| 1 | /* |
| 2 | * Copyright © 2009 Red Hat, Inc. |
| 3 | * Copyright © 2011 Google, Inc. |
| 4 | * |
| 5 | * This is part of HarfBuzz, a text shaping library. |
| 6 | * |
| 7 | * Permission is hereby granted, without written agreement and without |
| 8 | * license or royalty fees, to use, copy, modify, and distribute this |
| 9 | * software and its documentation for any purpose, provided that the |
| 10 | * above copyright notice and the following two paragraphs appear in |
| 11 | * all copies of this software. |
| 12 | * |
| 13 | * IN NO EVENT SHALL THE COPYRIGHT HOLDER BE LIABLE TO ANY PARTY FOR |
| 14 | * DIRECT, INDIRECT, SPECIAL, INCIDENTAL, OR CONSEQUENTIAL DAMAGES |
| 15 | * ARISING OUT OF THE USE OF THIS SOFTWARE AND ITS DOCUMENTATION, EVEN |
| 16 | * IF THE COPYRIGHT HOLDER HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH |
| 17 | * DAMAGE. |
| 18 | * |
| 19 | * THE COPYRIGHT HOLDER SPECIFICALLY DISCLAIMS ANY WARRANTIES, INCLUDING, |
| 20 | * BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND |
| 21 | * FITNESS FOR A PARTICULAR PURPOSE. THE SOFTWARE PROVIDED HEREUNDER IS |
| 22 | * ON AN "AS IS" BASIS, AND THE COPYRIGHT HOLDER HAS NO OBLIGATION TO |
| 23 | * PROVIDE MAINTENANCE, SUPPORT, UPDATES, ENHANCEMENTS, OR MODIFICATIONS. |
| 24 | * |
| 25 | * Red Hat Author(s): Behdad Esfahbod |
| 26 | * Google Author(s): Behdad Esfahbod, Roozbeh Pournader |
| 27 | */ |
| 28 | |
| 29 | #include "hb.hh" |
| 30 | |
| 31 | #ifndef HB_NO_OT_TAG |
| 32 | |
| 33 | |
| 34 | /* hb_script_t */ |
| 35 | |
| 36 | static hb_tag_t |
| 37 | hb_ot_old_tag_from_script (hb_script_t script) |
| 38 | { |
| 39 | /* This seems to be accurate as of end of 2012. */ |
| 40 | |
| 41 | switch ((hb_tag_t) script) |
| 42 | { |
| 43 | case HB_SCRIPT_INVALID: return HB_OT_TAG_DEFAULT_SCRIPT; |
| 44 | case HB_SCRIPT_MATH: return HB_OT_TAG_MATH_SCRIPT; |
| 45 | |
| 46 | /* KATAKANA and HIRAGANA both map to 'kana' */ |
| 47 | case HB_SCRIPT_HIRAGANA: return HB_TAG('k','a','n','a'); |
| 48 | |
| 49 | /* Spaces at the end are preserved, unlike ISO 15924 */ |
| 50 | case HB_SCRIPT_LAO: return HB_TAG('l','a','o',' '); |
| 51 | case HB_SCRIPT_YI: return HB_TAG('y','i',' ',' '); |
| 52 | /* Unicode-5.0 additions */ |
| 53 | case HB_SCRIPT_NKO: return HB_TAG('n','k','o',' '); |
| 54 | /* Unicode-5.1 additions */ |
| 55 | case HB_SCRIPT_VAI: return HB_TAG('v','a','i',' '); |
| 56 | } |
| 57 | |
| 58 | /* Else, just change first char to lowercase and return */ |
| 59 | return ((hb_tag_t) script) | 0x20000000u; |
| 60 | } |
| 61 | |
| 62 | static hb_script_t |
| 63 | hb_ot_old_tag_to_script (hb_tag_t tag) |
| 64 | { |
| 65 | if (unlikely (tag == HB_OT_TAG_DEFAULT_SCRIPT)) |
| 66 | return HB_SCRIPT_INVALID; |
| 67 | if (unlikely (tag == HB_OT_TAG_MATH_SCRIPT)) |
| 68 | return HB_SCRIPT_MATH; |
| 69 | |
| 70 | /* This side of the conversion is fully algorithmic. */ |
| 71 | |
