Mercurial > hgrepos > Python2 > PyMuPDF

comparison mupdf-source/thirdparty/harfbuzz/src/hb-ot-name-table.hh @ 2:b50eed0cc0ef upstream

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
ADD: MuPDF v1.26.7: the MuPDF source as downloaded by a default build of PyMuPDF 1.26.4. The directory name has changed: no version number in the expanded directory now.
author Franz Glasner <fzglas.hg@dom66.de>
date Mon, 15 Sep 2025 11:43:07 +0200
parents
children
comparison
equal deleted inserted replaced
1:1d09e1dec1d9 2:b50eed0cc0ef
1 /*
2 * Copyright © 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_NAME_TABLE_HH
28 #define HB_OT_NAME_TABLE_HH
29
30 #include "hb-open-type.hh"
31 #include "hb-ot-name-language.hh"
32 #include "hb-aat-layout.hh"
33 #include "hb-utf.hh"
34
35
36 namespace OT {
37
38 template <typename in_utf_t, typename out_utf_t>
39 inline unsigned int
40 hb_ot_name_convert_utf (hb_bytes_t bytes,
41 unsigned int *text_size /* IN/OUT */,
42 typename out_utf_t::codepoint_t *text /* OUT */)
43 {
44 unsigned int src_len = bytes.length / sizeof (typename in_utf_t::codepoint_t);
45 const typename in_utf_t::codepoint_t *src = (const typename in_utf_t::codepoint_t *) bytes.arrayZ;
46 const typename in_utf_t::codepoint_t *src_end = src + src_len;
47
48 typename out_utf_t::codepoint_t *dst = text;
49
50 hb_codepoint_t unicode;
51 const hb_codepoint_t replacement = HB_BUFFER_REPLACEMENT_CODEPOINT_DEFAULT;
52
53 if (text_size && *text_size)
54 {
55 (*text_size)--; /* Save room for NUL-termination. */
56 const typename out_utf_t::codepoint_t *dst_end = text + *text_size;
57
58 while (src < src_end && dst < dst_end)
59 {
60 const typename in_utf_t::codepoint_t *src_next = in_utf_t::next (src, src_end, &unicode, replacement);
61 typename out_utf_t::codepoint_t *dst_next = out_utf_t::encode (dst, dst_end, unicode);
62 if (dst_next == dst)
63 break; /* Out-of-room. */
64
65 dst = dst_next;
66 src = src_next;
67 }
68
69 *text_size = dst - text;
70 *dst = 0; /* NUL-terminate. */
71 }
72
73 /* Accumulate length of rest. */
74 unsigned int dst_len = dst - text;
75 while (src < src_end)
76 {
77 src = in_utf_t::next (src, src_end, &unicode, replacement);
78 dst_len += out_utf_t::encode_len (unicode);
79 }
80 return dst_len;
81 }
82
83 #define entry_score var.u16[0]
84 #define entry_index var.u16[1]
85
86
87 /*
88 * name -- Naming
89 * https://docs.microsoft.com/en-us/typography/opentype/spec/name
90 */
91 #define HB_OT_TAG_name HB_TAG('n','a','m','e')
92
93 #define UNSUPPORTED 42
94
95 struct NameRecord
96 {
97 hb_language_t language (hb_face_t *face) const
98 {
99 #ifndef HB_NO_OT_NAME_LANGUAGE
100 unsigned int p = platformID;
101 unsigned int l = languageID;
102
103 if (p == 3)
104 return _hb_ot_name_language_for_ms_code (l);
105
106 if (p == 1)
107 return _hb_ot_name_language_for_mac_code (l);
108
109 #ifndef HB_NO_OT_NAME_LANGUAGE_AAT
110 if (p == 0)
111 return face->table.ltag->get_language (l);
112 #endif
113
114 #endif
115 return HB_LANGUAGE_INVALID;
116 }
117
118 uint16_t score () const
119 {
120 /* Same order as in cmap::find_best_subtable(). */
121 unsigned int p = platformID;
122 unsigned int e = encodingID;
123
124 /* 32-bit. */
125 if (p == 3 && e == 10) return 0;
126 if (p == 0 && e == 6) return 1;
127 if (p == 0 && e == 4) return 2;
128
129 /* 16-bit. */
130 if (p == 3 && e == 1) return 3;
131 if (p == 0 && e == 3) return 4;
132 if (p == 0 && e == 2) return 5;
133 if (p == 0 && e == 1) return 6;
