hb-ot-shaper.hh (11979B)
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 44 enum 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 67 struct 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; 172 HB_OT_SHAPERS_IMPLEMENT_SHAPERS 173 #undef HB_OT_SHAPER_IMPLEMENT 174 175 176 static inline const hb_ot_shaper_t * 177 hb_ot_shaper_categorize (hb_script_t script, 178 hb_direction_t direction, 179 hb_tag_t gsub_script) 180 { 181 switch ((hb_tag_t) script) 182 { 183 default: 184 return &_hb_ot_shaper_default; 185 186 187 /* Unicode-1.1 additions */ 188 case HB_SCRIPT_ARABIC: 189 190 /* Unicode-3.0 additions */ 191 case HB_SCRIPT_SYRIAC: 192 193 /* For Arabic script, use the Arabic shaper even if no OT script tag was found. 194 * This is because we do fallback shaping for Arabic script (and not others). 195 * But note that Arabic shaping is applicable only to horizontal layout; for 196 * vertical text, just use the generic shaper instead. */ 197 if ((gsub_script != HB_OT_TAG_DEFAULT_SCRIPT || script == HB_SCRIPT_ARABIC) && 198 HB_DIRECTION_IS_HORIZONTAL (direction)) 199 return &_hb_ot_shaper_arabic; 200 else 201 return &_hb_ot_shaper_default; 202 203 204 /* Unicode-1.1 additions */ 205 case HB_SCRIPT_THAI: 206 case HB_SCRIPT_LAO: 207 208 return &_hb_ot_shaper_thai; 209 210 211 /* Unicode-1.1 additions */ 212 case HB_SCRIPT_HANGUL: 213 214 return &_hb_ot_shaper_hangul; 215 216 217 /* Unicode-1.1 additions */ 218 case HB_SCRIPT_HEBREW: 219 220 return &_hb_ot_shaper_hebrew; 221 222 223 /* Unicode-1.1 additions */ 224 case HB_SCRIPT_BENGALI: 225 case HB_SCRIPT_DEVANAGARI: 226 case HB_SCRIPT_GUJARATI: 227 case HB_SCRIPT_GURMUKHI: 228 case HB_SCRIPT_KANNADA: 229 case HB_SCRIPT_MALAYALAM: 230 case HB_SCRIPT_ORIYA: 231 case HB_SCRIPT_TAMIL: 232 case HB_SCRIPT_TELUGU: 233 234 /* If the designer designed the font for the 'DFLT' script, 235 * (or we ended up arbitrarily pick 'latn'), use the default shaper. 236 * Otherwise, use the specific shaper. 237 * 238 * If it's indy3 tag, send to USE. */ 239 if (gsub_script == HB_TAG ('D','F','L','T') || 240 gsub_script == HB_TAG ('l','a','t','n')) 241 return &_hb_ot_shaper_default; 242 else if ((gsub_script & 0x000000FF) == '3') 243 return &_hb_ot_shaper_use; 244 else 245 return &_hb_ot_shaper_indic; 246 247 case HB_SCRIPT_KHMER: 248 return &_hb_ot_shaper_khmer; 249 250 case HB_SCRIPT_MYANMAR: 251 /* If the designer designed the font for the 'DFLT' script, 252 * (or we ended up arbitrarily pick 'latn'), use the default shaper. 253 * Otherwise, use the specific shaper. 254 * 255 * If designer designed for 'mymr' tag, also send to default 256 * shaper. That's tag used from before Myanmar shaping spec 257 * was developed. The shaping spec uses 'mym2' tag. */ 258 if (gsub_script == HB_TAG ('D','F','L','T') || 259 gsub_script == HB_TAG ('l','a','t','n') || 260 gsub_script == HB_TAG ('m','y','m','r')) 261 return &_hb_ot_shaper_default; 262 else 263 return &_hb_ot_shaper_myanmar; 264 265 266 #ifndef HB_NO_OT_SHAPER_MYANMAR_ZAWGYI 267 #define HB_SCRIPT_MYANMAR_ZAWGYI ((hb_script_t) HB_TAG ('Q','a','a','g')) 268 case HB_SCRIPT_MYANMAR_ZAWGYI: 269 /* https://github.com/harfbuzz/harfbuzz/issues/1162 */ 270 271 return &_hb_ot_shaper_myanmar_zawgyi; 272 #endif 273 274 275 /* Unicode-2.0 additions */ 276 case HB_SCRIPT_TIBETAN: 277 278 /* Unicode-3.0 additions */ 279 case HB_SCRIPT_MONGOLIAN: 280 case HB_SCRIPT_SINHALA: 281 282 /* Unicode-3.2 additions */ 283 case HB_SCRIPT_BUHID: 284 case HB_SCRIPT_HANUNOO: 285 case HB_SCRIPT_TAGALOG: 286 case HB_SCRIPT_TAGBANWA: 287 288 /* Unicode-4.0 additions */ 289 case HB_SCRIPT_LIMBU: 290 case HB_SCRIPT_TAI_LE: 291 292 /* Unicode-4.1 additions */ 293 case HB_SCRIPT_BUGINESE: 294 case HB_SCRIPT_KHAROSHTHI: 295 case HB_SCRIPT_SYLOTI_NAGRI: 296 case HB_SCRIPT_TIFINAGH: 297 298 /* Unicode-5.0 additions */ 299 case HB_SCRIPT_BALINESE: 300 case HB_SCRIPT_NKO: 301 case HB_SCRIPT_PHAGS_PA: 302 303 /* Unicode-5.1 additions */ 304 case HB_SCRIPT_CHAM: 305 case HB_SCRIPT_KAYAH_LI: 306 case HB_SCRIPT_LEPCHA: 307 case HB_SCRIPT_REJANG: 308 case HB_SCRIPT_SAURASHTRA: 309 case HB_SCRIPT_SUNDANESE: 310 311 /* Unicode-5.2 additions */ 312 case HB_SCRIPT_EGYPTIAN_HIEROGLYPHS: 313 case HB_SCRIPT_JAVANESE: 314 case HB_SCRIPT_KAITHI: 315 case HB_SCRIPT_MEETEI_MAYEK: 316 case HB_SCRIPT_TAI_THAM: 317 case HB_SCRIPT_TAI_VIET: 318 319 /* Unicode-6.0 additions */ 320 case HB_SCRIPT_BATAK: 321 case HB_SCRIPT_BRAHMI: 322 case HB_SCRIPT_MANDAIC: 323 324 /* Unicode-6.1 additions */ 325 case HB_SCRIPT_CHAKMA: 326 case HB_SCRIPT_MIAO: 327 case HB_SCRIPT_SHARADA: 328 case HB_SCRIPT_TAKRI: 329 330 /* Unicode-7.0 additions */ 331 case HB_SCRIPT_DUPLOYAN: 332 case HB_SCRIPT_GRANTHA: 333 case HB_SCRIPT_KHOJKI: 334 case HB_SCRIPT_KHUDAWADI: 335 case HB_SCRIPT_MAHAJANI: 336 case HB_SCRIPT_MANICHAEAN: 337 case HB_SCRIPT_MODI: 338 case HB_SCRIPT_PAHAWH_HMONG: 339 case HB_SCRIPT_PSALTER_PAHLAVI: 340 case HB_SCRIPT_SIDDHAM: 341 case HB_SCRIPT_TIRHUTA: 342 343 /* Unicode-8.0 additions */ 344 case HB_SCRIPT_AHOM: 345 case HB_SCRIPT_MULTANI: 346 347 /* Unicode-9.0 additions */ 348 case HB_SCRIPT_ADLAM: 349 case HB_SCRIPT_BHAIKSUKI: 350 case HB_SCRIPT_MARCHEN: 351 case HB_SCRIPT_NEWA: 352 353 /* Unicode-10.0 additions */ 354 case HB_SCRIPT_MASARAM_GONDI: 355 case HB_SCRIPT_SOYOMBO: 356 case HB_SCRIPT_ZANABAZAR_SQUARE: 357 358 /* Unicode-11.0 additions */ 359 case HB_SCRIPT_DOGRA: 360 case HB_SCRIPT_GUNJALA_GONDI: 361 case HB_SCRIPT_HANIFI_ROHINGYA: 362 case HB_SCRIPT_MAKASAR: 363 case HB_SCRIPT_MEDEFAIDRIN: 364 case HB_SCRIPT_OLD_SOGDIAN: 365 case HB_SCRIPT_SOGDIAN: 366 367 /* Unicode-12.0 additions */ 368 case HB_SCRIPT_ELYMAIC: 369 case HB_SCRIPT_NANDINAGARI: 370 case HB_SCRIPT_NYIAKENG_PUACHUE_HMONG: 371 case HB_SCRIPT_WANCHO: 372 373 /* Unicode-13.0 additions */ 374 case HB_SCRIPT_CHORASMIAN: 375 case HB_SCRIPT_DIVES_AKURU: 376 case HB_SCRIPT_KHITAN_SMALL_SCRIPT: 377 case HB_SCRIPT_YEZIDI: 378 379 /* Unicode-14.0 additions */ 380 case HB_SCRIPT_CYPRO_MINOAN: 381 case HB_SCRIPT_OLD_UYGHUR: 382 case HB_SCRIPT_TANGSA: 383 case HB_SCRIPT_TOTO: 384 case HB_SCRIPT_VITHKUQI: 385 386 /* Unicode-15.0 additions */ 387 case HB_SCRIPT_KAWI: 388 case HB_SCRIPT_NAG_MUNDARI: 389 390 /* Unicode-16.0 additions */ 391 case HB_SCRIPT_GARAY: 392 case HB_SCRIPT_GURUNG_KHEMA: 393 case HB_SCRIPT_KIRAT_RAI: 394 case HB_SCRIPT_OL_ONAL: 395 case HB_SCRIPT_SUNUWAR: 396 case HB_SCRIPT_TODHRI: 397 case HB_SCRIPT_TULU_TIGALARI: 398 399 /* Unicode-17.0 additions */ 400 case HB_SCRIPT_BERIA_ERFE: 401 case HB_SCRIPT_SIDETIC: 402 case HB_SCRIPT_TAI_YO: 403 case HB_SCRIPT_TOLONG_SIKI: 404 405 /* If the designer designed the font for the 'DFLT' script, 406 * (or we ended up arbitrarily pick 'latn'), use the default shaper. 407 * Otherwise, use the specific shaper. 408 * Note that for some simple scripts, there may not be *any* 409 * GSUB/GPOS needed, so there may be no scripts found! */ 410 if (gsub_script == HB_TAG ('D','F','L','T') || 411 gsub_script == HB_TAG ('l','a','t','n')) 412 return &_hb_ot_shaper_default; 413 else 414 return &_hb_ot_shaper_use; 415 } 416 } 417 418 419 #endif /* HB_OT_SHAPER_HH */