Coverage Report

Created: 2025-01-28 06:17

/src/mupdf/source/fitz/text.c
Line
Count
Source (jump to first uncovered line)
1
// Copyright (C) 2004-2024 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
156k
{
30
156k
  fz_text *text = fz_malloc_struct(ctx, fz_text);
31
156k
  text->refs = 1;
32
156k
  return text;
33
156k
}
34
35
fz_text *
36
fz_keep_text(fz_context *ctx, const fz_text *textc)
37
936
{
38
936
  fz_text *text = (fz_text *)textc; /* Explicit cast away of const */
39
40
936
  return fz_keep_imp(ctx, text, &text->refs);
41
936
}
42
43
void
44
fz_drop_text(fz_context *ctx, const fz_text *textc)
45
188k
{
46
188k
  fz_text *text = (fz_text *)textc; /* Explicit cast away of const */
47
48
188k
  if (fz_drop_imp(ctx, text, &text->refs))
49
156k
  {
50
156k
    fz_text_span *span = text->head;
51
448k
    while (span)
52
292k
    {
53
292k
      fz_text_span *next = span->next;
54
292k
      fz_drop_font(ctx, span->font);
55
292k
      fz_free(ctx, span->items);
56
292k
      fz_free(ctx, span);
57
292k
      span = next;
58
292k
    }
59
156k
    fz_free(ctx, text);
60
156k
  }
61
188k
}
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
292k
{
66
292k
  fz_text_span *span = fz_malloc_struct(ctx, fz_text_span);
67
292k
  span->font = fz_keep_font(ctx, font);
68
292k
  span->wmode = wmode;
69
292k
  span->bidi_level = bidi_level;
70
292k
  span->markup_dir = markup_dir;
71
292k
  span->language = language;
72
292k
  span->trm = trm;
73
292k
  span->trm.e = 0;
74
292k
  span->trm.f = 0;
75
292k
  return span;
76
292k
}
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
5.91M
{
81
5.91M
  if (!text->tail)
82
156k
  {
83
156k
    text->head = text->tail = fz_new_text_span(ctx, font, wmode, bidi_level, markup_dir, language, trm);
84
156k
  }
85
5.76M
  else if (text->tail->font != font ||
86
5.76M
    text->tail->wmode != wmode ||
87
5.76M
    text->tail->bidi_level != bidi_level ||
88
5.76M
    text->tail->markup_dir != markup_dir ||
89
5.76M
    text->tail->language != language ||
90
5.76M
    text->tail->trm.a != trm.a ||
91
5.76M
    text->tail->trm.b != trm.b ||
92
5.76M
    text->tail->trm.c != trm.c ||
93
5.76M
    text->tail->trm.d != trm.d)
94
135k
  {
95
135k
    text->tail = text->tail->next = fz_new_text_span(ctx, font, wmode, bidi_level, markup_dir, language, trm);
96
135k
  }
97
5.91M
  return text->tail;
98
5.91M
}
99
100
static void
101
fz_grow_text_span(fz_context *ctx, fz_text_span *span, int n)
102
5.91M
{
103
5.91M
  int new_cap = span->cap;
104
5.91M
  if (span->len + n < new_cap)
105
5.38M
    return;
106
945k
  while (span->len + n > new_cap)
107
412k
    new_cap = new_cap + 36;
108
532k
  span->items = fz_realloc_array(ctx, span->items, new_cap, fz_text_item);
109
532k
  span->cap = new_cap;
110
532k
}
111
112
void
113
fz_show_glyph_aux(fz_context *ctx, fz_text *text, fz_font *font, fz_matrix trm, float adv, int gid, int ucs, int cid, int wmode, int bidi_level, fz_bidi_direction markup_dir, fz_text_language lang)
