/src/xpdf-4.06/xpdf/CMap.cc
Line | Count | Source |
1 | | //======================================================================== |
2 | | // |
3 | | // CMap.cc |
4 | | // |
5 | | // Copyright 2001-2003 Glyph & Cog, LLC |
6 | | // |
7 | | //======================================================================== |
8 | | |
9 | | #include <aconf.h> |
10 | | |
11 | | #include <stdio.h> |
12 | | #include <stdlib.h> |
13 | | #include <string.h> |
14 | | #include <ctype.h> |
15 | | #include "gmem.h" |
16 | | #include "gmempp.h" |
17 | | #include "gfile.h" |
18 | | #include "GString.h" |
19 | | #include "GHash.h" |
20 | | #include "Error.h" |
21 | | #include "GlobalParams.h" |
22 | | #include "PSTokenizer.h" |
23 | | #include "Object.h" |
24 | | #include "Stream.h" |
25 | | #include "CMap.h" |
26 | | |
27 | | //------------------------------------------------------------------------ |
28 | | |
29 | | struct CMapVectorEntry { |
30 | | GBool isVector; |
31 | | union { |
32 | | CMapVectorEntry *vector; |
33 | | CID cid; |
34 | | }; |
35 | | }; |
36 | | |
37 | | //------------------------------------------------------------------------ |
38 | | |
39 | 0 | static int getCharFromFile(void *data) { |
40 | 0 | return fgetc((FILE *)data); |
41 | 0 | } |
42 | | |
43 | 0 | static int getCharFromStream(void *data) { |
44 | 0 | return ((Stream *)data)->getChar(); |
45 | 0 | } |
46 | | |
47 | | //------------------------------------------------------------------------ |
48 | | |
49 | | CMap *CMap::parse(CMapCache *cache, GString *collectionA, Object *obj, |
50 | 0 | GHash *usedCMaps) { |
51 | 0 | CMap *cMap; |
52 | 0 | GString *cMapNameA; |
53 | |
|
54 | 0 | if (obj->isName()) { |
55 | 0 | cMapNameA = new GString(obj->getName()); |
56 | 0 | if (!(cMap = globalParams->getCMap(collectionA, cMapNameA))) { |
57 | 0 | error(errSyntaxError, -1, |
58 | 0 | "Unknown CMap '{0:t}' for character collection '{1:t}'", |
59 | 0 | cMapNameA, collectionA); |
60 | 0 | } |
61 | 0 | delete cMapNameA; |
62 | 0 | } else if (obj->isStream()) { |
63 | 0 | GHash *newUsedCMaps = NULL; |
64 | 0 | if (!usedCMaps) { |
65 | 0 | usedCMaps = newUsedCMaps = new GHash(gTrue); |
66 | 0 | } |
67 | 0 | if (!(cMap = CMap::parse(NULL, collectionA, obj->getStream(), |
68 | 0 | usedCMaps))) { |
69 | 0 | error(errSyntaxError, -1, "Invalid CMap in Type 0 font"); |
70 | 0 | } |
71 | 0 | if (newUsedCMaps) { |
72 | 0 | delete newUsedCMaps; |
73 | 0 | } |
74 | 0 | } else { |
75 | 0 | error(errSyntaxError, -1, "Invalid Encoding in Type 0 font"); |
76 | 0 | return NULL; |
77 | 0 | } |
78 | 0 | return cMap; |
79 | 0 | } |
80 | | |
81 | | CMap *CMap::parse(CMapCache *cache, GString *collectionA, |
82 | 0 | GString *cMapNameA, GHash *usedCMaps) { |
83 | 0 | FILE *f; |
84 | 0 | CMap *cMap; |
85 | |
|
