1/*
2 * Copyright © 2010,2011,2012 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#ifndef HB_OT_SHAPER_HH
28#define HB_OT_SHAPER_HH
29
30#include "hb.hh"
31
32#include "hb-ot-layout.hh"
33#include "hb-ot-shape.hh"
34#include "hb-ot-shape-normalize.hh"
35
36
37/* buffer var allocations, used by all OT shapers */
38#define ot_shaper_var_u8_category() var2.u8[2]
39#define ot_shaper_var_u8_auxiliary() var2.u8[3]
40
41
42#define HB_OT_SHAPE_MAX_COMBINING_MARKS 32
43
44enum hb_ot_shape_zero_width_marks_type_t {
45 HB_OT_SHAPE_ZERO_WIDTH_MARKS_NONE,
46 HB_OT_SHAPE_ZERO_WIDTH_MARKS_BY_GDEF_EARLY,
47 HB_OT_SHAPE_ZERO_WIDTH_MARKS_BY_GDEF_LATE
48};
49
50
51/* Master OT shaper list */
52#define HB_OT_SHAPERS_IMPLEMENT_SHAPERS \
53 HB_OT_SHAPER_IMPLEMENT (arabic) \
54 HB_OT_SHAPER_IMPLEMENT (default) \
55 HB_OT_SHAPER_IMPLEMENT (dumber) \
56 HB_OT_SHAPER_IMPLEMENT (hangul) \
57 HB_OT_SHAPER_IMPLEMENT (hebrew) \
58 HB_OT_SHAPER_IMPLEMENT (indic) \
59 HB_OT_SHAPER_IMPLEMENT (khmer) \
60 HB_OT_SHAPER_IMPLEMENT (myanmar) \
61 HB_OT_SHAPER_IMPLEMENT (myanmar_zawgyi) \
62 HB_OT_SHAPER_IMPLEMENT (thai) \
63 HB_OT_SHAPER_IMPLEMENT (use) \
64 /* ^--- Add new shapers here; keep sorted. */
65
66
67struct hb_ot_shaper_t
68{
69 /* collect_features()
70 * Called during shape_plan().
71 * Shapers should use plan->map to add their features and callbacks.
72 * May be NULL.
73 */
74 void (*collect_features) (hb_ot_shape_planner_t *plan);
75
76 /* override_features()
77 * Called during shape_plan().
78 * Shapers should use plan->map to override features and add callbacks after
79 * common features are added.
80 * May be NULL.
81 */
82 void (*override_features) (hb_ot_shape_planner_t *plan);
83
84
85 /* data_create()
86 * Called at the end of shape_plan().
87 * Whatever shapers return will be accessible through plan->data later.
88 * If nullptr is returned, means a plan failure.
89 */
90 void *(*data_create) (const hb_ot_shape_plan_t *plan);
91
92 /* data_destroy()
93 * Called when the shape_plan is being destroyed.
94 * plan->data is passed here for destruction.
95 * If nullptr is returned, means a plan failure.
96 * May be NULL.
97 */
98 void (*data_destroy) (void *data);
99
100
101 /* preprocess_text()
102 * Called during shape().
103 * Shapers can use to modify text before shaping starts.
104 * May be NULL.
105 */
106 void (*preprocess_text) (const hb_ot_shape_plan_t *plan,
107 hb_buffer_t *buffer,
108 hb_font_t *font);
109
110 /* postprocess_glyphs()
111 * Called during shape().
112 * Shapers can use to modify glyphs after shaping ends.
113 * May be NULL.
114 */
115 void (*postprocess_glyphs) (const hb_ot_shape_plan_t *plan,
116 hb_buffer_t *buffer,
117 hb_font_t *font);
118
119
120 /* decompose()
121 * Called during shape()'s normalization.
122 * May be NULL.
123 */
124 bool (*decompose) (const hb_ot_shape_normalize_context_t *c,
125 hb_codepoint_t ab,
126 hb_codepoint_t *a,
127 hb_codepoint_t *b);
128
129 /* compose()
130 * Called during shape()'s normalization.