134 if (p == 0 && e == 0) return 7;
135
136 /* Symbol. */
137 if (p == 3 && e == 0) return 8;
138
139 /* We treat all Mac Latin names as ASCII only. */
140 if (p == 1 && e == 0) return 10; /* 10 is magic number :| */
141
142 return UNSUPPORTED;
143 }
144
145 NameRecord* copy (hb_serialize_context_t *c, const void *base
146 #ifdef HB_EXPERIMENTAL_API
147 , const hb_hashmap_t<hb_ot_name_record_ids_t, hb_bytes_t> *name_table_overrides
148 #endif
149 ) const
150 {
151 TRACE_SERIALIZE (this);
152 HB_UNUSED auto snap = c->snapshot ();
153 auto *out = c->embed (this);
154 if (unlikely (!out)) return_trace (nullptr);
155 #ifdef HB_EXPERIMENTAL_API
156 hb_ot_name_record_ids_t record_ids (platformID, encodingID, languageID, nameID);
157 hb_bytes_t* name_bytes;
158
159 if (name_table_overrides->has (record_ids, &name_bytes)) {
160 hb_bytes_t encoded_bytes = *name_bytes;
161 char *name_str_utf16_be = nullptr;
162
163 if (platformID != 1)
164 {
165 unsigned text_size = hb_ot_name_convert_utf<hb_utf8_t, hb_utf16_be_t> (*name_bytes, nullptr, nullptr);
166
167 text_size++; // needs to consider NULL terminator for use in hb_ot_name_convert_utf()
168 unsigned byte_len = text_size * hb_utf16_be_t::codepoint_t::static_size;
169 name_str_utf16_be = (char *) hb_calloc (byte_len, 1);
170 if (!name_str_utf16_be)
171 {
172 c->revert (snap);
173 return_trace (nullptr);
174 }
175 hb_ot_name_convert_utf<hb_utf8_t, hb_utf16_be_t> (*name_bytes, &text_size,
176 (hb_utf16_be_t::codepoint_t *) name_str_utf16_be);
177
178 unsigned encoded_byte_len = text_size * hb_utf16_be_t::codepoint_t::static_size;
179 if (!encoded_byte_len || !c->check_assign (out->length, encoded_byte_len, HB_SERIALIZE_ERROR_INT_OVERFLOW)) {
180 c->revert (snap);
181 hb_free (name_str_utf16_be);
182 return_trace (nullptr);
183 }
184
185 encoded_bytes = hb_bytes_t (name_str_utf16_be, encoded_byte_len);
186 }
187 else
188 {
189 // mac platform, copy the UTF-8 string(all ascii characters) as is
190 if (!c->check_assign (out->length, encoded_bytes.length, HB_SERIALIZE_ERROR_INT_OVERFLOW)) {
191 c->revert (snap);
192 return_trace (nullptr);
193 }
194 }
195
196 out->offset = 0;
197 c->push ();
198 encoded_bytes.copy (c);
199 c->add_link (out->offset, c->pop_pack (), hb_serialize_context_t::Tail, 0);
200 hb_free (name_str_utf16_be);
201 }
202 else
203 #endif
204 {
205 out->offset.serialize_copy (c, offset, base, 0, hb_serialize_context_t::Tail, length);
206 }
207 return_trace (out);
208 }
209
210 bool isUnicode () const
211 {
212 unsigned int p = platformID;
213 unsigned int e = encodingID;
214
215 return (p == 0 ||
216 (p == 3 && (e == 0 || e == 1 || e == 10)));
217 }
218
219 static int cmp (const void *pa, const void *pb)
220 {
221 const NameRecord *a = (const NameRecord *)pa;
222 const NameRecord *b = (const NameRecord *)pb;
223
224 if (a->platformID != b->platformID)
225 return a->platformID - b->platformID;
226
227 if (a->encodingID != b->encodingID)
228 return a->encodingID - b->encodingID;
229
230 if (a->languageID != b->languageID)
231 return a->languageID - b->languageID;
232
233 if (a->nameID != b->nameID)
234 return a->nameID - b->nameID;
235
236 if (a->length != b->length)
237 return a->length - b->length;
238
239 return 0;
240 }
241
242 bool sanitize (hb_sanitize_context_t *c, const void *base) const
243 {
244 TRACE_SANITIZE (this);
245 return_trace (c->check_struct (this) && offset.sanitize (c, base, length));
246 }
247
