 0xe3, 0xe4, 0xe5, 0xe6, 0xe7,
64
  0xe8, 0xe9, 0xea, 0xeb, 0xec, 0xed, 0x1a, 0xef,
65
  0xf0, 0xf1, 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xf7,
66
  0xf8, 0xf9, 0x1a, 0x1a, 0x1a, 0x1a, 0x1a, 0x1a
67
};
68
69
const uint8_t libuna_codepage_mac_thai_unicode_to_byte_stream_base_0x2008[ 32 ] = {
70
  0x1a, 0x1a, 0x1a, 0xdc, 0x1a, 0x1a, 0x1a, 0x1a,
71
  0x1a, 0x1a, 0x1a, 0xdd, 0xde, 0x1a, 0x1a, 0x1a,
72
  0x9d, 0x9e, 0x1a, 0x1a, 0x8d, 0x8e, 0x1a, 0x1a,
73
  0x1a, 0x1a, 0x91, 0x1a, 0x1a, 0x1a, 0x82, 0x1a
74
};
75
76
/* Determines the size of a MacThai encoded byte stream from an Unicode character
77
 * Adds the size to the byte stream character size value
78
 * Returns 1 if successful, 0 if the byte stream character is valid but not supported since it requires special handling or -1 on error
79
 */
80
int libuna_codepage_mac_thai_unicode_character_size_to_byte_stream(
81
     libuna_unicode_character_t unicode_character,
82
     size_t *byte_stream_character_size,
83
     libcerror_error_t **error )
84
0
{
85
0
  static char *function = "libuna_codepage_mac_thai_unicode_character_size_to_byte_stream";
86
0
  int result            = 0;
87
88
0
  if( byte_stream_character_size == NULL )
89
0
  {
90
0
    libcerror_error_set(
91
0
     error,
92
0
     LIBCERROR_ERROR_DOMAIN_ARGUMENTS,
93
0
     LIBCERROR_ARGUMENT_ERROR_INVALID_VALUE,
94
0
     "%s: invalid byte stream character size.",
95
0
     function );
96
97
0
    return( -1 );
98
0
  }
99
0
  switch( unicode_character )
100
0
  {
101
0
    case 0x00000e31UL:
102
0
    case 0x00000e34UL:
103
0
    case 0x00000e35UL:
104
0
    case 0x00000e36UL:
105
0
    case 0x00000e37UL:
106
0
    case 0x00000e47UL:
107
0
    case 0x00000e48UL:
108
0
    case 0x00000e49UL:
109
0
    case 0x00000e4aUL:
110
0
    case 0x00000e4bUL:
111
0
    case 0x00000e4cUL:
112
0
    case 0x00000e4dUL:
113
0
      result = 0;
114
0
      break;
115
116
0
    default:
117
0
      *byte_stream_character_size += 1;
118
119
0
      result = 1;
120
0
      break;
121
0
  }
122
0
  return( result );
123
0
}
124
125
/* Copies an Unicode character from a MacThai encoded byte stream
126
 * Returns 1 if successful or -1 on error
127
 */
128
int libuna_codepage_mac_thai_copy_from_byte_stream(
129
     libuna_unicode_character_t *unicode_character,
130
     const uint8_t *byte_stream,
131
     size_t byte_stream_size,
132
     size_t *byte_stream_index,
133
     libcerror_error_t **error )
134
0
{
135
0
  static char *function                             = "libuna_codepage_mac_thai_copy_from_byte_stream";
136
0
  libuna_unicode_character_t safe_unicode_character = 0xfffd;
137
0
  size_t safe_byte_stream_index                     = 0;
138
0
  uint8_t byte_stream_character                     = 0;
139
140
0
  if( unicode_character == NULL )
141
0
  {
142
0
    libcerror_error_set(
143
0
     error,
144
0
     LIBCERROR_ERROR_DOMAIN_ARGUMENTS,
145
0
     LIBCERROR_ARGUMENT_ERROR_INVALID_VALUE,
146
0
     "%s: invalid Unicode character.",
147
0
     function );
148
149
0
    return( -1 );
150
0
  }
151
0
  if( byte_stream == NULL )
152
0
  {
153
0
    libcerror_error_set(
154
0
     error,
155
0
     LIBCERROR_ERROR_DOMAIN_ARGUMENTS,
156
0
     LIBCERROR_ARGUMENT_ERROR_INVALID_VALUE,
157
0
     "%s: invalid byte stream.",
158
0
     function );
159
160
0
    return( -1 );
161
0
  }
162
0
  if( byte_stream_size > (size_t) SSIZE_MAX )
163
0
  {
164
0
    libcerror_error_set(
165
0
     error,
166
0
     LIBCERROR_ERROR_DOMAIN_ARGUMENTS,
167
0
     LIBCERROR_ARGUMENT_ERROR_VALUE_EXCEEDS_MAXIMUM,
168
0
     "%s: invalid byte stream size value exceeds maximum.",
169
0
     function );
170
171
0
    return( -1 );
172
0
  }
173
0
  if( byte_stream_index == NULL )
174
0
  {
175
0
    libcerror_error_set(
176
0
     error,
177
0
     LIBCERROR_ERROR_DOMAIN_ARGUMENTS,
178
0
     LIBCERROR_ARGUMENT_ERROR_INVALID_VALUE,
179
0
     "%s: invalid byte stream index.",
180
0
     function );
181
182
0
    return( -1 );
183
0
  }
184
0
  safe_byte_stream_index = *byte_stream_index;
185
186
0
  if( safe_byte_stream_index >= byte_stream_size )
187
0
  {
188
0
    libcerror_error_set(
189