| 72 | /* Any spaces at the end of the tag are replaced by repeating the last |
| 73 | * letter. Eg 'nko ' -> 'Nkoo' */ |
| 74 | if (unlikely ((tag & 0x0000FF00u) == 0x00002000u)) |
| 75 | tag |= (tag >> 8) & 0x0000FF00u; /* Copy second letter to third */ |
| 76 | if (unlikely ((tag & 0x000000FFu) == 0x00000020u)) |
| 77 | tag |= (tag >> 8) & 0x000000FFu; /* Copy third letter to fourth */ |
| 78 | |
| 79 | /* Change first char to uppercase and return */ |
| 80 | return (hb_script_t) (tag & ~0x20000000u); |
| 81 | } |
| 82 | |
| 83 | static hb_tag_t |
| 84 | hb_ot_new_tag_from_script (hb_script_t script) |
| 85 | { |
| 86 | switch ((hb_tag_t) script) { |
| 87 | case HB_SCRIPT_BENGALI: return HB_TAG('b','n','g','2'); |
| 88 | case HB_SCRIPT_DEVANAGARI: return HB_TAG('d','e','v','2'); |
| 89 | case HB_SCRIPT_GUJARATI: return HB_TAG('g','j','r','2'); |
| 90 | case HB_SCRIPT_GURMUKHI: return HB_TAG('g','u','r','2'); |
| 91 | case HB_SCRIPT_KANNADA: return HB_TAG('k','n','d','2'); |
| 92 | case HB_SCRIPT_MALAYALAM: return HB_TAG('m','l','m','2'); |
| 93 | case HB_SCRIPT_ORIYA: return HB_TAG('o','r','y','2'); |
| 94 | case HB_SCRIPT_TAMIL: return HB_TAG('t','m','l','2'); |
| 95 | case HB_SCRIPT_TELUGU: return HB_TAG('t','e','l','2'); |
| 96 | case HB_SCRIPT_MYANMAR: return HB_TAG('m','y','m','2'); |
| 97 | } |
| 98 | |
| 99 | return HB_OT_TAG_DEFAULT_SCRIPT; |
| 100 | } |
| 101 | |
| 102 | static hb_script_t |
| 103 | hb_ot_new_tag_to_script (hb_tag_t tag) |
| 104 | { |
| 105 | switch (tag) { |
| 106 | case HB_TAG('b','n','g','2'): return HB_SCRIPT_BENGALI; |
| 107 | case HB_TAG('d','e','v','2'): return HB_SCRIPT_DEVANAGARI; |
| 108 | case HB_TAG('g','j','r','2'): return HB_SCRIPT_GUJARATI; |
| 109 | case HB_TAG('g','u','r','2'): return HB_SCRIPT_GURMUKHI; |
| 110 | case HB_TAG('k','n','d','2'): return HB_SCRIPT_KANNADA; |
| 111 | case HB_TAG('m','l','m','2'): return HB_SCRIPT_MALAYALAM; |
| 112 | case HB_TAG('o','r','y','2'): return HB_SCRIPT_ORIYA; |
| 113 | case HB_TAG('t','m','l','2'): return HB_SCRIPT_TAMIL; |
| 114 | case HB_TAG('t','e','l','2'): return HB_SCRIPT_TELUGU; |
| 115 | case HB_TAG('m','y','m','2'): return HB_SCRIPT_MYANMAR; |
| 116 | } |
| 117 | |
| 118 | return HB_SCRIPT_UNKNOWN; |
| 119 | } |
| 120 | |
| 121 | #ifndef HB_DISABLE_DEPRECATED |
| 122 | /** |
| 123 | * hb_ot_tags_from_script: |
| 124 | * @script: an #hb_script_t to convert. |
| 125 | * @script_tag_1: (out): output #hb_tag_t. |
| 126 | * @script_tag_2: (out): output #hb_tag_t. |
| 127 | * |
| 128 | * Converts an #hb_script_t to script tags. |
| 129 | * |
| 130 | * Since: 0.6.0 |
| 131 | * Deprecated: 2.0.0: use hb_ot_tags_from_script_and_language() instead |
| 132 | **/ |
| 133 | void |
| 134 | hb_ot_tags_from_script (hb_script_t script, |
| 135 | hb_tag_t *script_tag_1, |
| 136 | hb_tag_t *script_tag_2) |
| 137 | { |
| 138 | unsigned int count = 2; |
| 139 | hb_tag_t tags[2]; |