248 HBUINT16 platformID; /* Platform ID. */
249 HBUINT16 encodingID; /* Platform-specific encoding ID. */
250 HBUINT16 languageID; /* Language ID. */
251 HBUINT16 nameID; /* Name ID. */
252 HBUINT16 length; /* String length (in bytes). */
253 NNOffset16To<UnsizedArrayOf<HBUINT8>>
254 offset; /* String offset from start of storage area (in bytes). */
255 public:
256 DEFINE_SIZE_STATIC (12);
257 };
258
259 static int
260 _hb_ot_name_entry_cmp_key (const void *pa, const void *pb, bool exact)
261 {
262 const hb_ot_name_entry_t *a = (const hb_ot_name_entry_t *) pa;
263 const hb_ot_name_entry_t *b = (const hb_ot_name_entry_t *) pb;
264
265 /* Compare by name_id, then language. */
266
267 if (a->name_id != b->name_id)
268 return a->name_id - b->name_id;
269
270 if (a->language == b->language) return 0;
271 if (!a->language) return -1;
272 if (!b->language) return +1;
273
274 const char *astr = hb_language_to_string (a->language);
275 const char *bstr = hb_language_to_string (b->language);
276
277 signed c = strcmp (astr, bstr);
278
279 // 'a' is the user request, and 'b' is string in the font.
280 // If eg. user asks for "en-us" and font has "en", approve.
281 if (!exact && c &&
282 hb_language_matches (b->language, a->language))
283 return 0;
284
285 return c;
286 }
287
288 static int
289 _hb_ot_name_entry_cmp (const void *pa, const void *pb)
290 {
291 /* Compare by name_id, then language, then score, then index. */
292
293 int v = _hb_ot_name_entry_cmp_key (pa, pb, true);
294 if (v)
295 return v;
296
297 const hb_ot_name_entry_t *a = (const hb_ot_name_entry_t *) pa;
298 const hb_ot_name_entry_t *b = (const hb_ot_name_entry_t *) pb;
299
300 if (a->entry_score != b->entry_score)
301 return a->entry_score - b->entry_score;
302
303 if (a->entry_index != b->entry_index)
304 return a->entry_index - b->entry_index;
305
306 return 0;
307 }
308
309 struct name
310 {
311 static constexpr hb_tag_t tableTag = HB_OT_TAG_name;
312
313 unsigned int get_size () const
314 { return min_size + count * nameRecordZ.item_size; }
315
316 template <typename Iterator,
317 hb_requires (hb_is_source_of (Iterator, const NameRecord &))>
318 bool serialize (hb_serialize_context_t *c,
319 Iterator it,
320 const void *src_string_pool
321 #ifdef HB_EXPERIMENTAL_API
322 , const hb_vector_t<hb_ot_name_record_ids_t>& insert_name_records
323 , const hb_hashmap_t<hb_ot_name_record_ids_t, hb_bytes_t> *name_table_overrides
324 #endif
325 )
326 {
327 TRACE_SERIALIZE (this);
328
329 if (unlikely (!c->extend_min ((*this)))) return_trace (false);
330
331 unsigned total_count = it.len ()
332 #ifdef HB_EXPERIMENTAL_API
333 + insert_name_records.length
334 #endif
335 ;
336 this->format = 0;
337 if (!c->check_assign (this->count, total_count, HB_SERIALIZE_ERROR_INT_OVERFLOW))
338 return false;
339
340 NameRecord *name_records = (NameRecord *) hb_calloc (total_count, NameRecord::static_size);
341 if (unlikely (!name_records)) return_trace (false);
342
343 hb_array_t<NameRecord> records (name_records, total_count);
344
345 for (const NameRecord& record : it)
346 {
347 hb_memcpy (name_records, &record, NameRecord::static_size);
348 name_records++;
349 }
350
351 #ifdef HB_EXPERIMENTAL_API
352 for (unsigned i = 0; i < insert_name_records.length; i++)
353 {
354 const hb_ot_name_record_ids_t& ids = insert_name_records[i];
355 NameRecord record;
356 record.platformID = ids.platform_id;
357 record.encodingID = ids.encoding_id;