131 * May be NULL.
132 */
133 bool (*compose) (const hb_ot_shape_normalize_context_t *c,
134 hb_codepoint_t a,
135 hb_codepoint_t b,
136 hb_codepoint_t *ab);
137
138 /* setup_masks()
139 * Called during shape().
140 * Shapers should use map to get feature masks and set on buffer.
141 * Shapers may NOT modify characters.
142 * May be NULL.
143 */
144 void (*setup_masks) (const hb_ot_shape_plan_t *plan,
145 hb_buffer_t *buffer,
146 hb_font_t *font);
147
148 /* reorder_marks()
149 * Called during shape().
150 * Shapers can use to modify ordering of combining marks.
151 * May be NULL.
152 */
153 void (*reorder_marks) (const hb_ot_shape_plan_t *plan,
154 hb_buffer_t *buffer,
155 unsigned int start,
156 unsigned int end);
157
158 /* gpos_tag()
159 * If not HB_TAG_NONE, then must match found GPOS script tag for
160 * GPOS to be applied. Otherwise, fallback positioning will be used.
161 */
162 hb_tag_t gpos_tag;
163
164 hb_ot_shape_normalization_mode_t normalization_preference;
165
166 hb_ot_shape_zero_width_marks_type_t zero_width_marks;
167
168 bool fallback_position;
169};
170
171#define HB_OT_SHAPER_IMPLEMENT(name) extern HB_INTERNAL const hb_ot_shaper_t _hb_ot_shaper_##name;
172HB_OT_SHAPERS_IMPLEMENT_SHAPERS
173#undef HB_OT_SHAPER_IMPLEMENT
174
175
176static inline const hb_ot_shaper_t *
177hb_ot_shaper_categorize (const hb_ot_shape_planner_t *planner)
178{
179 switch ((hb_tag_t) planner->props.script)
180 {
181 default:
182 return &_hb_ot_shaper_default;
183
184
185 /* Unicode-1.1 additions */
186 case HB_SCRIPT_ARABIC:
187
188 /* Unicode-3.0 additions */
189 case HB_SCRIPT_SYRIAC:
190
191 /* For Arabic script, use the Arabic shaper even if no OT script tag was found.
192 * This is because we do fallback shaping for Arabic script (and not others).
193 * But note that Arabic shaping is applicable only to horizontal layout; for
194 * vertical text, just use the generic shaper instead. */
195 if ((planner->map.chosen_script[0] != HB_OT_TAG_DEFAULT_SCRIPT ||
196 planner->props.script == HB_SCRIPT_ARABIC) &&
197 HB_DIRECTION_IS_HORIZONTAL(planner->props.direction))
198 return &_hb_ot_shaper_arabic;
199 else
200 return &_hb_ot_shaper_default;
201
202
203 /* Unicode-1.1 additions */
204 case HB_SCRIPT_THAI:
205 case HB_SCRIPT_LAO:
206
207 return &_hb_ot_shaper_thai;
208
209
210 /* Unicode-1.1 additions */
211 case HB_SCRIPT_HANGUL:
212
213 return &_hb_ot_shaper_hangul;
214
215
216 /* Unicode-1.1 additions */
217 case HB_SCRIPT_HEBREW:
218
219 return &_hb_ot_shaper_hebrew;
220
221
222 /* Unicode-1.1 additions */
223 case HB_SCRIPT_BENGALI:
224 case HB_SCRIPT_DEVANAGARI:
225 case HB_SCRIPT_GUJARATI:
226 case HB_SCRIPT_GURMUKHI:
227 case HB_SCRIPT_KANNADA:
228 case HB_SCRIPT_MALAYALAM:
229 case HB_SCRIPT_ORIYA:
230 case HB_SCRIPT_TAMIL:
231 case HB_SCRIPT_TELUGU:
232
233 /* If the designer designed the font for the 'DFLT' script,
234 * (or we ended up arbitrarily pick 'latn'), use the default shaper.