0
     error,
190
0
     LIBCERROR_ERROR_DOMAIN_ARGUMENTS,
191
0
     LIBCERROR_ARGUMENT_ERROR_VALUE_TOO_SMALL,
192
0
     "%s: byte stream too small.",
193
0
     function );
194
195
0
    return( -1 );
196
0
  }
197
0
  byte_stream_character = byte_stream[ safe_byte_stream_index++ ];
198
199
0
  if( byte_stream_character < 0x80 )
200
0
  {
201
0
    safe_unicode_character = byte_stream_character;
202
0
  }
203
0
  else
204
0
  {
205
0
    byte_stream_character -= 0x80;
206
207
0
    safe_unicode_character = libuna_codepage_mac_thai_byte_stream_to_unicode_base_0x80[ byte_stream_character ];
208
0
  }
209
0
  *unicode_character = safe_unicode_character;
210
0
  *byte_stream_index = safe_byte_stream_index;
211
212
0
  return( 1 );
213
0
}
214
215
/* Copies an Unicode character to a MacThai encoded byte stream
216
 * Returns 1 if successful or -1 on error
217
 */
218
int libuna_codepage_mac_thai_copy_to_byte_stream(
219
     libuna_unicode_character_t unicode_character,
220
     uint8_t *byte_stream,
221
     size_t byte_stream_size,
222
     size_t *byte_stream_index,
223
     libcerror_error_t **error )
224
0
{
225
0
  static char *function         = "libuna_codepage_mac_thai_copy_to_byte_stream";
226
0
  size_t safe_byte_stream_index = 0;
227
0
  uint16_t byte_stream_value    = 0x001a;
228
229
0
  if( byte_stream == NULL )
230
0
  {
231
0
    libcerror_error_set(
232
0
     error,
233
0
     LIBCERROR_ERROR_DOMAIN_ARGUMENTS,
234
0
     LIBCERROR_ARGUMENT_ERROR_INVALID_VALUE,
235
0
     "%s: invalid byte stream.",
236
0
     function );
237
238
0
    return( -1 );
239
0
  }
240
0
  if( byte_stream_size > (size_t) SSIZE_MAX )
241
0
  {
242
0
    libcerror_error_set(
243
0
     error,
244
0
     LIBCERROR_ERROR_DOMAIN_ARGUMENTS,
245
0
     LIBCERROR_ARGUMENT_ERROR_VALUE_EXCEEDS_MAXIMUM,
246
0
     "%s: invalid byte stream size value exceeds maximum.",
247
0
     function );
248
249
0
    return( -1 );
250
0
  }
251
0
  if( byte_stream_index == NULL )
252
0
  {
253
0
    libcerror_error_set(
254
0
     error,
255
0
     LIBCERROR_ERROR_DOMAIN_ARGUMENTS,
256
0
     LIBCERROR_ARGUMENT_ERROR_INVALID_VALUE,
257
0
     "%s: invalid byte stream index.",
258
0
     function );
259
260
0
    return( -1 );
261
0
  }
262
0
  safe_byte_stream_index = *byte_stream_index;
263
264
0
  if( safe_byte_stream_index >= byte_stream_size )
265
0
  {
266
0
    libcerror_error_set(
267
0
     error,
268
0
     LIBCERROR_ERROR_DOMAIN_ARGUMENTS,
269
0
     LIBCERROR_ARGUMENT_ERROR_VALUE_TOO_SMALL,
270
0
     "%s: byte stream too small.",
271
0
     function );
272
273
0
    return( -1 );
274
0
  }
275
0
  if( ( unicode_character < 0x0080 )
276
0
   || ( unicode_character == 0x00a0 ) )
277
0
  {
278
0
    byte_stream_value = (uint16_t) unicode_character;
279
0
  }
280
0
  else if( ( unicode_character >= 0x0e00 )
281
0
        && ( unicode_character < 0x0e60 ) )
282
0
  {
283
0
    unicode_character -= 0x0e00;
284
285
0
    byte_stream_value = libuna_codepage_mac_thai_unicode_to_byte_stream_base_0x0e00[ unicode_character ];
286
0
  }
287
0
  else if( ( unicode_character >= 0x2008 )
288
0
        && ( unicode_character < 0x2028 ) )
289
0
  {
290
0
    unicode_character -= 0x2008;
291
292
0
    byte_stream_value = libuna_codepage_mac_thai_unicode_to_byte_stream_base_0x2008[ unicode_character ];
293
0
  }
294
0
  else switch( unicode_character )
295
0
  {
296
0
    case 0x00a9:
297
0
      byte_stream_value = 0xfb;
298
0
      break;
299
300
0
    case 0x00ab:
301
0
      byte_stream_value = 0x80;
302
0
      break;
303
304
0
    case 0x00ae:
305
0
      byte_stream_value = 0xfa;
306
0
      break;
307
308
0
    case 0x00bb:
309
0
      byte_stream_value = 0x81;
310
0
      break;
311
312
0
    case 0x2060:
313
0
      byte_stream_value = 0xdb;
314
0
      break;
315
316
0
    case 0x2122:
317
0
      byte_stream_value = 0xee;
318
0
      break;
319
320
0
    default:
321
0
      byte_stream_value = 0x1a;
322
0
      break;
323
0
  }
324
0
  byte_stream[ safe_byte_stream_index++ ] = (uint8_t) ( byte_stream_value & 0x00ff );
325
326
0
  *byte_stream_index = safe_byte_stream_index;
327
328
0
  return( 1 );
329
0
}
330