358 record.languageID = ids.language_id;
359 record.nameID = ids.name_id;
360 record.length = 0; // handled in NameRecord copy()
361 record.offset = 0;
362 memcpy (name_records, &record, NameRecord::static_size);
363 name_records++;
364 }
365 #endif
366
367 records.qsort ();
368
369 c->copy_all (records,
370 src_string_pool
371 #ifdef HB_EXPERIMENTAL_API
372 , name_table_overrides
373 #endif
374 );
375 hb_free (records.arrayZ);
376
377
378 if (unlikely (c->ran_out_of_room ())) return_trace (false);
379
380 this->stringOffset = c->length ();
381
382 return_trace (true);
383 }
384
385 bool subset (hb_subset_context_t *c) const
386 {
387 TRACE_SUBSET (this);
388
389 name *name_prime = c->serializer->start_embed<name> ();
390 if (unlikely (!name_prime)) return_trace (false);
391
392 #ifdef HB_EXPERIMENTAL_API
393 const hb_hashmap_t<hb_ot_name_record_ids_t, hb_bytes_t> *name_table_overrides =
394 c->plan->name_table_overrides;
395 #endif
396
397 auto it =
398 + nameRecordZ.as_array (count)
399 | hb_filter (c->plan->name_ids, &NameRecord::nameID)
400 | hb_filter (c->plan->name_languages, &NameRecord::languageID)
401 | hb_filter ([&] (const NameRecord& namerecord) {
402 return
403 (c->plan->flags & HB_SUBSET_FLAGS_NAME_LEGACY)
404 || namerecord.isUnicode ();
405 })
406 #ifdef HB_EXPERIMENTAL_API
407 | hb_filter ([&] (const NameRecord& namerecord) {
408 if (name_table_overrides->is_empty ())
409 return true;
410 hb_ot_name_record_ids_t rec_ids (namerecord.platformID,
411 namerecord.encodingID,
412 namerecord.languageID,
413 namerecord.nameID);
414
415 hb_bytes_t *p;
416 if (name_table_overrides->has (rec_ids, &p) &&
417 (*p).length == 0)
418 return false;
419 return true;
420 })
421 #endif
422 ;
423
424 #ifdef HB_EXPERIMENTAL_API
425 hb_vector_t<hb_ot_name_record_ids_t> insert_name_records;
426 if (!name_table_overrides->is_empty ())
427 {
428 if (unlikely (!insert_name_records.alloc (name_table_overrides->get_population ())))
429 return_trace (false);
430 for (const auto& record_ids : name_table_overrides->keys ())
431 {
432 if (name_table_overrides->get (record_ids).length == 0)
433 continue;
434 if (has_name_record_with_ids (record_ids))
435 continue;
436 insert_name_records.push (record_ids);
437 }
438 }
439 #endif
440
441 return (name_prime->serialize (c->serializer, it,
442 std::addressof (this + stringOffset)
443 #ifdef HB_EXPERIMENTAL_API
444 , insert_name_records
445 , name_table_overrides
446 #endif
447 ));
448 }
449
450 bool sanitize_records (hb_sanitize_context_t *c) const
451 {
452 TRACE_SANITIZE (this);
453 const void *string_pool = (this+stringOffset).arrayZ;
454 return_trace (nameRecordZ.sanitize (c, count, string_pool));
455 }
456
457 bool sanitize (hb_sanitize_context_t *c) const
458 {
459 TRACE_SANITIZE (this);
460 return_trace (c->check_struct (this) &&
461 likely (format == 0 || format == 1) &&
462 c->check_array (nameRecordZ.arrayZ, count) &&
463 c->check_range (this, stringOffset) &&
464 sanitize_records (c));
465 }
466
467 struct accelerator_t
468 {
469 accelerator_t (hb_face_t *face)
470 {
471 this->table = hb_sanitize_context_t ().reference_table<name> (face);
472 assert (this->table.get_length () >= this->table->stringOffset);
473 this->pool = (const char *) (const void *) (this->table+this->table->stringOffset);
474 this->pool_len = this->table.get_length () - this->table->stringOffset;
475 const hb_array_t<const NameRecord> all_names (this->table->nameRecordZ.arrayZ,
