Coverage Report

Created: 2023-06-07 06:20

/src/mupdf/source/fitz/text.c
Line
Count
Source (jump to first uncovered line)
1
// Copyright (C) 2004-2021 Artifex Software, Inc.
2
//
3
// This file is part of MuPDF.
4
//
5
// MuPDF is free software: you can redistribute it and/or modify it under the
6
// terms of the GNU Affero General Public License as published by the Free
7
// Software Foundation, either version 3 of the License, or (at your option)
8
// any later version.
9
//
10
// MuPDF is distributed in the hope that it will be useful, but WITHOUT ANY
11
// WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
12
// FOR A PARTICULAR PURPOSE. See the GNU Affero General Public License for more
13
// details.
14
//
15
// You should have received a copy of the GNU Affero General Public License
16
// along with MuPDF. If not, see <https://www.gnu.org/licenses/agpl-3.0.en.html>
17
//
18
// Alternative licensing terms are available from the licensor.
19
// For commercial licensing, see <https://www.artifex.com/> or contact
20
// Artifex Software, Inc., 39 Mesa Street, Suite 108A, San Francisco,
21
// CA 94129, USA, for further information.
22
23
#include "mupdf/fitz.h"
24
25
#include <string.h>
26
27
fz_text *
28
fz_new_text(fz_context *ctx)
29
236k
{
30
236k
  fz_text *text = fz_malloc_struct(ctx, fz_text);
31
236k
  text->refs = 1;
32
236k
  return text;
33
236k
}
34
35
fz_text *
36
fz_keep_text(fz_context *ctx, const fz_text *textc)
37
908
{
38
908
  fz_text *text = (fz_text *)textc; /* Explicit cast away of const */
39
40
908
  return fz_keep_imp(ctx, text, &text->refs);
41
908
}
42
43
void
44
fz_drop_text(fz_context *ctx, const fz_text *textc)
45
273k
{
46
273k
  fz_text *text = (fz_text *)textc; /* Explicit cast away of const */
47
48
273k
  if (fz_drop_imp(ctx, text, &text->refs))
49
236k
  {
50
236k
    fz_text_span *span = text->head;
51
531k
    while (span)
52
295k
    {
53
295k
      fz_text_span *next = span->next;
54
295k
      fz_drop_font(ctx, span->font);
55
295k
      fz_free(ctx, span->items);
56
295k
      fz_free(ctx, span);
57
295k
      span = next;
58
295k
    }
59
236k
    fz_free(ctx, text);
60
236k
  }
61
273k
}
62
63
static fz_text_span *
64
fz_new_text_span(fz_context *ctx, fz_font *font, int wmode, int bidi_level, fz_bidi_direction markup_dir, fz_text_language language, fz_matrix trm)
65
295k
{
66
295k
  fz_text_span *span = fz_malloc_struct(ctx, fz_text_span);
67
295k
  span->font = fz_keep_font(ctx, font);
68
295k
  span->wmode = wmode;
69
295k
  span->bidi_level = bidi_level;
70
295k
  span->markup_dir = markup_dir;
71
295k
  span->language = language;
72
295k
  span->trm = trm;
73
295k
  span->trm.e = 0;
74
295k
  span->trm.f = 0;
75
295k
  return span;
76
295k
}
77
78
static fz_text_span *
79
fz_add_text_span(fz_context *ctx, fz_text *text, fz_font *font, int wmode, int bidi_level, fz_bidi_direction markup_dir, fz_text_language language, fz_matrix trm)
80
6.98M
{
81
6.98M
  if (!text->tail)
82
236k
  {
83
236k
    text->head = text->tail = fz_new_text_span(ctx, font, wmode, bidi_level, markup_dir, language, trm);
84
236k
  }
85
6.74M
  else if (text->tail->font != font ||
86
6.74M
    text->tail->wmode != wmode ||
87
6.74M
    text->tail->bidi_level != bidi_level ||
88
6.74M
