1 /*
2 * Copyright © 2018 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): Garret Rieger, Rod Sheeter, Behdad Esfahbod
25 */
26
27 #include "hb.hh"
28 #include "hb-open-type.hh"
29
30 #include "hb-subset.hh"
31
32 #include "hb-open-file.hh"
33 #include "hb-ot-cmap-table.hh"
34 #include "hb-ot-glyf-table.hh"
35 #include "hb-ot-hdmx-table.hh"
36 #include "hb-ot-head-table.hh"
37 #include "hb-ot-hhea-table.hh"
38 #include "hb-ot-hmtx-table.hh"
39 #include "hb-ot-maxp-table.hh"
40 #include "hb-ot-color-sbix-table.hh"
41 #include "hb-ot-color-colr-table.hh"
42 #include "hb-ot-color-cpal-table.hh"
43 #include "hb-ot-os2-table.hh"
44 #include "hb-ot-post-table.hh"
45 #include "hb-ot-post-table-v2subset.hh"
46 #include "hb-ot-cff1-table.hh"
47 #include "hb-ot-cff2-table.hh"
48 #include "hb-ot-vorg-table.hh"
49 #include "hb-ot-name-table.hh"
50 #include "hb-ot-color-cbdt-table.hh"
51 #include "hb-ot-layout-gsub-table.hh"
52 #include "hb-ot-layout-gpos-table.hh"
53 #include "hb-ot-var-gvar-table.hh"
54 #include "hb-ot-var-hvar-table.hh"
55 #include "hb-ot-math-table.hh"
56 #include "hb-repacker.hh"
57
58 /**
59 * SECTION:hb-subset
60 * @title: hb-subset
61 * @short_description: Subsets font files.
62 * @include: hb-subset.h
63 *
64 * Subsetting reduces the codepoint coverage of font files and removes all data
65 * that is no longer needed. A subset input describes the desired subset. The input is
66 * provided along with a font to the subsetting operation. Output is a new font file
67 * containing only the data specified in the input.
68 *
69 * Currently most outline and bitmap tables are supported: glyf, CFF, CFF2, sbix,
70 * COLR, and CBDT/CBLC. This also includes fonts with variable outlines via OpenType
71 * variations. Notably EBDT/EBLC and SVG are not supported. Layout subsetting is supported
72 * only for OpenType Layout tables (GSUB, GPOS, GDEF). Notably subsetting of graphite or AAT tables
73 * is not yet supported.
74 *
75 * Fonts with graphite or AAT tables may still be subsetted but will likely need to use the
76 * retain glyph ids option and configure the subset to pass through the layout tables untouched.
77 */
78
79 static unsigned
_plan_estimate_subset_table_size(hb_subset_plan_t * plan,unsigned table_len)80 _plan_estimate_subset_table_size (hb_subset_plan_t *plan, unsigned table_len)
81 {
82 unsigned src_glyphs = plan->source->get_num_glyphs ();
83 unsigned dst_glyphs = plan->glyphset ()->get_population ();
84
85 if (unlikely (!src_glyphs))
86 return 512 + table_len;
87
88 return 512 + (unsigned) (table_len * sqrt ((double) dst_glyphs / src_glyphs));
89 }
90
91 /*
92 * Repack the serialization buffer if any offset overflows exist.
93 */
94 static hb_blob_t*
_repack(hb_tag_t tag,const hb_serialize_context_t & c)95 _repack (hb_tag_t tag, const hb_serialize_context_t& c)
96 {
97 if (tag != HB_OT_TAG_GPOS
98 && tag != HB_OT_TAG_GSUB)
99 {
100 // Check for overflow in a non-handled table.
101 return c.successful () ? c.copy_blob () : nullptr;
102 }
103
104 if (!c.offset_overflow ())
105 return c.copy_blob ();
106
107 hb_vector_t<char> buf;
108 int buf_size = c.end - c.start;
109 if (unlikely (!buf.alloc (buf_size)))
110 return nullptr;
111
112 hb_serialize_context_t repacked ((void *) buf, buf_size);
113 hb_resolve_overflows (c.object_graph (), tag, &repacked);
114
115 if (unlikely (repacked.in_error ()))
116 // TODO(garretrieger): refactor so we can share the resize/retry logic with the subset
117 // portion.