| 140 | hb_ot_tags_from_script_and_language (script, HB_LANGUAGE_INVALID, &count, tags, nullptr, nullptr); |
| 141 | *script_tag_1 = count > 0 ? tags[0] : HB_OT_TAG_DEFAULT_SCRIPT; |
| 142 | *script_tag_2 = count > 1 ? tags[1] : HB_OT_TAG_DEFAULT_SCRIPT; |
| 143 | } |
| 144 | #endif |
| 145 | |
| 146 | /* |
| 147 | * Complete list at: |
| 148 | * https://docs.microsoft.com/en-us/typography/opentype/spec/scripttags |
| 149 | * |
| 150 | * Most of the script tags are the same as the ISO 15924 tag but lowercased. |
| 151 | * So we just do that, and handle the exceptional cases in a switch. |
| 152 | */ |
| 153 | |
| 154 | static void |
| 155 | hb_ot_all_tags_from_script (hb_script_t script, |
| 156 | unsigned int *count /* IN/OUT */, |
| 157 | hb_tag_t *tags /* OUT */) |
| 158 | { |
| 159 | unsigned int i = 0; |
| 160 | |
| 161 | hb_tag_t new_tag = hb_ot_new_tag_from_script (script); |
| 162 | if (unlikely (new_tag != HB_OT_TAG_DEFAULT_SCRIPT)) |
| 163 | { |
| 164 | /* HB_SCRIPT_MYANMAR maps to 'mym2', but there is no 'mym3'. */ |
| 165 | if (new_tag != HB_TAG('m','y','m','2')) |
| 166 | tags[i++] = new_tag | '3'; |
| 167 | if (*count > i) |
| 168 | tags[i++] = new_tag; |
| 169 | } |
| 170 | |
| 171 | if (*count > i) |
| 172 | { |
| 173 | hb_tag_t old_tag = hb_ot_old_tag_from_script (script); |
| 174 | if (old_tag != HB_OT_TAG_DEFAULT_SCRIPT) |
| 175 | tags[i++] = old_tag; |
| 176 | } |
| 177 | |
| 178 | *count = i; |
| 179 | } |
| 180 | |
| 181 | /** |
| 182 | * hb_ot_tag_to_script: |
| 183 | * @tag: a script tag |
| 184 | * |
| 185 | * Converts a script tag to an #hb_script_t. |
| 186 | * |
| 187 | * Return value: The #hb_script_t corresponding to @tag. |
| 188 | * |
| 189 | **/ |
| 190 | hb_script_t |
| 191 | hb_ot_tag_to_script (hb_tag_t tag) |
| 192 | { |
| 193 | unsigned char digit = tag & 0x000000FFu; |
| 194 | if (unlikely (digit == '2' || digit == '3')) |
| 195 | return hb_ot_new_tag_to_script (tag & 0xFFFFFF32); |
| 196 | |
| 197 | return hb_ot_old_tag_to_script (tag); |
| 198 | } |
| 199 | |
| 200 | |
| 201 | /* hb_language_t */ |
| 202 | |
| 203 | static inline bool |
| 204 | subtag_matches (const char *lang_str, |
| 205 | const char *limit, |
| 206 | const char *subtag, |
| 207 | unsigned subtag_len) |
| 208 | { |
| 209 | if (likely ((unsigned) (limit - lang_str) < subtag_len)) |
| 210 | return false; |
| 211 | |
| 212 | do { |
| 213 | const char *s = strstr (lang_str, subtag); |
| 214 | if (!s || s >= limit) |
| 215 | return false; |
| 216 | if (!ISALNUM (s[subtag_len])) |
| 217 | return true; |
| 218 | lang_str = s + subtag_len; |
| 219 | } while (true); |
| 220 | } |
| 221 | |
| 222 | static bool |
| 223 | lang_matches (const char *lang_str, |
| 224 | const char *limit, |
| 225 | const char *spec, |
| 226 | unsigned spec_len) |
| 227 | { |
| 228 | /* Same as hb_language_matches(); duplicated. */ |
| 229 | |
| 230 | if (likely ((unsigned) (limit - lang_str) < spec_len)) |
| 231 | return false; |
| 232 | |