114
5.91M
{
115
5.91M
  fz_text_span *span;
116
117
5.91M
  if (text->refs != 1)
118
0
    fz_throw(ctx, FZ_ERROR_ARGUMENT, "cannot modify shared text objects");
119
120
5.91M
  span = fz_add_text_span(ctx, text, font, wmode, bidi_level, markup_dir, lang, trm);
121
122
5.91M
  fz_grow_text_span(ctx, span, 1);
123
124
5.91M
  span->items[span->len].ucs = ucs;
125
5.91M
  span->items[span->len].gid = gid;
126
5.91M
  span->items[span->len].cid = cid;
127
5.91M
  span->items[span->len].x = trm.e;
128
5.91M
  span->items[span->len].y = trm.f;
129
5.91M
  span->items[span->len].adv = adv;
130
5.91M
  span->len++;
131
5.91M
}
132
133
void
134
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)
135
91.9k
{
136
91.9k
  float adv = (gid >= 0) ? fz_advance_glyph(ctx, font, gid, wmode) : 0;
137
91.9k
  fz_show_glyph_aux(ctx, text, font, trm, adv, gid, ucs, ucs, wmode, bidi_level, markup_dir, lang);
138
91.9k
}
139
140
fz_matrix
141
fz_show_string(fz_context *ctx, fz_text *text, fz_font *user_font, fz_matrix trm, const char *s,
142
  int wmode, int bidi_level, fz_bidi_direction markup_dir, fz_text_language language)
143
295
{
144
295
  fz_font *font;
145
295
  int gid, ucs;
146
295
  float adv;
147
148
61.8k
  while (*s)
149
61.5k
  {
150
61.5k
    s += fz_chartorune(&ucs, s);
151
61.5k
    gid = fz_encode_character_with_fallback(ctx, user_font, ucs, 0, language, &font);
152
61.5k
    if (gid >= 0)
153
61.5k
      adv = fz_advance_glyph(ctx, font, gid, wmode);
154
0
    else
155
0
      adv = 0;
156
61.5k
    fz_show_glyph_aux(ctx, text, font, trm, adv, gid, ucs, ucs, wmode, bidi_level, markup_dir, language);
157
61.5k
    if (wmode == 0)
158
61.5k
      trm = fz_pre_translate(trm, adv, 0);
159
0
    else
160
0
      trm = fz_pre_translate(trm, 0, -adv);
161
61.5k
  }
162
163
295
  return trm;
164
295
}
165
166
fz_matrix
167
fz_measure_string(fz_context *ctx, fz_font *user_font, fz_matrix trm, const char *s,
168
  int wmode, int bidi_level, fz_bidi_direction markup_dir, fz_text_language language)
169
0
{
170
0
  fz_font *font;
171
0
  int gid, ucs;
172
0
  float adv;
173
174
0
  while (*s)
175
0
  {
176
0
    s += fz_chartorune(&ucs, s);
177
0
    gid = fz_encode_character_with_fallback(ctx, user_font, ucs, 0, language, &font);
178
0
    adv = fz_advance_glyph(ctx, font, gid, wmode);
179
0
    if (wmode == 0)
180
0
      trm = fz_pre_translate(trm, adv, 0);
181
0
    else
182
0
      trm = fz_pre_translate(trm, 0, -adv);
183
0
  }
184
185
0
  return trm;
186
0
}
187
188
fz_rect
189
fz_bound_text(fz_context *ctx, const fz_text *text, const fz_stroke_state *stroke, fz_matrix ctm)
190
2.65k
{
191
2.65k
  fz_text_span *span;
192
2.65k
  fz_matrix tm, trm;
193
2.65k
  fz_rect gbox;
194
2.65k
  fz_rect bbox;
195
2.65k
  int i;
196
197
2.65k
  bbox = fz_empty_rect;
198
199
5.43k
  for (span = text->head; span; span = span->next)
200
2.78k
  {
201
2.78k
    if (span->len > 0)
202
2.78k
    {
203
2.78k
      tm = span->trm;
204
248k
      for (i = 0; i < span->len; i++)