86 | 0 | if (!(f = globalParams->findCMapFile(collectionA, cMapNameA))) { |
87 | | |
88 | | // Check for an identity CMap. |
89 | 0 | if (!cMapNameA->cmp("Identity") || !cMapNameA->cmp("Identity-H")) { |
90 | 0 | return new CMap(collectionA->copy(), cMapNameA->copy(), 0); |
91 | 0 | } |
92 | 0 | if (!cMapNameA->cmp("Identity-V")) { |
93 | 0 | return new CMap(collectionA->copy(), cMapNameA->copy(), 1); |
94 | 0 | } |
95 | | |
96 | 0 | error(errSyntaxError, -1, |
97 | 0 | "Couldn't find '{0:t}' CMap file for '{1:t}' collection", |
98 | 0 | cMapNameA, collectionA); |
99 | 0 | return NULL; |
100 | 0 | } |
101 | | |
102 | 0 | GHash *newUsedCMaps = NULL; |
103 | 0 | if (!usedCMaps) { |
104 | 0 | usedCMaps = newUsedCMaps = new GHash(gTrue); |
105 | 0 | } |
106 | |
|
107 | 0 | cMap = new CMap(collectionA->copy(), cMapNameA->copy()); |
108 | 0 | cMap->parse2(cache, &getCharFromFile, f, usedCMaps); |
109 | |
|
110 | 0 | if (newUsedCMaps) { |
111 | 0 | delete newUsedCMaps; |
112 | 0 | } |
113 | |
|
114 | 0 | fclose(f); |
115 | |
|
116 | 0 | return cMap; |
117 | 0 | } |
118 | | |
119 | | CMap *CMap::parse(CMapCache *cache, GString *collectionA, Stream *str, |
120 | 0 | GHash *usedCMaps) { |
121 | 0 | Object obj1; |
122 | 0 | CMap *cMap; |
123 | | |
124 | | // check for a loop |
125 | 0 | if (usedCMaps) { |
126 | 0 | GString *name; |
127 | 0 | if (str->getDict()->lookup("CMapName", &obj1)->isName()) { |
128 | 0 | name = new GString(obj1.getName()); |
129 | 0 | } else { |
130 | 0 | name = new GString(); |
131 | 0 | } |
132 | 0 | obj1.free(); |
133 | 0 | if (usedCMaps->lookupInt(name)) { |
134 | 0 | error(errSyntaxError, -1, "Loop in usecmap"); |
135 | 0 | delete name; |
136 | 0 | return NULL; |
137 | 0 | } |
138 | 0 | usedCMaps->add(name, 1); |
139 | 0 | } |
140 | | |
141 | 0 | cMap = new CMap(collectionA->copy(), NULL); |
142 | |
|
143 | 0 | if (!str->getDict()->lookup("UseCMap", &obj1)->isNull()) { |
144 | 0 | cMap->useCMap(cache, &obj1, usedCMaps); |
145 | 0 | } |
146 | 0 | obj1.free(); |
147 | |
|
148 | 0 | str->reset(); |
149 | 0 | cMap->parse2(cache, &getCharFromStream, str, usedCMaps); |
150 | 0 | str->close(); |
151 | 0 | return cMap; |
152 | 0 | } |
153 | | |
154 | | void CMap::parse2(CMapCache *cache, int (*getCharFunc)(void *), void *data, |
155 | 0 | GHash *usedCMaps) { |
156 | 0 | PSTokenizer *pst; |
157 | 0 | char tok1[256], tok2[256], tok3[256]; |
158 | 0 | int n1, n2, n3; |
159 | 0 | Guint start, end, code; |
160 | |
|
161 | 0 | pst = new PSTokenizer(getCharFunc, data); |
162 | 0 | pst->getToken(tok1, sizeof(tok1), &n1); |
163 | 0 | while (pst->getToken(tok2, sizeof(tok2), &n2)) { |
164 | 0 | if (!strcmp(tok2, "usecmap")) { |
165 | 0 | if (tok1[0] == '/') { |
166 | 0 | useCMap(cache, tok1 + 1, usedCMaps); |
167 | 0 | } |
168 | 0 | pst->getToken(tok1, sizeof(tok1), &n1); |