| 233 | return strncmp (lang_str, spec, spec_len) == 0 && |
| 234 | (lang_str[spec_len] == '\0' || lang_str[spec_len] == '-'); |
| 235 | } |
| 236 | |
| 237 | struct LangTag |
| 238 | { |
| 239 | hb_tag_t language; |
| 240 | hb_tag_t tag; |
| 241 | |
| 242 | int cmp (hb_tag_t a) const |
| 243 | { |
| 244 | return a < this->language ? -1 : a > this->language ? +1 : 0; |
| 245 | } |
| 246 | int cmp (const LangTag *that) const |
| 247 | { return cmp (that->language); } |
| 248 | }; |
| 249 | |
| 250 | #include "hb-ot-tag-table.hh" |
| 251 | |
| 252 | /* The corresponding languages IDs for the following IDs are unclear, |
| 253 | * overlap, or are architecturally weird. Needs more research. */ |
| 254 | |
| 255 | /*{"??", {HB_TAG('B','C','R',' ')}},*/ /* Bible Cree */ |
| 256 | /*{"zh?", {HB_TAG('C','H','N',' ')}},*/ /* Chinese (seen in Microsoft fonts) */ |
| 257 | /*{"ar-Syrc?", {HB_TAG('G','A','R',' ')}},*/ /* Garshuni */ |
| 258 | /*{"??", {HB_TAG('N','G','R',' ')}},*/ /* Nagari */ |
| 259 | /*{"??", {HB_TAG('Y','I','C',' ')}},*/ /* Yi Classic */ |
| 260 | /*{"zh?", {HB_TAG('Z','H','P',' ')}},*/ /* Chinese Phonetic */ |
| 261 | |
| 262 | #ifndef HB_DISABLE_DEPRECATED |
| 263 | /** |
| 264 | * hb_ot_tag_from_language: |
| 265 | * @language: an #hb_language_t to convert. |
| 266 | * |
| 267 | * Converts an #hb_language_t to an #hb_tag_t. |
| 268 | * |
| 269 | * Since: 0.6.0 |
| 270 | * Deprecated: 2.0.0: use hb_ot_tags_from_script_and_language() instead |
| 271 | **/ |
| 272 | hb_tag_t |
| 273 | hb_ot_tag_from_language (hb_language_t language) |
| 274 | { |
| 275 | unsigned int count = 1; |
| 276 | hb_tag_t tags[1]; |
| 277 | hb_ot_tags_from_script_and_language (HB_SCRIPT_UNKNOWN, language, nullptr, nullptr, &count, tags); |
| 278 | return count > 0 ? tags[0] : HB_OT_TAG_DEFAULT_LANGUAGE; |
| 279 | } |
| 280 | #endif |
| 281 | |
| 282 | static void |
| 283 | hb_ot_tags_from_language (const char *lang_str, |
| 284 | const char *limit, |
| 285 | unsigned int *count, |
| 286 | hb_tag_t *tags) |
| 287 | { |
| 288 | |
| 289 | #ifndef HB_NO_LANGUAGE_LONG |
| 290 | /* Check for matches of multiple subtags. */ |
| 291 | if (hb_ot_tags_from_complex_language (lang_str, limit, count, tags)) |
| 292 | return; |
| 293 | #endif |
| 294 | |
| 295 | /* Find a language matching in the first component. */ |
| 296 | #ifndef HB_NO_LANGUAGE_LONG |
| 297 | const char *s; s = strchr (lang_str, '-'); |
| 298 | #endif |
| 299 | { |
| 300 | #ifndef HB_NO_LANGUAGE_LONG |
| 301 | if (s && limit - lang_str >= 6) |
| 302 | { |
| 303 | const char *extlang_end = strchr (s + 1, '-'); |
| 304 | /* If there is an extended language tag, use it. */ |
| 305 | if (3 == (extlang_end ? extlang_end - s - 1 : strlen (s + 1)) && |
| 306 | ISALPHA (s[1])) |
| 307 | lang_str = s + 1; |
| 308 | } |
| 309 | #endif |
| 310 | const LangTag *ot_languages = nullptr; |
| 311 | unsigned ot_languages_len = 0; |
