Coverage Report

Created: 2025-07-07 10:01

/work/workdir/UnpackedTarball/harfbuzz/src/hb-ot-shaper-hebrew.cc
Line
Count
Source (jump to first uncovered line)
1
/*
2
 * Copyright © 2010,2012  Google, Inc.
3
 *
4
 *  This is part of HarfBuzz, a text shaping library.
5
 *
6
 * Permission is hereby granted, without written agreement and without
7
 * license or royalty fees, to use, copy, modify, and distribute this
8
 * software and its documentation for any purpose, provided that the
9
 * above copyright notice and the following two paragraphs appear in
10
 * all copies of this software.
11
 *
12
 * IN NO EVENT SHALL THE COPYRIGHT HOLDER BE LIABLE TO ANY PARTY FOR
13
 * DIRECT, INDIRECT, SPECIAL, INCIDENTAL, OR CONSEQUENTIAL DAMAGES
14
 * ARISING OUT OF THE USE OF THIS SOFTWARE AND ITS DOCUMENTATION, EVEN
15
 * IF THE COPYRIGHT HOLDER HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH
16
 * DAMAGE.
17
 *
18
 * THE COPYRIGHT HOLDER SPECIFICALLY DISCLAIMS ANY WARRANTIES, INCLUDING,
19
 * BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND
20
 * FITNESS FOR A PARTICULAR PURPOSE.  THE SOFTWARE PROVIDED HEREUNDER IS
21
 * ON AN "AS IS" BASIS, AND THE COPYRIGHT HOLDER HAS NO OBLIGATION TO
22
 * PROVIDE MAINTENANCE, SUPPORT, UPDATES, ENHANCEMENTS, OR MODIFICATIONS.
23
 *
24
 * Google Author(s): Behdad Esfahbod
25
 */
26
27
#include "hb.hh"
28
29
#ifndef HB_NO_OT_SHAPE
30
31
#include "hb-ot-shaper.hh"
32
33
34
static bool
35
compose_hebrew (const hb_ot_shape_normalize_context_t *c,
36
    hb_codepoint_t  a,
37
    hb_codepoint_t  b,
38
    hb_codepoint_t *ab)
39
64.6k
{
40
  /* Hebrew presentation-form shaping.
41
   * https://bugzilla.mozilla.org/show_bug.cgi?id=728866
42
   * Hebrew presentation forms with dagesh, for characters U+05D0..05EA;
43
   * Note that some letters do not have a dagesh presForm encoded.
44
   */
45
64.6k
  static const hb_codepoint_t sDageshForms[0x05EAu - 0x05D0u + 1] = {
46
64.6k
    0xFB30u, /* ALEF */
47
64.6k
    0xFB31u, /* BET */
48
64.6k
    0xFB32u, /* GIMEL */
49
64.6k
    0xFB33u, /* DALET */
50
64.6k
    0xFB34u, /* HE */
51
64.6k
    0xFB35u, /* VAV */
52
64.6k
    0xFB36u, /* ZAYIN */
53
64.6k
    0x0000u, /* HET */
54
64.6k
    0xFB38u, /* TET */
55
64.6k
    0xFB39u, /* YOD */
56
64.6k
    0xFB3Au, /* FINAL KAF */
57
64.6k
    0xFB3Bu, /* KAF */
58
64.6k
    0xFB3Cu, /* LAMED */
59
64.6k
    0x0000u, /* FINAL MEM */
60
64.6k
    0xFB3Eu, /* MEM */
61
64.6k
    0x0000u, /* FINAL NUN */
62
64.6k
    0xFB40u, /* NUN */
63
64.6k
    0xFB41u, /* SAMEKH */
64
64.6k
    0x0000u, /* AYIN */
65
64.6k
    0xFB43u, /* FINAL PE */
66
64.6k
    0xFB44u, /* PE */
67
64.6k
    0x0000u, /* FINAL TSADI */
68
64.6k
    0xFB46u, /* TSADI */
69
64.6k
    0xFB47u, /* QOF */
70
64.6k
    0xFB48u, /* RESH */
71
64.6k
    0xFB49u, /* SHIN */
72
64.6k
    0xFB4Au /* TAV */
73
64.6k
  };
74
75
64.6k
  bool found = (bool) c->unicode->compose (a, b, ab);
76
77
#ifdef HB_NO_OT_SHAPER_HEBREW_FALLBACK
78
  return found;
79
#endif
80
81
64.6k
  if (!found && (c->plan && !c->plan->has_gpos_mark))
82
0
  {
83
      /* Special-case Hebrew presentation forms that are excluded from
84
       * standard normalization, but wanted for old fonts. */
85
0
      switch (b) {
86
0
      case 0x05B4u: /* HIRIQ */
87
0
    if (a == 0x05D9u) { /* YOD */
88
0
        *ab = 0xFB1Du;
89
0
        found = true;
90
0
    }
91
0
    break;
92
0
      case 0x05B7u: /* PATAH */
93
0
    if (a == 0x05F2u) { /* YIDDISH YOD YOD */
94
0
        *ab = 0xFB1Fu;
95
0
        found = true;
96
0
    } else if (a == 0x05D0u) { /* ALEF */
97
0
        *ab = 0xFB2Eu;
