2 * Copyright © 2009 Red Hat, Inc.
3 * Copyright © 2011 Google, Inc.
5 * This is part of HarfBuzz, a text shaping library.
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.
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
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.
25 * Red Hat Author(s): Behdad Esfahbod
26 * Google Author(s): Behdad Esfahbod, Roozbeh Pournader
35 hb_ot_old_tag_from_script (hb_script_t script)
37 /* This seems to be accurate as of end of 2012. */
39 switch ((hb_tag_t) script)
41 case HB_SCRIPT_INVALID: return HB_OT_TAG_DEFAULT_SCRIPT;
43 /* KATAKANA and HIRAGANA both map to 'kana' */
44 case HB_SCRIPT_HIRAGANA: return HB_TAG('k','a','n','a');
46 /* Spaces at the end are preserved, unlike ISO 15924 */
47 case HB_SCRIPT_LAO: return HB_TAG('l','a','o',' ');
48 case HB_SCRIPT_YI: return HB_TAG('y','i',' ',' ');
49 /* Unicode-5.0 additions */
50 case HB_SCRIPT_NKO: return HB_TAG('n','k','o',' ');
51 /* Unicode-5.1 additions */
52 case HB_SCRIPT_VAI: return HB_TAG('v','a','i',' ');
55 /* Else, just change first char to lowercase and return */
56 return ((hb_tag_t) script) | 0x20000000u;
60 hb_ot_old_tag_to_script (hb_tag_t tag)
62 if (unlikely (tag == HB_OT_TAG_DEFAULT_SCRIPT))
63 return HB_SCRIPT_INVALID;
65 /* This side of the conversion is fully algorithmic. */
67 /* Any spaces at the end of the tag are replaced by repeating the last
68 * letter. Eg 'nko ' -> 'Nkoo' */
69 if (unlikely ((tag & 0x0000FF00u) == 0x00002000u))
70 tag |= (tag >> 8) & 0x0000FF00u; /* Copy second letter to third */
71 if (unlikely ((tag & 0x000000FFu) == 0x00000020u))
72 tag |= (tag >> 8) & 0x000000FFu; /* Copy third letter to fourth */
74 /* Change first char to uppercase and return */
75 return (hb_script_t) (tag & ~0x20000000u);
79 hb_ot_new_tag_from_script (hb_script_t script)
81 switch ((hb_tag_t) script) {
82 case HB_SCRIPT_BENGALI: return HB_TAG('b','n','g','2');
83 case HB_SCRIPT_DEVANAGARI: return HB_TAG('d','e','v','2');
84 case HB_SCRIPT_GUJARATI: return HB_TAG('g','j','r','2');
85 case HB_SCRIPT_GURMUKHI: return HB_TAG('g','u','r','2');
86 case HB_SCRIPT_KANNADA: return HB_TAG('k','n','d','2');
87 case HB_SCRIPT_MALAYALAM: return HB_TAG('m','l','m','2');
88 case HB_SCRIPT_ORIYA: return HB_TAG('o','r','y','2');
89 case HB_SCRIPT_TAMIL: return HB_TAG('t','m','l','2');
90 case HB_SCRIPT_TELUGU: return HB_TAG('t','e','l','2');
91 case HB_SCRIPT_MYANMAR: return HB_TAG('m','y','m','2');
94 return HB_OT_TAG_DEFAULT_SCRIPT;
98 hb_ot_new_tag_to_script (hb_tag_t tag)
101 case HB_TAG('b','n','g','2'): return HB_SCRIPT_BENGALI;
102 case HB_TAG('d','e','v','2'): return HB_SCRIPT_DEVANAGARI;
103 case HB_TAG('g','j','r','2'): return HB_SCRIPT_GUJARATI;
104 case HB_TAG('g','u','r','2'): return HB_SCRIPT_GURMUKHI;
105 case HB_TAG('k','n','d','2'): return HB_SCRIPT_KANNADA;
106 case HB_TAG('m','l','m','2'): return HB_SCRIPT_MALAYALAM;
107 case HB_TAG('o','r','y','2'): return HB_SCRIPT_ORIYA;
108 case HB_TAG('t','m','l','2'): return HB_SCRIPT_TAMIL;
109 case HB_TAG('t','e','l','2'): return HB_SCRIPT_TELUGU;
110 case HB_TAG('m','y','m','2'): return HB_SCRIPT_MYANMAR;
113 return HB_SCRIPT_UNKNOWN;
117 hb_ot_tags_from_script (hb_script_t script,
118 hb_tag_t *script_tag_1,
119 hb_tag_t *script_tag_2)
121 unsigned int count = 2;
123 hb_ot_tags_from_script_and_language (script, HB_LANGUAGE_INVALID, &count, tags, nullptr, nullptr);
124 *script_tag_1 = count > 0 ? tags[0] : HB_OT_TAG_DEFAULT_SCRIPT;
125 *script_tag_2 = count > 1 ? tags[1] : HB_OT_TAG_DEFAULT_SCRIPT;
130 * https://docs.microsoft.com/en-us/typography/opentype/spec/scripttags
132 * Most of the script tags are the same as the ISO 15924 tag but lowercased.