| 312 | const char *dash = strchr (lang_str, '-'); |
| 313 | unsigned first_len = dash ? dash - lang_str : limit - lang_str; |
| 314 | if (first_len == 2) |
| 315 | { |
| 316 | ot_languages = ot_languages2; |
| 317 | ot_languages_len = ARRAY_LENGTH (ot_languages2); |
| 318 | } |
| 319 | #ifndef HB_NO_LANGUAGE_LONG |
| 320 | else if (first_len == 3) |
| 321 | { |
| 322 | ot_languages = ot_languages3; |
| 323 | ot_languages_len = ARRAY_LENGTH (ot_languages3); |
| 324 | } |
| 325 | #endif |
| 326 | |
| 327 | hb_tag_t lang_tag = hb_tag_from_string (lang_str, first_len); |
| 328 | |
| 329 | static hb_atomic_int_t last_tag_idx; /* Poor man's cache. */ |
| 330 | unsigned tag_idx = last_tag_idx; |
| 331 | |
| 332 | if (likely (tag_idx < ot_languages_len && ot_languages[tag_idx].language == lang_tag) || |
| 333 | hb_sorted_array (ot_languages, ot_languages_len).bfind (lang_tag, &tag_idx)) |
| 334 | { |
| 335 | last_tag_idx = tag_idx; |
| 336 | unsigned int i; |
| 337 | while (tag_idx != 0 && |
| 338 | ot_languages[tag_idx].language == ot_languages[tag_idx - 1].language) |
| 339 | tag_idx--; |
| 340 | for (i = 0; |
| 341 | i < *count && |
| 342 | tag_idx + i < ot_languages_len && |
| 343 | ot_languages[tag_idx + i].tag != HB_TAG_NONE && |
| 344 | ot_languages[tag_idx + i].language == ot_languages[tag_idx].language; |
| 345 | i++) |
| 346 | tags[i] = ot_languages[tag_idx + i].tag; |
| 347 | *count = i; |
| 348 | return; |
| 349 | } |
| 350 | } |
| 351 | |
| 352 | #ifndef HB_NO_LANGUAGE_LONG |
| 353 | if (!s) |
| 354 | s = lang_str + strlen (lang_str); |
| 355 | if (s - lang_str == 3) { |
| 356 | /* Assume it's ISO-639-3 and upper-case and use it. */ |
| 357 | tags[0] = hb_tag_from_string (lang_str, s - lang_str) & ~0x20202000u; |
| 358 | *count = 1; |
| 359 | return; |
| 360 | } |
| 361 | #endif |
| 362 | |
| 363 | *count = 0; |
| 364 | } |
| 365 | |
| 366 | static bool |
| 367 | parse_private_use_subtag (const char *private_use_subtag, |
| 368 | unsigned int *count, |
| 369 | hb_tag_t *tags, |
| 370 | const char *prefix, |
| 371 | unsigned char (*normalize) (unsigned char)) |
| 372 | { |
| 373 | #ifdef HB_NO_LANGUAGE_PRIVATE_SUBTAG |
| 374 | return false; |
| 375 | #endif |
| 376 | |
| 377 | if (!(private_use_subtag && count && tags && *count)) return false; |
| 378 | |
| 379 | const char *s = strstr (private_use_subtag, prefix); |
| 380 | if (!s) return false; |
| 381 | |
| 382 | char tag[4]; |
| 383 | int i; |
| 384 | s += strlen (prefix); |
| 385 | if (s[0] == '-') { |
| 386 | s += 1; |
| 387 | char c; |
| 388 | for (i = 0; i < 8 && ISHEX (s[i]); i++) |
| 389 | { |
| 390 | c = FROMHEX (s[i]); |
| 391 | if (i % 2 == 0) |
| 392 | tag[i / 2] = c << 4; |
| 393 | else |
| 394 | tag[i / 2] += c; |
| 395 | } |
| 396 | if (i != 8) return false; |
| 397 | } else { |
| 398 | for (i = 0; i < 4 && ISALNUM (s[i]); i++) |
| 399 | tag[i] = normalize (s[i]); |
| 400 | if (!i) return false; |
| 401 | |