98
0
        found = true;
99
0
    }
100
0
    break;
101
0
      case 0x05B8u: /* QAMATS */
102
0
    if (a == 0x05D0u) { /* ALEF */
103
0
        *ab = 0xFB2Fu;
104
0
        found = true;
105
0
    }
106
0
    break;
107
0
      case 0x05B9u: /* HOLAM */
108
0
    if (a == 0x05D5u) { /* VAV */
109
0
        *ab = 0xFB4Bu;
110
0
        found = true;
111
0
    }
112
0
    break;
113
0
      case 0x05BCu: /* DAGESH */
114
0
    if (a >= 0x05D0u && a <= 0x05EAu) {
115
0
        *ab = sDageshForms[a - 0x05D0u];
116
0
        found = (*ab != 0);
117
0
    } else if (a == 0xFB2Au) { /* SHIN WITH SHIN DOT */
118
0
        *ab = 0xFB2Cu;
119
0
        found = true;
120
0
    } else if (a == 0xFB2Bu) { /* SHIN WITH SIN DOT */
121
0
        *ab = 0xFB2Du;
122
0
        found = true;
123
0
    }
124
0
    break;
125
0
      case 0x05BFu: /* RAFE */
126
0
    switch (a) {
127
0
    case 0x05D1u: /* BET */
128
0
        *ab = 0xFB4Cu;
129
0
        found = true;
130
0
        break;
131
0
    case 0x05DBu: /* KAF */
132
0
        *ab = 0xFB4Du;
133
0
        found = true;
134
0
        break;
135
0
    case 0x05E4u: /* PE */
136
0
        *ab = 0xFB4Eu;
137
0
        found = true;
138
0
        break;
139
0
    }
140
0
    break;
141
0
      case 0x05C1u: /* SHIN DOT */
142
0
    if (a == 0x05E9u) { /* SHIN */
143
0
        *ab = 0xFB2Au;
144
0
        found = true;
145
0
    } else if (a == 0xFB49u) { /* SHIN WITH DAGESH */
146
0
        *ab = 0xFB2Cu;
147
0
        found = true;
148
0
    }
149
0
    break;
150
0
      case 0x05C2u: /* SIN DOT */
151
0
    if (a == 0x05E9u) { /* SHIN */
152
0
        *ab = 0xFB2Bu;
153
0
        found = true;
154
0
    } else if (a == 0xFB49u) { /* SHIN WITH DAGESH */
155
0
        *ab = 0xFB2Du;
156
0
        found = true;
157
0
    }
158
0
    break;
159
0
      }
160
0
  }
161
162
64.6k
  return found;
163
64.6k
}
164
165
static void
166
reorder_marks_hebrew (const hb_ot_shape_plan_t *plan HB_UNUSED,
167
          hb_buffer_t              *buffer,
168
          unsigned int              start,
169
          unsigned int              end)
170
62.7k
{
171
62.7k
  hb_glyph_info_t *info = buffer->info;
172
173
75.3k
  for (unsigned i = start + 2; i < end; i++)
174
12.5k
  {
175
12.5k
    unsigned c0 = info_cc (info[i - 2]);
176
12.5k
    unsigned c1 = info_cc (info[i - 1]);
177
12.5k
    unsigned c2 = info_cc (info[i - 0]);
178
179
12.5k
    if ((c0 == HB_MODIFIED_COMBINING_CLASS_CCC17 || c0 == HB_MODIFIED_COMBINING_CLASS_CCC18) /* patach or qamats */ &&
180
12.5k
  (c1 == HB_MODIFIED_COMBINING_CLASS_CCC10 || c1 == HB_MODIFIED_COMBINING_CLASS_CCC14) /* sheva or hiriq */ &&
181
12.5k
  (c2 == HB_MODIFIED_COMBINING_CLASS_CCC22 || c2 == HB_UNICODE_COMBINING_CLASS_BELOW) /* meteg or below */)
182
0
    {
183
0
      buffer->merge_clusters (i - 1, i + 1);
184
0
      hb_swap (info[i - 1], info[i]);
185
0
      break;
186
0
    }
187
12.5k
  }
188
189
190
62.7k
}
191
192
const hb_ot_shaper_t _hb_ot_shaper_hebrew =
193
{
194
  nullptr, /* collect_features */
195
  nullptr, /* override_features */
196
  nullptr, /* data_create */
197
  nullptr, /* data_destroy */
198
  nullptr, /* preprocess_text */
199
  nullptr, /* postprocess_glyphs */
200
  nullptr, /* decompose */
201
  compose_hebrew,
202
  nullptr, /* setup_masks */
203
  reorder_marks_hebrew,
204
  HB_TAG ('h','e','b','r'), /* gpos_tag. https://github.com/harfbuzz/harfbuzz/issues/347#issuecomment-267838368 */
205
  HB_OT_SHAPE_NORMALIZATION_MODE_DEFAULT,
206
  HB_OT_SHAPE_ZERO_WIDTH_MARKS_BY_GDEF_LATE,
207
  true, /* fallback_position */
208
};
209
210
211
#endif