118 return nullptr;
119
120 return repacked.copy_blob ();
121 }
122
123 template<typename TableType>
124 static
125 bool
_try_subset(const TableType * table,hb_vector_t<char> * buf,unsigned buf_size,hb_subset_context_t * c)126 _try_subset (const TableType *table,
127 hb_vector_t<char>* buf,
128 unsigned buf_size,
129 hb_subset_context_t* c /* OUT */)
130 {
131 c->serializer->start_serialize<TableType> ();
132 if (c->serializer->in_error ()) return false;
133
134 bool needed = table->subset (c);
135 if (!c->serializer->ran_out_of_room ())
136 {
137 c->serializer->end_serialize ();
138 return needed;
139 }
140
141 buf_size += (buf_size >> 1) + 32;
142 DEBUG_MSG (SUBSET, nullptr, "OT::%c%c%c%c ran out of room; reallocating to %u bytes.",
143 HB_UNTAG (c->table_tag), buf_size);
144
145 if (unlikely (!buf->alloc (buf_size)))
146 {
147 DEBUG_MSG (SUBSET, nullptr, "OT::%c%c%c%c failed to reallocate %u bytes.",
148 HB_UNTAG (c->table_tag), buf_size);
149 return needed;
150 }
151
152 c->serializer->reset (buf->arrayZ, buf_size);
153 return _try_subset (table, buf, buf_size, c);
154 }
155
156 template<typename TableType>
157 static bool
_subset(hb_subset_plan_t * plan)158 _subset (hb_subset_plan_t *plan)
159 {
160 hb_blob_t *source_blob = hb_sanitize_context_t ().reference_table<TableType> (plan->source);
161 const TableType *table = source_blob->as<TableType> ();
162
163 hb_tag_t tag = TableType::tableTag;
164 if (!source_blob->data)
165 {
166 DEBUG_MSG (SUBSET, nullptr,
167 "OT::%c%c%c%c::subset sanitize failed on source table.", HB_UNTAG (tag));
168 hb_blob_destroy (source_blob);
169 return false;
170 }
171
172 hb_vector_t<char> buf;
173 /* TODO Not all tables are glyph-related. 'name' table size for example should not be
174 * affected by number of glyphs. Accommodate that. */
175 unsigned buf_size = _plan_estimate_subset_table_size (plan, source_blob->length);
176 DEBUG_MSG (SUBSET, nullptr,
177 "OT::%c%c%c%c initial estimated table size: %u bytes.", HB_UNTAG (tag), buf_size);
178 if (unlikely (!buf.alloc (buf_size)))
179 {
180 DEBUG_MSG (SUBSET, nullptr, "OT::%c%c%c%c failed to allocate %u bytes.", HB_UNTAG (tag), buf_size);
181 hb_blob_destroy (source_blob);
182 return false;
183 }
184
185 bool needed = false;
186 hb_serialize_context_t serializer (buf.arrayZ, buf_size);
187 {
188 hb_subset_context_t c (source_blob, plan, &serializer, tag);
189 needed = _try_subset (table, &buf, buf_size, &c);
190 }
191 hb_blob_destroy (source_blob);
192
193 if (serializer.in_error () && !serializer.only_offset_overflow ())
194 {
195 DEBUG_MSG (SUBSET, nullptr, "OT::%c%c%c%c::subset FAILED!", HB_UNTAG (tag));
196 return false;
197 }
198
199 if (!needed)
200 {
201 DEBUG_MSG (SUBSET, nullptr, "OT::%c%c%c%c::subset table subsetted to empty.", HB_UNTAG (tag));
202 return true;
203 }
204
205 bool result = false;
206 hb_blob_t *dest_blob = _repack (tag, serializer);
207 if (dest_blob)
208 {
209 DEBUG_MSG (SUBSET, nullptr,
210 "OT::%c%c%c%c final subset table size: %u bytes.",