| 402 | for (; i < 4; i++) |
| 403 | tag[i] = ' '; |
| 404 | } |
| 405 | tags[0] = HB_TAG (tag[0], tag[1], tag[2], tag[3]); |
| 406 | if ((tags[0] & 0xDFDFDFDF) == HB_OT_TAG_DEFAULT_SCRIPT) |
| 407 | tags[0] ^= ~0xDFDFDFDF; |
| 408 | *count = 1; |
| 409 | return true; |
| 410 | } |
| 411 | |
| 412 | /** |
| 413 | * hb_ot_tags_from_script_and_language: |
| 414 | * @script: an #hb_script_t to convert. |
| 415 | * @language: (nullable): an #hb_language_t to convert. |
| 416 | * @script_count: (inout) (optional): maximum number of script tags to retrieve (IN) |
| 417 | * and actual number of script tags retrieved (OUT) |
| 418 | * @script_tags: (out) (optional): array of size at least @script_count to store the |
| 419 | * script tag results |
| 420 | * @language_count: (inout) (optional): maximum number of language tags to retrieve |
| 421 | * (IN) and actual number of language tags retrieved (OUT) |
| 422 | * @language_tags: (out) (optional): array of size at least @language_count to store |
| 423 | * the language tag results |
| 424 | * |
| 425 | * Converts an #hb_script_t and an #hb_language_t to script and language tags. |
| 426 | * |
| 427 | * Since: 2.0.0 |
| 428 | **/ |
| 429 | void |
| 430 | hb_ot_tags_from_script_and_language (hb_script_t script, |
| 431 | hb_language_t language, |
| 432 | unsigned int *script_count /* IN/OUT */, |
| 433 | hb_tag_t *script_tags /* OUT */, |
| 434 | unsigned int *language_count /* IN/OUT */, |
| 435 | hb_tag_t *language_tags /* OUT */) |
| 436 | { |
| 437 | bool needs_script = true; |
| 438 | |
| 439 | if (language == HB_LANGUAGE_INVALID) |
| 440 | { |
| 441 | if (language_count && language_tags && *language_count) |
| 442 | *language_count = 0; |
| 443 | } |
| 444 | else |
| 445 | { |
| 446 | const char *lang_str, *s, *limit, *private_use_subtag; |
| 447 | bool needs_language; |
| 448 | |
| 449 | lang_str = hb_language_to_string (language); |
| 450 | limit = nullptr; |
| 451 | private_use_subtag = nullptr; |
| 452 | if (lang_str[0] == 'x' && lang_str[1] == '-') |
| 453 | { |
| 454 | private_use_subtag = lang_str; |
| 455 | } else { |
| 456 | for (s = lang_str + 1; *s; s++) |
| 457 | { |
| 458 | if (s[-1] == '-' && s[1] == '-') |
| 459 | { |
| 460 | if (s[0] == 'x') |
| 461 | { |
| 462 | private_use_subtag = s; |
| 463 | if (!limit) |
| 464 | limit = s - 1; |
| 465 | break; |
| 466 | } else if (!limit) |
| 467 | { |
| 468 | limit = s - 1; |
| 469 | } |
| 470 | } |
| 471 | } |
| 472 | if (!limit) |
| 473 | limit = s; |
| 474 | } |
| 475 | |
| 476 | needs_script = !parse_private_use_subtag (private_use_subtag, script_count, script_tags, "-hbsc" , TOLOWER); |
| 477 | needs_language = !parse_private_use_subtag (private_use_subtag, language_count, language_tags, "-hbot" , TOUPPER); |
| 478 | |
| 479 | if (needs_language && language_count && language_tags && *language_count) |
| 480 | hb_ot_tags_from_language (lang_str, limit, language_count, language_tags); |