205
245k
      {
206
245k
        if (span->items[i].gid >= 0)
207
245k
        {
208
245k
          tm.e = span->items[i].x;
209
245k
          tm.f = span->items[i].y;
210
245k
          trm = fz_concat(tm, ctm);
211
245k
          gbox = fz_bound_glyph(ctx, span->font, span->items[i].gid, trm);
212
245k
          bbox = fz_union_rect(bbox, gbox);
213
245k
        }
214
245k
      }
215
2.78k
    }
216
2.78k
  }
217
218
2.65k
  if (!fz_is_empty_rect(bbox))
219
2.08k
  {
220
2.08k
    if (stroke)
221
147
      bbox = fz_adjust_rect_for_stroke(ctx, bbox, stroke, ctm);
222
223
    /* Compensate for the glyph cache limited positioning precision */
224
2.08k
    bbox.x0 -= 1;
225
2.08k
    bbox.y0 -= 1;
226
2.08k
    bbox.x1 += 1;
227
2.08k
    bbox.y1 += 1;
228
2.08k
  }
229
230
2.65k
  return bbox;
231
2.65k
}
232
233
fz_text_language fz_text_language_from_string(const char *str)
234
81.4k
{
235
81.4k
  fz_text_language lang;
236
237
81.4k
  if (str == NULL)
238
0
    return FZ_LANG_UNSET;
239
240
81.4k
  if (!strcmp(str, "zh-Hant") ||
241
81.4k
      !strcmp(str, "zh-HK") ||
242
81.4k
      !strcmp(str, "zh-MO") ||
243
81.4k
      !strcmp(str, "zh-SG") ||
244
81.4k
      !strcmp(str, "zh-TW"))
245
0
    return FZ_LANG_zh_Hant;
246
81.4k
  if (!strcmp(str, "zh-Hans") ||
247
81.4k
      !strcmp(str, "zh-CN"))
248
0
    return FZ_LANG_zh_Hans;
249
250
  /* 1st char */
251
81.4k
  if (str[0] >= 'a' && str[0] <= 'z')
252
4.71k
    lang = str[0] - 'a' + 1;
253
76.7k
  else if (str[0] >= 'A' && str[0] <= 'Z')
254
48
    lang = str[0] - 'A' + 1;
255
76.7k
  else
256
76.7k
    return 0;
257
258
  /* 2nd char */
259
4.76k
  if (str[1] >= 'a' && str[1] <= 'z')
260
4.71k
    lang += 27*(str[1] - 'a' + 1);
261
48
  else if (str[1] >= 'A' && str[1] <= 'Z')
262
48
    lang += 27*(str[1] - 'A' + 1);
263
0
  else
264
0
    return 0; /* There are no valid 1 char language codes */
265
266
  /* 3nd char */
267
4.76k
  if (str[2] >= 'a' && str[2] <= 'z')
268
0
    lang += 27*27*(str[2] - 'a' + 1);
269
4.76k
  else if (str[2] >= 'A' && str[2] <= 'Z')
270
0
    lang += 27*27*(str[2] - 'A' + 1);
271
272
  /* We don't support iso 639-6 4 char codes, cos the standard
273
   * has been withdrawn, and no one uses them. */
274
4.76k
  return lang;
275
4.76k
}
276
277
char *fz_string_from_text_language(char str[8], fz_text_language lang)
278
0
{
279
0
  int c;
280
281
  /* str is supposed to be at least 8 chars in size */
282
0
  if (str == NULL)
283
0
    return NULL;
284
285
0
  if (lang == FZ_LANG_zh_Hant)
286
0
    fz_strlcpy(str, "zh-Hant", 8);
287
0
  else if (lang == FZ_LANG_zh_Hans)
288
0
    fz_strlcpy(str, "zh-Hans", 8);
289
0
  else
290
0
  {
291
0
    c = lang % 27;
292
0
    lang = lang / 27;
293
0
    str[0] = c == 0 ? 0 : c - 1 + 'a';
294
0
    c = lang % 27;
295
0
    lang = lang / 27;
296
0
    str[1] = c == 0 ? 0 : c - 1 + 'a';
297
0
    c = lang % 27;
298
0
    str[2] = c == 0 ? 0 : c - 1 + 'a';
299
0
    str[3] = 0;
300
0
  }
301
302
0
  return str;
303
0
}