133 * So we just do that, and handle the exceptional cases in a switch.
137 hb_ot_all_tags_from_script (hb_script_t script,
138 unsigned int *count /* IN/OUT */,
139 hb_tag_t *tags /* OUT */)
143 hb_tag_t new_tag = hb_ot_new_tag_from_script (script);
144 if (unlikely (new_tag != HB_OT_TAG_DEFAULT_SCRIPT))
146 tags[i++] = new_tag | '3';
153 hb_tag_t old_tag = hb_ot_old_tag_from_script (script);
154 if (old_tag != HB_OT_TAG_DEFAULT_SCRIPT)
162 hb_ot_tag_to_script (hb_tag_t tag)
164 unsigned char digit = tag & 0x000000FFu;
165 if (unlikely (digit == '2' || digit == '3'))
166 return hb_ot_new_tag_to_script (tag & 0xFFFFFF32);
168 return hb_ot_old_tag_to_script (tag);
175 subtag_matches (const char *lang_str,
180 const char *s = strstr (lang_str, subtag);
181 if (!s || s >= limit)
183 if (!ISALNUM (s[strlen (subtag)]))
185 lang_str = s + strlen (subtag);
190 lang_matches (const char *lang_str, const char *spec)
192 unsigned int len = strlen (spec);
194 return strncmp (lang_str, spec, len) == 0 &&
195 (lang_str[len] == '\0' || lang_str[len] == '-');
201 hb_tag_t tags[HB_OT_MAX_TAGS_PER_LANGUAGE];
203 int cmp (const char *a) const
205 const char *b = this->language;
210 da = p ? (unsigned int) (p - a) : strlen (a);
213 db = p ? (unsigned int) (p - b) : strlen (b);
215 return strncmp (a, b, MAX (da, db));
217 int cmp (const LangTag *that) const
218 { return cmp (that->language); }
221 #include "hb-ot-tag-table.hh"
223 /* The corresponding languages IDs for the following IDs are unclear,
224 * overlap, or are architecturally weird. Needs more research. */
226 /*{"??", {HB_TAG('B','C','R',' ')}},*/ /* Bible Cree */
227 /*{"zh?", {HB_TAG('C','H','N',' ')}},*/ /* Chinese (seen in Microsoft fonts) */
228 /*{"ar-Syrc?", {HB_TAG('G','A','R',' ')}},*/ /* Garshuni */
229 /*{"??", {HB_TAG('N','G','R',' ')}},*/ /* Nagari */
230 /*{"??", {HB_TAG('Y','I','C',' ')}},*/ /* Yi Classic */
231 /*{"zh?", {HB_TAG('Z','H','P',' ')}},*/ /* Chinese Phonetic */
234 hb_ot_tag_from_language (hb_language_t language)
236 unsigned int count = 1;
238 hb_ot_tags_from_script_and_language (HB_SCRIPT_UNKNOWN, language, nullptr, nullptr, &count, tags);
239 return count > 0 ? tags[0] : HB_OT_TAG_DEFAULT_LANGUAGE;
243 hb_ot_tags_from_language (const char *lang_str,
250 /* Check for matches of multiple subtags. */
251 if (hb_ot_tags_from_complex_language (lang_str, limit, count, tags))
254 /* Find a language matching in the first component. */
255 s = strchr (lang_str, '-');
257 const LangTag *lang_tag;
258 if (s && limit - lang_str >= 6)
260 const char *extlang_end = strchr (s + 1, '-');
261 /* If there is an extended language tag, use it. */
262 if (3 == (extlang_end ? extlang_end - s - 1 : strlen (s + 1)) &&
266 lang_tag = hb_sorted_array (ot_languages).bsearch (lang_str);
270 for (i = 0; i < *count && lang_tag->tags[i] != HB_TAG_NONE; i++)
271 tags[i] = lang_tag->tags[i];
278 s = lang_str + strlen (lang_str);
279 if (s - lang_str == 3) {
280 /* Assume it's ISO-639-3 and upper-case and use it. */
281 tags[0] = hb_tag_from_string (lang_str, s - lang_str) & ~0x20202000u;
290 parse_private_use_subtag (const char *private_use_subtag,
294 unsigned char (*normalize) (unsigned char))
296 if (private_use_subtag && count && tags && *count)
298 const char *s = strstr (private_use_subtag, prefix);
303 s += strlen (prefix);
304 for (i = 0; i < 4 && ISALNUM (s[i]); i++)
305 tag[i] = normalize (s[i]);
310 tags[0] = HB_TAG (tag[0], tag[1], tag[2], tag[3]);
311 if ((tags[0] & 0xDFDFDFDF) == HB_OT_TAG_DEFAULT_SCRIPT)
312 tags[0] ^= ~0xDFDFDFDF;
322 * hb_ot_tags_from_script_and_language:
323 * @script: an #hb_script_t to convert.