| 481 | } |
| 482 | |
| 483 | if (needs_script && script_count && script_tags && *script_count) |
| 484 | hb_ot_all_tags_from_script (script, script_count, script_tags); |
| 485 | } |
| 486 | |
| 487 | /** |
| 488 | * hb_ot_tag_to_language: |
| 489 | * @tag: an language tag |
| 490 | * |
| 491 | * Converts a language tag to an #hb_language_t. |
| 492 | * |
| 493 | * Return value: (transfer none) (nullable): |
| 494 | * The #hb_language_t corresponding to @tag. |
| 495 | * |
| 496 | * Since: 0.9.2 |
| 497 | **/ |
| 498 | hb_language_t |
| 499 | hb_ot_tag_to_language (hb_tag_t tag) |
| 500 | { |
| 501 | unsigned int i; |
| 502 | |
| 503 | if (tag == HB_OT_TAG_DEFAULT_LANGUAGE) |
| 504 | return nullptr; |
| 505 | |
| 506 | #ifndef HB_NO_LANGUAGE_LONG |
| 507 | { |
| 508 | hb_language_t disambiguated_tag = hb_ot_ambiguous_tag_to_language (tag); |
| 509 | if (disambiguated_tag != HB_LANGUAGE_INVALID) |
| 510 | return disambiguated_tag; |
| 511 | } |
| 512 | #endif |
| 513 | |
| 514 | char buf[4]; |
| 515 | for (i = 0; i < ARRAY_LENGTH (ot_languages2); i++) |
| 516 | if (ot_languages2[i].tag == tag) |
| 517 | { |
| 518 | hb_tag_to_string (ot_languages2[i].language, buf); |
| 519 | return hb_language_from_string (buf, 2); |
| 520 | } |
| 521 | #ifndef HB_NO_LANGUAGE_LONG |
| 522 | for (i = 0; i < ARRAY_LENGTH (ot_languages3); i++) |
| 523 | if (ot_languages3[i].tag == tag) |
| 524 | { |
| 525 | hb_tag_to_string (ot_languages3[i].language, buf); |
| 526 | return hb_language_from_string (buf, 3); |
| 527 | } |
| 528 | #endif |
| 529 | |
| 530 | /* Return a custom language in the form of "x-hbot-AABBCCDD". |
| 531 | * If it's three letters long, also guess it's ISO 639-3 and lower-case and |
| 532 | * prepend it (if it's not a registered tag, the private use subtags will |
| 533 | * ensure that calling hb_ot_tag_from_language on the result will still return |
| 534 | * the same tag as the original tag). |
| 535 | */ |
| 536 | { |
| 537 | char buf[20]; |
| 538 | char *str = buf; |
| 539 | if (ISALPHA (tag >> 24) |
| 540 | && ISALPHA ((tag >> 16) & 0xFF) |
| 541 | && ISALPHA ((tag >> 8) & 0xFF) |
| 542 | && (tag & 0xFF) == ' ') |
| 543 | { |
| 544 | buf[0] = TOLOWER (tag >> 24); |
| 545 | buf[1] = TOLOWER ((tag >> 16) & 0xFF); |
| 546 | buf[2] = TOLOWER ((tag >> 8) & 0xFF); |
| 547 | buf[3] = '-'; |
| 548 | str += 4; |
| 549 | } |
| 550 | snprintf (str, 16, "x-hbot-%08x" , tag); |
| 551 | return hb_language_from_string (&*buf, -1); |
| 552 | } |
| 553 | } |
| 554 | |
| 555 | /** |
| 556 | * hb_ot_tags_to_script_and_language: |
| 557 | * @script_tag: a script tag |
| 558 | * @language_tag: a language tag |
| 559 | * @script: (out) (optional): the #hb_script_t corresponding to @script_tag. |
| 560 | * @language: (out) (optional): the #hb_language_t corresponding to @script_tag and |
| 561 | * @language_tag. |
| 562 | * |
| 563 | * Converts a script tag and a language tag to an #hb_script_t and an |