211 HB_UNTAG (tag), dest_blob->length);
212 result = plan->add_table (tag, dest_blob);
213 hb_blob_destroy (dest_blob);
214 }
215
216 DEBUG_MSG (SUBSET, nullptr, "OT::%c%c%c%c::subset %s",
217 HB_UNTAG (tag), result ? "success" : "FAILED!");
218 return result;
219 }
220
221 static bool
_is_table_present(hb_face_t * source,hb_tag_t tag)222 _is_table_present (hb_face_t *source, hb_tag_t tag)
223 {
224 hb_tag_t table_tags[32];
225 unsigned offset = 0, num_tables = ARRAY_LENGTH (table_tags);
226 while ((hb_face_get_table_tags (source, offset, &num_tables, table_tags), num_tables))
227 {
228 for (unsigned i = 0; i < num_tables; ++i)
229 if (table_tags[i] == tag)
230 return true;
231 offset += num_tables;
232 }
233 return false;
234 }
235
236 static bool
_should_drop_table(hb_subset_plan_t * plan,hb_tag_t tag)237 _should_drop_table (hb_subset_plan_t *plan, hb_tag_t tag)
238 {
239 if (plan->drop_tables->has (tag))
240 return true;
241
242 switch (tag)
243 {
244 case HB_TAG ('c','v','a','r'): /* hint table, fallthrough */
245 case HB_TAG ('c','v','t',' '): /* hint table, fallthrough */
246 case HB_TAG ('f','p','g','m'): /* hint table, fallthrough */
247 case HB_TAG ('p','r','e','p'): /* hint table, fallthrough */
248 case HB_TAG ('h','d','m','x'): /* hint table, fallthrough */
249 case HB_TAG ('V','D','M','X'): /* hint table, fallthrough */
250 return plan->flags & HB_SUBSET_FLAGS_NO_HINTING;
251
252 #ifdef HB_NO_SUBSET_LAYOUT
253 // Drop Layout Tables if requested.
254 case HB_OT_TAG_GDEF:
255 case HB_OT_TAG_GPOS:
256 case HB_OT_TAG_GSUB:
257 case HB_TAG ('m','o','r','x'):
258 case HB_TAG ('m','o','r','t'):
259 case HB_TAG ('k','e','r','x'):
260 case HB_TAG ('k','e','r','n'):
261 return true;
262 #endif
263
264 default:
265 return false;
266 }
267 }
268
269 static bool
_passthrough(hb_subset_plan_t * plan,hb_tag_t tag)270 _passthrough (hb_subset_plan_t *plan, hb_tag_t tag)
271 {
272 hb_blob_t *source_table = hb_face_reference_table (plan->source, tag);
273 bool result = plan->add_table (tag, source_table);
274 hb_blob_destroy (source_table);
275 return result;
276 }
277
278 static bool
_subset_table(hb_subset_plan_t * plan,hb_tag_t tag)279 _subset_table (hb_subset_plan_t *plan, hb_tag_t tag)
280 {
281 if (plan->no_subset_tables->has (tag)) {
282 return _passthrough (plan, tag);
283 }
284
285 DEBUG_MSG (SUBSET, nullptr, "subset %c%c%c%c", HB_UNTAG (tag));
286 switch (tag)
287 {
288 case HB_OT_TAG_glyf: return _subset<const OT::glyf> (plan);
289 case HB_OT_TAG_hdmx: return _subset<const OT::hdmx> (plan);
290 case HB_OT_TAG_name: return _subset<const OT::name> (plan);
291 case HB_OT_TAG_head:
292 if (_is_table_present (plan->source, HB_OT_TAG_glyf) && !_should_drop_table (plan, HB_OT_TAG_glyf))
293 return true; /* skip head, handled by glyf */
294 return _subset<const OT::head> (plan);
295 case HB_OT_TAG_hhea: return true; /* skip hhea, handled by hmtx */
296 case HB_OT_TAG_hmtx: return _subset<const OT::hmtx> (plan);
297 case HB_OT_TAG_vhea: return true; /* skip vhea, handled by vmtx */