    text->tail->markup_dir != markup_dir ||
89
6.74M
    text->tail->language != language ||
90
6.74M
    text->tail->trm.a != trm.a ||
91
6.74M
    text->tail->trm.b != trm.b ||
92
6.74M
    text->tail->trm.c != trm.c ||
93
6.74M
    text->tail->trm.d != trm.d)
94
58.8k
  {
95
58.8k
    text->tail = text->tail->next = fz_new_text_span(ctx, font, wmode, bidi_level, markup_dir, language, trm);
96
58.8k
  }
97
6.98M
  return text->tail;
98
6.98M
}
99
100
static void
101
fz_grow_text_span(fz_context *ctx, fz_text_span *span, int n)
102
6.98M
{
103
6.98M
  int new_cap = span->cap;
104
6.98M
  if (span->len + n < new_cap)
105
6.39M
    return;
106
1.02M
  while (span->len + n > new_cap)
107
438k
    new_cap = new_cap + 36;
108
583k
  span->items = fz_realloc_array(ctx, span->items, new_cap, fz_text_item);
109
583k
  span->cap = new_cap;
110
583k
}
111
112
void
113
fz_show_glyph(fz_context *ctx, fz_text *text, fz_font *font, fz_matrix trm, int gid, int ucs, int wmode, int bidi_level, fz_bidi_direction markup_dir, fz_text_language lang)
114
6.98M
{
115
6.98M
  fz_text_span *span;
116
117
6.98M
  if (text->refs != 1)
118
0
    fz_throw(ctx, FZ_ERROR_GENERIC, "cannot modify shared text objects");
119
120
6.98M
  span = fz_add_text_span(ctx, text, font, wmode, bidi_level, markup_dir, lang, trm);
121
122
6.98M
  fz_grow_text_span(ctx, span, 1);
123
124
6.98M
  span->items[span->len].ucs = ucs;
125
6.98M
  span->items[span->len].gid = gid;
126
6.98M
  span->items[span->len].x = trm.e;
127
6.98M
  span->items[span->len].y = trm.f;
128
6.98M
  span->len++;
129
6.98M
}
130
131
fz_matrix
132
fz_show_string(fz_context *ctx, fz_text *text, fz_font *user_font, fz_matrix trm, const char *s,
133
  int wmode, int bidi_level, fz_bidi_direction markup_dir, fz_text_language language)
134
0
{
135
0
  fz_font *font;
136
0
  int gid, ucs;
137
0
  float adv;
138
139
0
  while (*s)
140
0
  {
141
0
    s += fz_chartorune(&ucs, s);
142
0
    gid = fz_encode_character_with_fallback(ctx, user_font, ucs, 0, language, &font);
143
0
    fz_show_glyph(ctx, text, font, trm, gid, ucs, wmode, bidi_level, markup_dir, language);
144
0
    adv = fz_advance_glyph(ctx, font, gid, wmode);
145
0
    if (wmode == 0)
146
0
      trm = fz_pre_translate(trm, adv, 0);
147
0
    else
148
0
      trm = fz_pre_translate(trm, 0, -adv);
149
0
  }
150
151
0
  return trm;
152
0
}
153
154
fz_matrix
155
fz_measure_string(fz_context *ctx, fz_font *user_font, fz_matrix trm, const char *s,
156
  int wmode, int bidi_level, fz_bidi_direction markup_dir, fz_text_language language)
157
0
{
158
0
  fz_font *font;
159
0
  int gid, ucs;
160
0
  float adv;
161
162
0
  while (*s)
163
0
  {
164
0
    s += fz_chartorune(&ucs, s);
165
0
    gid = fz_encode_character_with_fallback(ctx, user_font, ucs, 0, language, &font);
166
0
    adv = fz_advance_glyph(ctx, font, gid, wmode);
167
0
    if (wmode == 0)
168
0
      trm = fz_pre_translate(trm, adv, 0);
169
0
    else
170
0
      trm = fz_pre_translate(trm, 0, -adv);
171
0
  }
172
173
0
  return trm;
174
0
}
175
176
fz_rect
177
fz_bound_text(fz_context *ctx, const fz_text *text, const fz_stroke_state *stroke, fz_matrix ctm)
178
4.08k
{
179
4.08k
  fz_text_span *span;
180
4.08k
  fz_matrix tm, trm;
181
4.08k
  fz_rect gbox;
182
4.08k