| 564 | * #hb_language_t. |
| 565 | * |
| 566 | * Since: 2.0.0 |
| 567 | **/ |
| 568 | void |
| 569 | hb_ot_tags_to_script_and_language (hb_tag_t script_tag, |
| 570 | hb_tag_t language_tag, |
| 571 | hb_script_t *script /* OUT */, |
| 572 | hb_language_t *language /* OUT */) |
| 573 | { |
| 574 | hb_script_t script_out = hb_ot_tag_to_script (script_tag); |
| 575 | if (script) |
| 576 | *script = script_out; |
| 577 | if (language) |
| 578 | { |
| 579 | unsigned int script_count = 1; |
| 580 | hb_tag_t primary_script_tag[1]; |
| 581 | hb_ot_tags_from_script_and_language (script_out, |
| 582 | HB_LANGUAGE_INVALID, |
| 583 | &script_count, |
| 584 | primary_script_tag, |
| 585 | nullptr, nullptr); |
| 586 | *language = hb_ot_tag_to_language (language_tag); |
| 587 | if (script_count == 0 || primary_script_tag[0] != script_tag) |
| 588 | { |
| 589 | unsigned char *buf; |
| 590 | const char *lang_str = hb_language_to_string (*language); |
| 591 | size_t len = strlen (lang_str); |
| 592 | buf = (unsigned char *) hb_malloc (len + 16); |
| 593 | if (unlikely (!buf)) |
| 594 | { |
| 595 | *language = nullptr; |
| 596 | } |
| 597 | else |
| 598 | { |
| 599 | int shift; |
| 600 | hb_memcpy (buf, lang_str, len); |
| 601 | if (lang_str[0] != 'x' || lang_str[1] != '-') { |
| 602 | buf[len++] = '-'; |
| 603 | buf[len++] = 'x'; |
| 604 | } |
| 605 | buf[len++] = '-'; |
| 606 | buf[len++] = 'h'; |
| 607 | buf[len++] = 'b'; |
| 608 | buf[len++] = 's'; |
| 609 | buf[len++] = 'c'; |
| 610 | buf[len++] = '-'; |
| 611 | for (shift = 28; shift >= 0; shift -= 4) |
| 612 | buf[len++] = TOHEX (script_tag >> shift); |
| 613 | *language = hb_language_from_string ((char *) buf, len); |
| 614 | hb_free (buf); |
| 615 | } |
| 616 | } |
| 617 | } |
| 618 | } |
| 619 | |
| 620 | #ifdef MAIN |
| 621 | static inline void |
| 622 | test_langs_sorted () |
| 623 | { |
| 624 | for (unsigned int i = 1; i < ARRAY_LENGTH (ot_languages2); i++) |
| 625 | { |
| 626 | int c = ot_languages2[i].cmp (&ot_languages2[i - 1]); |
| 627 | if (c > 0) |
| 628 | { |
| 629 | fprintf (stderr, "ot_languages2 not sorted at index %u: %08x %d %08x\n" , |
| 630 | i, ot_languages2[i-1].language, c, ot_languages2[i].language); |
| 631 | abort(); |
| 632 | } |
| 633 | } |
| 634 | #ifndef HB_NO_LANGUAGE_LONG |
| 635 | for (unsigned int i = 1; i < ARRAY_LENGTH (ot_languages3); i++) |
| 636 | { |
| 637 | int c = ot_languages3[i].cmp (&ot_languages3[i - 1]); |
| 638 | if (c > 0) |
| 639 | { |
| 640 | fprintf (stderr, "ot_languages3 not sorted at index %u: %08x %d %08x\n" , |
| 641 | i, ot_languages3[i-1].language, c, ot_languages3[i].language); |
| 642 | abort(); |
| 643 | } |
| 644 | } |
| 645 | #endif |
| 646 | } |
| 647 | |
| 648 | int |
| 649 | main () |
| 650 | { |
| 651 | test_langs_sorted (); |
| 652 | return 0; |
| 653 | } |
| 654 | |
| 655 | #endif |
| 656 | |
| 657 | |
| 658 | #endif |
| 659 | |