476 this->table->count);
477
478 this->names.alloc (all_names.length);
479
480 for (unsigned int i = 0; i < all_names.length; i++)
481 {
482 hb_ot_name_entry_t *entry = this->names.push ();
483
484 entry->name_id = all_names[i].nameID;
485 entry->language = all_names[i].language (face);
486 entry->entry_score = all_names[i].score ();
487 entry->entry_index = i;
488 }
489
490 this->names.qsort (_hb_ot_name_entry_cmp);
491 /* Walk and pick best only for each name_id,language pair,
492 * while dropping unsupported encodings. */
493 unsigned int j = 0;
494 for (unsigned int i = 0; i < this->names.length; i++)
495 {
496 if (this->names[i].entry_score == UNSUPPORTED ||
497 this->names[i].language == HB_LANGUAGE_INVALID)
498 continue;
499 if (i &&
500 this->names[i - 1].name_id == this->names[i].name_id &&
501 this->names[i - 1].language == this->names[i].language)
502 continue;
503 this->names[j++] = this->names[i];
504 }
505 this->names.resize (j);
506 }
507 ~accelerator_t ()
508 {
509 this->table.destroy ();
510 }
511
512 int get_index (hb_ot_name_id_t name_id,
513 hb_language_t language,
514 unsigned int *width=nullptr) const
515 {
516 const hb_ot_name_entry_t key = {name_id, {0}, language};
517 const hb_ot_name_entry_t *entry = hb_bsearch (key, (const hb_ot_name_entry_t *) this->names,
518 this->names.length,
519 sizeof (hb_ot_name_entry_t),
520 _hb_ot_name_entry_cmp_key,
521 true);
522
523 if (!entry)
524 {
525 entry = hb_bsearch (key, (const hb_ot_name_entry_t *) this->names,
526 this->names.length,
527 sizeof (hb_ot_name_entry_t),
528 _hb_ot_name_entry_cmp_key,
529 false);
530 }
531
532 if (!entry)
533 return -1;
534
535 if (width)
536 *width = entry->entry_score < 10 ? 2 : 1;
537
538 return entry->entry_index;
539 }
540
541 hb_bytes_t get_name (unsigned int idx) const
542 {
543 const hb_array_t<const NameRecord> all_names (table->nameRecordZ.arrayZ, table->count);
544 const NameRecord &record = all_names[idx];
545 const hb_bytes_t string_pool (pool, pool_len);
546 return string_pool.sub_array (record.offset, record.length);
547 }
548
549 private:
550 const char *pool;
551 unsigned int pool_len;
552 public:
553 hb_blob_ptr_t<name> table;
554 hb_vector_t<hb_ot_name_entry_t> names;
555 };
556
557 private:
558 // sometimes NameRecords are not sorted in the font file, so use linear search
559 // here
560 bool has_name_record_with_ids (const hb_ot_name_record_ids_t& record_ids) const
561 {
562 for (const auto& record : nameRecordZ.as_array (count))
563 {
564 if (record.platformID == record_ids.platform_id &&
565 record.encodingID == record_ids.encoding_id &&
566 record.languageID == record_ids.language_id &&
567 record.nameID == record_ids.name_id)
568 return true;
569 }
570 return false;
571 }
572
573 public:
574 /* We only implement format 0 for now. */
575 HBUINT16 format; /* Format selector (=0/1). */
576 HBUINT16 count; /* Number of name records. */
577 NNOffset16To<UnsizedArrayOf<HBUINT8>>
578 stringOffset; /* Offset to start of string storage (from start of table). */
579 UnsizedArrayOf<NameRecord>
580 nameRecordZ; /* The name records where count is the number of records. */
581 public:
582 DEFINE_SIZE_ARRAY (6, nameRecordZ);
583 };
584
585 #undef entry_index
586 #undef entry_score
587
588 struct name_accelerator_t : name::accelerator_t {
589 name_accelerator_t (hb_face_t *face) : name::accelerator_t (face) {}
590 };
591
592 } /* namespace OT */
593
594
595 #endif /* HB_OT_NAME_TABLE_HH */