  fz_rect bbox;
183
4.08k
  int i;
184
185
4.08k
  bbox = fz_empty_rect;
186
187
8.47k
  for (span = text->head; span; span = span->next)
188
4.38k
  {
189
4.38k
    if (span->len > 0)
190
4.38k
    {
191
4.38k
      tm = span->trm;
192
377k
      for (i = 0; i < span->len; i++)
193
373k
      {
194
373k
        if (span->items[i].gid >= 0)
195
373k
        {
196
373k
          tm.e = span->items[i].x;
197
373k
          tm.f = span->items[i].y;
198
373k
          trm = fz_concat(tm, ctm);
199
373k
          gbox = fz_bound_glyph(ctx, span->font, span->items[i].gid, trm);
200
373k
          bbox = fz_union_rect(bbox, gbox);
201
373k
        }
202
373k
      }
203
4.38k
    }
204
4.38k
  }
205
206
4.08k
  if (!fz_is_empty_rect(bbox))
207
3.36k
  {
208
3.36k
    if (stroke)
209
251
      bbox = fz_adjust_rect_for_stroke(ctx, bbox, stroke, ctm);
210
211
    /* Compensate for the glyph cache limited positioning precision */
212
3.36k
    bbox.x0 -= 1;
213
3.36k
    bbox.y0 -= 1;
214
3.36k
    bbox.x1 += 1;
215
3.36k
    bbox.y1 += 1;
216
3.36k
  }
217
218
4.08k
  return bbox;
219
4.08k
}
220
221
fz_text_language fz_text_language_from_string(const char *str)
222
123k
{
223
123k
  fz_text_language lang;
224
225
123k
  if (str == NULL)
226
0
    return FZ_LANG_UNSET;
227
228
123k
  if (!strcmp(str, "zh-Hant") ||
229
123k
      !strcmp(str, "zh-HK") ||
230
123k
      !strcmp(str, "zh-MO") ||
231
123k
      !strcmp(str, "zh-SG") ||
232
123k
      !strcmp(str, "zh-TW"))
233
0
    return FZ_LANG_zh_Hant;
234
123k
  if (!strcmp(str, "zh-Hans") ||
235
123k
      !strcmp(str, "zh-CN"))
236
0
    return FZ_LANG_zh_Hans;
237
238
  /* 1st char */
239
123k
  if (str[0] >= 'a' && str[0] <= 'z')
240
11.9k
    lang = str[0] - 'a' + 1;
241
111k
  else if (str[0] >= 'A' && str[0] <= 'Z')
242
36
    lang = str[0] - 'A' + 1;
243
111k
  else
244
111k
    return 0;
245
246
  /* 2nd char */
247
11.9k
  if (str[1] >= 'a' && str[1] <= 'z')
248
11.9k
    lang += 27*(str[1] - 'a' + 1);
249
38
  else if (str[1] >= 'A' && str[1] <= 'Z')
250
36
    lang += 27*(str[1] - 'A' + 1);
251
2
  else
252
2
    return 0; /* There are no valid 1 char language codes */
253
254
  /* 3nd char */
255
11.9k
  if (str[2] >= 'a' && str[2] <= 'z')
256
0
    lang += 27*27*(str[2] - 'a' + 1);
257
11.9k
  else if (str[2] >= 'A' && str[2] <= 'Z')
258
0
    lang += 27*27*(str[2] - 'A' + 1);
259
260
  /* We don't support iso 639-6 4 char codes, cos the standard
261
   * has been withdrawn, and no one uses them. */
262
11.9k
  return lang;
263
11.9k
}
264
265
char *fz_string_from_text_language(char str[8], fz_text_language lang)
266
0
{
267
0
  int c;
268
269
  /* str is supposed to be at least 8 chars in size */
270
0
  if (str == NULL)
271
0
    return NULL;
272
273
0
  if (lang == FZ_LANG_zh_Hant)
274
0
    fz_strlcpy(str, "zh-Hant", 8);
275
0
  else if (lang == FZ_LANG_zh_Hans)
276
0
    fz_strlcpy(str, "zh-Hans", 8);
277
0
  else
278
0
  {
279
0
    c = lang % 27;
280
0
    lang = lang / 27;
281
0
    str[0] = c == 0 ? 0 : c - 1 + 'a';
282
0
    c = lang % 27;
283
0
    lang = lang / 27;
284
0
    str[1] = c == 0 ? 0 : c - 1 + 'a';
285
0
    c = lang % 27;
286
0
    str[2] = c == 0 ? 0 : c - 1 + 'a';
287
0
    str[3] = 0;
288
0
  }
289
290
0
  return str;
291
0
}