Coverage Report

Created: 2025-01-28 06:58

/src/libidn2/lib/context.c
Line
Count
Source (jump to first uncovered line)
1
/* context.c - check contextual rule on label
2
   Copyright (C) 2011-2024 Simon Josefsson
3
4
   Libidn2 is free software: you can redistribute it and/or modify it
5
   under the terms of either:
6
7
     * the GNU Lesser General Public License as published by the Free
8
       Software Foundation; either version 3 of the License, or (at
9
       your option) any later version.
10
11
   or
12
13
     * the GNU General Public License as published by the Free
14
       Software Foundation; either version 2 of the License, or (at
15
       your option) any later version.
16
17
   or both in parallel, as here.
18
19
   This program is distributed in the hope that it will be useful,
20
   but WITHOUT ANY WARRANTY; without even the implied warranty of
21
   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
22
   GNU General Public License for more details.
23
24
   You should have received copies of the GNU General Public License and
25
   the GNU Lesser General Public License along with this program.  If
26
   not, see <http://www.gnu.org/licenses/>.
27
*/
28
29
#include <config.h>
30
31
#include "idn2.h"
32
#include "tables.h"
33
#include <unictype.h>   /* uc_combining_class, UC_CCC_VR */
34
#include "context.h"
35
36
int
37
_idn2_contextj_rule (const uint32_t *label, size_t llen, size_t pos)
38
45.5k
{
39
45.5k
  uint32_t cp;
40
41
45.5k
  if (llen == 0)
42
0
    return IDN2_OK;
43
44
45.5k
  cp = label[pos];
45
46
45.5k
  if (!_idn2_contextj_p (cp))
47
44.1k
    return IDN2_OK;
48
49
1.37k
  switch (cp)
50
1.37k
    {
51
1.07k
    case 0x200C:    /* ZERO WIDTH NON-JOINER */
52
1.07k
      if (pos > 0)
53
1.04k
  {
54
    /* If Canonical_Combining_Class(Before(cp)) .eq.  Virama Then True; */
55
1.04k
    uint32_t before_cp = label[pos - 1];
56
1.04k
    int cc = uc_combining_class (before_cp);
57
1.04k
    if (cc == UC_CCC_VR)
58
217
      return IDN2_OK;
59
1.04k
  }
60
61
      /* See http://permalink.gmane.org/gmane.ietf.idnabis/6980 for
62
         clarified rule. */
63
64
861
      if (pos == 0 || pos == llen - 1)
65
79
  return IDN2_CONTEXTJ;
66
67
782
      {
68
782
  int jt;
69
782
  size_t tmp;
70
71
  /* Search backwards. */
72
782
  for (tmp = pos - 1;; tmp--)
73
1.34k
    {
74
1.34k
      jt = uc_joining_type (label[tmp]);
75
1.34k
      if (jt == UC_JOINING_TYPE_L || jt == UC_JOINING_TYPE_D)
76
626
        break;
77
715
      if (tmp == 0)
78
130
        return IDN2_CONTEXTJ;
79
585
      if (jt == UC_JOINING_TYPE_T)
80
559
        continue;
81
26
      return IDN2_CONTEXTJ;
82
585
    }
83
84
  /* Search forward. */
85
972
  for (tmp = pos + 1; tmp < llen; tmp++)
86
972
    {
87
972
      jt = uc_joining_type (label[tmp]);
88
972
      if (jt == UC_JOINING_TYPE_R || jt == UC_JOINING_TYPE_D)
89
517
        break;
90
455
      if (tmp == llen - 1)
91
95
        return IDN2_CONTEXTJ;
92
360
      if (jt == UC_JOINING_TYPE_T)
93
346
        continue;
94
14
      return IDN2_CONTEXTJ;
95
360
    }
96
626
      }
97
98
517
      return IDN2_OK;
99
0
      break;
100
101
292
    case 0x200D:    /* ZERO WIDTH JOINER */
102
292
      if (pos > 0)
103
281
  {
104
281
    uint32_t before_cp = label[pos - 1];
105
281
    int cc = uc_combining_class (before_cp);
106
281
    if (cc == UC_CCC_VR)
107
241
      return IDN2_OK;
108
281
  }
109
51
      return IDN2_CONTEXTJ;
110
1.37k
    }
111
112
0
  return IDN2_CONTEXTJ_NO_RULE;
113
1.37k
}
114
115
static const char *
116
_uc_script_name (ucs4_t uc)
117
0
{
118
0
  const uc_script_t *ucs = uc_script (uc);
119
120
0
  if (!ucs)
121
0
    return "";
122
123
0
  return ucs->name;
124
0
}
125
126
int
127
_idn2_contexto_rule (const uint32_t *label, size_t llen, size_t pos)