235 * Otherwise, use the specific shaper.
236 *
237 * If it's indy3 tag, send to USE. */
238 if (planner->map.chosen_script[0] == HB_TAG ('D','F','L','T') ||
239 planner->map.chosen_script[0] == HB_TAG ('l','a','t','n'))
240 return &_hb_ot_shaper_default;
241 else if ((planner->map.chosen_script[0] & 0x000000FF) == '3')
242 return &_hb_ot_shaper_use;
243 else
244 return &_hb_ot_shaper_indic;
245
246 case HB_SCRIPT_KHMER:
247 return &_hb_ot_shaper_khmer;
248
249 case HB_SCRIPT_MYANMAR:
250 /* If the designer designed the font for the 'DFLT' script,
251 * (or we ended up arbitrarily pick 'latn'), use the default shaper.
252 * Otherwise, use the specific shaper.
253 *
254 * If designer designed for 'mymr' tag, also send to default
255 * shaper. That's tag used from before Myanmar shaping spec
256 * was developed. The shaping spec uses 'mym2' tag. */
257 if (planner->map.chosen_script[0] == HB_TAG ('D','F','L','T') ||
258 planner->map.chosen_script[0] == HB_TAG ('l','a','t','n') ||
259 planner->map.chosen_script[0] == HB_TAG ('m','y','m','r'))
260 return &_hb_ot_shaper_default;
261 else
262 return &_hb_ot_shaper_myanmar;
263
264
265#ifndef HB_NO_OT_SHAPER_MYANMAR_ZAWGYI
266#define HB_SCRIPT_MYANMAR_ZAWGYI ((hb_script_t) HB_TAG ('Q','a','a','g'))
267 case HB_SCRIPT_MYANMAR_ZAWGYI:
268 /* https://github.com/harfbuzz/harfbuzz/issues/1162 */
269
270 return &_hb_ot_shaper_myanmar_zawgyi;
271#endif
272
273
274 /* Unicode-2.0 additions */
275 case HB_SCRIPT_TIBETAN:
276
277 /* Unicode-3.0 additions */
278 case HB_SCRIPT_MONGOLIAN:
279 case HB_SCRIPT_SINHALA:
280
281 /* Unicode-3.2 additions */
282 case HB_SCRIPT_BUHID:
283 case HB_SCRIPT_HANUNOO:
284 case HB_SCRIPT_TAGALOG:
285 case HB_SCRIPT_TAGBANWA:
286
287 /* Unicode-4.0 additions */
288 case HB_SCRIPT_LIMBU:
289 case HB_SCRIPT_TAI_LE:
290
291 /* Unicode-4.1 additions */
292 case HB_SCRIPT_BUGINESE:
293 case HB_SCRIPT_KHAROSHTHI:
294 case HB_SCRIPT_SYLOTI_NAGRI:
295 case HB_SCRIPT_TIFINAGH:
296
297 /* Unicode-5.0 additions */
298 case HB_SCRIPT_BALINESE:
299 case HB_SCRIPT_NKO:
300 case HB_SCRIPT_PHAGS_PA:
301
302 /* Unicode-5.1 additions */
303 case HB_SCRIPT_CHAM:
304 case HB_SCRIPT_KAYAH_LI:
305 case HB_SCRIPT_LEPCHA:
306 case HB_SCRIPT_REJANG:
307 case HB_SCRIPT_SAURASHTRA:
308 case HB_SCRIPT_SUNDANESE:
309
310 /* Unicode-5.2 additions */
311 case HB_SCRIPT_EGYPTIAN_HIEROGLYPHS:
312 case HB_SCRIPT_JAVANESE:
313 case HB_SCRIPT_KAITHI:
314 case HB_SCRIPT_MEETEI_MAYEK:
315 case HB_SCRIPT_TAI_THAM:
316 case HB_SCRIPT_TAI_VIET:
317
318 /* Unicode-6.0 additions */
319 case HB_SCRIPT_BATAK:
320 case HB_SCRIPT_BRAHMI:
321 case HB_SCRIPT_MANDAIC:
322
323 /* Unicode-6.1 additions */
324 case HB_SCRIPT_CHAKMA:
325 case HB_SCRIPT_MIAO:
326 case HB_SCRIPT_SHARADA:
327 case HB_SCRIPT_TAKRI:
328
329 /* Unicode-7.0 additions */
330 case HB_SCRIPT_DUPLOYAN:
331 case HB_SCRIPT_GRANTHA:
332 case HB_SCRIPT_KHOJKI:
333 case HB_SCRIPT_KHUDAWADI:
334 case HB_SCRIPT_MAHAJANI:
335 case HB_SCRIPT_MANICHAEAN:
336 case HB_SCRIPT_MODI:
337 case HB_SCRIPT_PAHAWH_HMONG:
338 case HB_SCRIPT_PSALTER_PAHLAVI:
339 case HB_SCRIPT_SIDDHAM:
340 case HB_SCRIPT_TIRHUTA:
341
342 /* Unicode-8.0 additions */
343 case HB_SCRIPT_AHOM:
344 case HB_SCRIPT_MULTANI:
345
346 /* Unicode-9.0 additions */
347 case HB_SCRIPT_ADLAM:
348 case HB_SCRIPT_BHAIKSUKI:
349 case HB_SCRIPT_MARCHEN:
350 case HB_SCRIPT_NEWA:
351
352 /* Unicode-10.0 additions */
353 case HB_SCRIPT_MASARAM_GONDI:
354 case HB_SCRIPT_SOYOMBO:
355 case HB_SCRIPT_ZANABAZAR_SQUARE:
356
357 /* Unicode-11.0 additions */
358 case HB_SCRIPT_DOGRA:
359 case HB_SCRIPT_GUNJALA_GONDI:
360 case HB_SCRIPT_HANIFI_ROHINGYA:
361 case HB_SCRIPT_MAKASAR:
362 case HB_SCRIPT_MEDEFAIDRIN:
363 case HB_SCRIPT_OLD_SOGDIAN:
364 case HB_SCRIPT_SOGDIAN:
365
366 /* Unicode-12.0 additions */
367 case HB_SCRIPT_ELYMAIC:
368 case HB_SCRIPT_NANDINAGARI:
369 case HB_SCRIPT_NYIAKENG_PUACHUE_HMONG:
370 case HB_SCRIPT_WANCHO:
371
372 /* Unicode-13.0 additions */
373 case HB_SCRIPT_CHORASMIAN:
374 case HB_SCRIPT_DIVES_AKURU:
375 case HB_SCRIPT_KHITAN_SMALL_SCRIPT:
376 case HB_SCRIPT_YEZIDI:
377
378 /* Unicode-14.0 additions */
379 case HB_SCRIPT_CYPRO_MINOAN:
380 case HB_SCRIPT_OLD_UYGHUR:
381 case HB_SCRIPT_TANGSA:
382 case HB_SCRIPT_TOTO:
383 case HB_SCRIPT_VITHKUQI:
384
385 /* Unicode-15.0 additions */
386 case HB_SCRIPT_KAWI:
387 case HB_SCRIPT_NAG_MUNDARI:
388
389 /* If the designer designed the font for the 'DFLT' script,
390 * (or we ended up arbitrarily pick 'latn'), use the default shaper.
391 * Otherwise, use the specific shaper.
392 * Note that for some simple scripts, there may not be *any*
393 * GSUB/GPOS needed, so there may be no scripts found! */
394 if (planner->map.chosen_script[0] == HB_TAG ('D','F','L','T') ||
395 planner->map.chosen_script[0] == HB_TAG ('l','a','t','n'))
396 return &_hb_ot_shaper_default;
397 else
398 return &_hb_ot_shaper_use;
399 }
400}
401
402
403#endif /* HB_OT_SHAPER_HH */
404