298 case HB_OT_TAG_vmtx: return _subset<const OT::vmtx> (plan);
299 case HB_OT_TAG_maxp: return _subset<const OT::maxp> (plan);
300 case HB_OT_TAG_sbix: return _subset<const OT::sbix> (plan);
301 case HB_OT_TAG_loca: return true; /* skip loca, handled by glyf */
302 case HB_OT_TAG_cmap: return _subset<const OT::cmap> (plan);
303 case HB_OT_TAG_OS2 : return _subset<const OT::OS2 > (plan);
304 case HB_OT_TAG_post: return _subset<const OT::post> (plan);
305 case HB_OT_TAG_COLR: return _subset<const OT::COLR> (plan);
306 case HB_OT_TAG_CPAL: return _subset<const OT::CPAL> (plan);
307 case HB_OT_TAG_CBLC: return _subset<const OT::CBLC> (plan);
308 case HB_OT_TAG_CBDT: return true; /* skip CBDT, handled by CBLC */
309 case HB_OT_TAG_MATH: return _subset<const OT::MATH> (plan);
310
311 #ifndef HB_NO_SUBSET_CFF
312 case HB_OT_TAG_cff1: return _subset<const OT::cff1> (plan);
313 case HB_OT_TAG_cff2: return _subset<const OT::cff2> (plan);
314 case HB_OT_TAG_VORG: return _subset<const OT::VORG> (plan);
315 #endif
316
317 #ifndef HB_NO_SUBSET_LAYOUT
318 case HB_OT_TAG_GDEF: return _subset<const OT::GDEF> (plan);
319 case HB_OT_TAG_GSUB: return _subset<const OT::GSUB> (plan);
320 case HB_OT_TAG_GPOS: return _subset<const OT::GPOS> (plan);
321 case HB_OT_TAG_gvar: return _subset<const OT::gvar> (plan);
322 case HB_OT_TAG_HVAR: return _subset<const OT::HVAR> (plan);
323 case HB_OT_TAG_VVAR: return _subset<const OT::VVAR> (plan);
324 #endif
325
326 default:
327 if (plan->flags & HB_SUBSET_FLAGS_PASSTHROUGH_UNRECOGNIZED)
328 return _passthrough (plan, tag);
329
330 // Drop table
331 return true;
332 }
333 }
334
335 /**
336 * hb_subset_or_fail:
337 * @source: font face data to be subset.
338 * @input: input to use for the subsetting.
339 *
340 * Subsets a font according to provided input. Returns nullptr
341 * if the subset operation fails.
342 *
343 * Since: 2.9.0
344 **/
345 hb_face_t *
hb_subset_or_fail(hb_face_t * source,const hb_subset_input_t * input)346 hb_subset_or_fail (hb_face_t *source, const hb_subset_input_t *input)
347 {
348 if (unlikely (!input || !source)) return hb_face_get_empty ();
349
350 hb_subset_plan_t *plan = hb_subset_plan_create (source, input);
351 if (unlikely (plan->in_error ())) {
352 hb_subset_plan_destroy (plan);
353 return nullptr;
354 }
355
356 hb_set_t tags_set;
357 bool success = true;
358 hb_tag_t table_tags[32];
359 unsigned offset = 0, num_tables = ARRAY_LENGTH (table_tags);
360 while ((hb_face_get_table_tags (source, offset, &num_tables, table_tags), num_tables))
361 {
362 for (unsigned i = 0; i < num_tables; ++i)
363 {
364 hb_tag_t tag = table_tags[i];
365 if (_should_drop_table (plan, tag) && !tags_set.has (tag)) continue;
366 tags_set.add (tag);
367 success = _subset_table (plan, tag);
368 if (unlikely (!success)) goto end;
369 }
370 offset += num_tables;
371 }
372 end:
373
374 hb_face_t *result = success ? hb_face_reference (plan->dest) : nullptr;
375
376 hb_subset_plan_destroy (plan);
377 return result;
378 }
379