128
0
{
129
0
  uint32_t cp = label[pos];
130
131
0
  if (!_idn2_contexto_p (cp))
132
0
    return IDN2_OK;
133
134
0
  switch (cp)
135
0
    {
136
0
    case 0x00B7:
137
      /* MIDDLE DOT */
138
0
      if (llen < 3)
139
0
  return IDN2_CONTEXTO;
140
0
      if (pos == 0 || pos == llen - 1)
141
0
  return IDN2_CONTEXTO;
142
0
      if (label[pos - 1] == 0x006C && label[pos + 1] == 0x006C)
143
0
  return IDN2_OK;
144
0
      return IDN2_CONTEXTO;
145
0
      break;
146
147
0
    case 0x0375:
148
      /* GREEK LOWER NUMERAL SIGN (KERAIA) */
149
0
      if (pos == llen - 1)
150
0
  return IDN2_CONTEXTO;
151
0
      if (strcmp (_uc_script_name (label[pos + 1]), "Greek") == 0)
152
0
  return IDN2_OK;
153
0
      return IDN2_CONTEXTO;
154
0
      break;
155
156
0
    case 0x05F3:
157
      /* HEBREW PUNCTUATION GERESH */
158
0
    case 0x05F4:
159
      /* HEBREW PUNCTUATION GERSHAYIM */
160
0
      if (pos == 0)
161
0
  return IDN2_CONTEXTO;
162
0
      if (strcmp (_uc_script_name (label[pos - 1]), "Hebrew") == 0)
163
0
  return IDN2_OK;
164
0
      return IDN2_CONTEXTO;
165
0
      break;
166
167
0
    case 0x0660:
168
0
    case 0x0661:
169
0
    case 0x0662:
170
0
    case 0x0663:
171
0
    case 0x0664:
172
0
    case 0x0665:
173
0
    case 0x0666:
174
0
    case 0x0667:
175
0
    case 0x0668:
176
0
    case 0x0669:
177
0
      {
178
  /* ARABIC-INDIC DIGITS */
179
0
  size_t i;
180
0
  for (i = 0; i < llen; i++)
181
0
    if (label[i] >= 0x6F0 && label[i] <= 0x06F9)
182
0
      return IDN2_CONTEXTO;
183
0
  return IDN2_OK;
184
0
  break;
185
0
      }
186
187
0
    case 0x06F0:
188
0
    case 0x06F1:
189
0
    case 0x06F2:
190
0
    case 0x06F3:
191
0
    case 0x06F4:
192
0
    case 0x06F5:
193
0
    case 0x06F6:
194
0
    case 0x06F7:
195
0
    case 0x06F8:
196
0
    case 0x06F9:
197
0
      {
198
  /* EXTENDED ARABIC-INDIC DIGITS */
199
0
  size_t i;
200
0
  for (i = 0; i < llen; i++)
201
0
    if (label[i] >= 0x660 && label[i] <= 0x0669)
202
0
      return IDN2_CONTEXTO;
203
0
  return IDN2_OK;
204
0
  break;
205
0
      }
206
0
    case 0x30FB:
207
0
      {
208
  /* KATAKANA MIDDLE DOT */
209
0
  size_t i;
210
0
  bool script_ok = false;
211
212
0
  for (i = 0; !script_ok && i < llen; i++)
213
0
    if (strcmp (_uc_script_name (label[i]), "Hiragana") == 0
214
0
        || strcmp (_uc_script_name (label[i]), "Katakana") == 0
215
0
        || strcmp (_uc_script_name (label[i]), "Han") == 0)
216
0
      script_ok = true;
217
218
0
  if (script_ok)
219
0
    return IDN2_OK;
220
0
  return IDN2_CONTEXTO;
221
0
  break;
222
0
      }
223
0
    }
224
225
0
  return IDN2_CONTEXTO_NO_RULE;
226
0
}
227
228
bool
229
_idn2_contexto_with_rule (uint32_t cp)
230
7.29k
{
231
7.29k
  switch (cp)
232
7.29k
    {
233
617
    case 0x00B7:
234
      /* MIDDLE DOT */
235
871
    case 0x0375:
236
      /* GREEK LOWER NUMERAL SIGN (KERAIA) */
237
1.25k
    case 0x05F3:
238
      /* HEBREW PUNCTUATION GERESH */
239
1.48k
    case 0x05F4:
240
      /* HEBREW PUNCTUATION GERSHAYIM */
241
1.83k
    case 0x0660:
242
2.04k
    case 0x0661:
243
2.28k
    case 0x0662:
244
2.51k
    case 0x0663:
245
2.73k
    case 0x0664:
246
2.97k
    case 0x0665:
247
3.19k
    case 0x0666:
248
3.44k
    case 0x0667:
249
3.65k
    case 0x0668:
250
3.87k
    case 0x0669:
251
      /* ARABIC-INDIC DIGITS */
252
4.20k
    case 0x06F0:
253
4.45k
    case 0x06F1:
254
4.69k
    case 0x06F2:
255
5.02k
    case 0x06F3:
256
5.44k
    case 0x06F4:
257
5.79k
    case 0x06F5:
258
6.00k
    case 0x06F6:
259
6.26k
    case 0x06F7:
260
6.46k
    case 0x06F8:
261
6.75k
    case 0x06F9:
262
      /* EXTENDED ARABIC-INDIC DIGITS */
263
7.29k
    case 0x30FB:
264
      /* KATAKANA MIDDLE DOT */
265
7.29k
      return true;
266
0
      break;
267
7.29k
    }
268
269
0
  return false;
270
7.29k
}