324 * @language: an #hb_language_t to convert.
325 * @script_count: (allow-none): maximum number of script tags to retrieve (IN)
326 * and actual number of script tags retrieved (OUT)
327 * @script_tags: (out) (allow-none): array of size at least @script_count to store the
329 * @language_count: (allow-none): maximum number of language tags to retrieve
330 * (IN) and actual number of language tags retrieved (OUT)
331 * @language_tags: (out) (allow-none): array of size at least @language_count to store
332 * the language tag results
334 * Converts an #hb_script_t and an #hb_language_t to script and language tags.
339 hb_ot_tags_from_script_and_language (hb_script_t script,
340 hb_language_t language,
341 unsigned int *script_count /* IN/OUT */,
342 hb_tag_t *script_tags /* OUT */,
343 unsigned int *language_count /* IN/OUT */,
344 hb_tag_t *language_tags /* OUT */)
346 bool needs_script = true;
348 if (language == HB_LANGUAGE_INVALID)
350 if (language_count && language_tags && *language_count)
355 const char *lang_str, *s, *limit, *private_use_subtag;
358 lang_str = hb_language_to_string (language);
360 private_use_subtag = nullptr;
361 if (lang_str[0] == 'x' && lang_str[1] == '-')
363 private_use_subtag = lang_str;
365 for (s = lang_str + 1; *s; s++)
367 if (s[-1] == '-' && s[1] == '-')
371 private_use_subtag = s;
385 needs_script = parse_private_use_subtag (private_use_subtag, script_count, script_tags, "-hbsc", TOLOWER);
386 needs_language = parse_private_use_subtag (private_use_subtag, language_count, language_tags, "-hbot", TOUPPER);
388 if (needs_language && language_count && language_tags && *language_count)
389 hb_ot_tags_from_language (lang_str, limit, language_count, language_tags);
392 if (needs_script && script_count && script_tags && *script_count)
393 hb_ot_all_tags_from_script (script, script_count, script_tags);
397 * hb_ot_tag_to_language:
401 * Return value: (transfer none):
406 hb_ot_tag_to_language (hb_tag_t tag)
410 if (tag == HB_OT_TAG_DEFAULT_LANGUAGE)
414 hb_language_t disambiguated_tag = hb_ot_ambiguous_tag_to_language (tag);
415 if (disambiguated_tag != HB_LANGUAGE_INVALID)
416 return disambiguated_tag;
419 for (i = 0; i < ARRAY_LENGTH (ot_languages); i++)
420 if (ot_languages[i].tags[0] == tag)
421 return hb_language_from_string (ot_languages[i].language, -1);
423 /* Else return a custom language in the form of "x-hbotABCD" */
425 unsigned char buf[11] = "x-hbot";
427 buf[7] = (tag >> 16) & 0xFF;
428 buf[8] = (tag >> 8) & 0xFF;
433 return hb_language_from_string ((char *) buf, -1);
438 * hb_ot_tags_to_script_and_language:
439 * @script_tag: a script tag
440 * @language_tag: a language tag
441 * @script: (allow-none): the #hb_script_t corresponding to @script_tag (OUT).
442 * @language: (allow-none): the #hb_language_t corresponding to @script_tag and
443 * @language_tag (OUT).
445 * Converts a script tag and a language tag to an #hb_script_t and an
451 hb_ot_tags_to_script_and_language (hb_tag_t script_tag,
452 hb_tag_t language_tag,
453 hb_script_t *script /* OUT */,
454 hb_language_t *language /* OUT */)
456 hb_script_t script_out = hb_ot_tag_to_script (script_tag);
458 *script = script_out;
461 unsigned int script_count = 1;
462 hb_tag_t primary_script_tag[1];
463 hb_ot_tags_from_script_and_language (script_out,
468 *language = hb_ot_tag_to_language (language_tag);
469 if (script_count == 0 || primary_script_tag[0] != script_tag)
472 const char *lang_str = hb_language_to_string (*language);
473 size_t len = strlen (lang_str);
474 buf = (unsigned char *) malloc (len + 11);
481 memcpy (buf, lang_str, len);
482 if (lang_str[0] != 'x' || lang_str[1] != '-') {
491 buf[len++] = script_tag >> 24;
492 buf[len++] = (script_tag >> 16) & 0xFF;
493 buf[len++] = (script_tag >> 8) & 0xFF;
494 buf[len++] = script_tag & 0xFF;
495 *language = hb_language_from_string ((char *) buf, len);
506 for (unsigned int i = 1; i < ARRAY_LENGTH (ot_languages); i++)
508 int c = ot_languages[i].cmp (&ot_languages[i - 1]);
511 fprintf (stderr, "ot_languages not sorted at index %d: %s %d %s\n",
512 i, ot_languages[i-1].language, c, ot_languages[i].language);
521 test_langs_sorted ();