169 | 0 | } else if (!strcmp(tok1, "/WMode")) { |
170 | 0 | wMode = atoi(tok2); |
171 | 0 | pst->getToken(tok1, sizeof(tok1), &n1); |
172 | 0 | } else if (!strcmp(tok2, "begincidchar")) { |
173 | 0 | while (pst->getToken(tok1, sizeof(tok1), &n1)) { |
174 | 0 | if (!strcmp(tok1, "endcidchar")) { |
175 | 0 | break; |
176 | 0 | } |
177 | 0 | if (!pst->getToken(tok2, sizeof(tok2), &n2) || |
178 | 0 | !strcmp(tok2, "endcidchar")) { |
179 | 0 | error(errSyntaxError, -1, "Illegal entry in cidchar block in CMap"); |
180 | 0 | break; |
181 | 0 | } |
182 | 0 | if (!(tok1[0] == '<' && tok1[n1 - 1] == '>' && |
183 | 0 | n1 >= 4 && (n1 & 1) == 0)) { |
184 | 0 | error(errSyntaxError, -1, "Illegal entry in cidchar block in CMap"); |
185 | 0 | continue; |
186 | 0 | } |
187 | 0 | tok1[n1 - 1] = '\0'; |
188 | 0 | if (sscanf(tok1 + 1, "%x", &code) != 1) { |
189 | 0 | error(errSyntaxError, -1, "Illegal entry in cidchar block in CMap"); |
190 | 0 | continue; |
191 | 0 | } |
192 | 0 | n1 = (n1 - 2) / 2; |
193 | 0 | addCIDs(code, code, n1, (CID)atoi(tok2)); |
194 | 0 | } |
195 | 0 | pst->getToken(tok1, sizeof(tok1), &n1); |
196 | 0 | } else if (!strcmp(tok2, "begincidrange")) { |
197 | 0 | while (pst->getToken(tok1, sizeof(tok1), &n1)) { |
198 | 0 | if (!strcmp(tok1, "endcidrange")) { |
199 | 0 | break; |
200 | 0 | } |
201 | 0 | if (!pst->getToken(tok2, sizeof(tok2), &n2) || |
202 | 0 | !strcmp(tok2, "endcidrange") || |
203 | 0 | !pst->getToken(tok3, sizeof(tok3), &n3) || |
204 | 0 | !strcmp(tok3, "endcidrange")) { |
205 | 0 | error(errSyntaxError, -1, "Illegal entry in cidrange block in CMap"); |
206 | 0 | break; |
207 | 0 | } |
208 | 0 | if (tok1[0] == '<' && tok2[0] == '<' && |
209 | 0 | n1 == n2 && n1 >= 4 && (n1 & 1) == 0) { |
210 | 0 | tok1[n1 - 1] = tok2[n1 - 1] = '\0'; |
211 | 0 | sscanf(tok1 + 1, "%x", &start); |
212 | 0 | sscanf(tok2 + 1, "%x", &end); |
213 | 0 | n1 = (n1 - 2) / 2; |
214 | 0 | addCIDs(start, end, n1, (CID)atoi(tok3)); |
215 | 0 | } |
216 | 0 | } |
217 | 0 | pst->getToken(tok1, sizeof(tok1), &n1); |
218 | 0 | } else { |
219 | 0 | strcpy(tok1, tok2); |
220 | 0 | } |
221 | 0 | } |
222 | 0 | delete pst; |
223 | 0 | } |
224 | | |
225 | 0 | CMap::CMap(GString *collectionA, GString *cMapNameA) { |
226 | 0 | int i; |
227 | |
|
228 | 0 | collection = collectionA; |
229 | 0 | cMapName = cMapNameA; |
230 | 0 | isIdent = gFalse; |
231 | 0 | wMode = 0; |
232 | 0 | vector = (CMapVectorEntry *)gmallocn(256, sizeof(CMapVectorEntry)); |
233 | 0 | for (i = 0; i < 256; ++i) { |
234 | 0 | vector[i].isVector = gFalse; |
235 | 0 | vector[i].cid = 0; |
236 | 0 | } |
237 | 0 | refCnt = 1; |
238 | 0 | } |
239 | | |
240 | 0 | CMap::CMap(GString *collectionA, GString *cMapNameA, int wModeA) { |
241 | 0 | collection = collectionA; |
242 | 0 | cMapName = cMapNameA; |
243 | 0 | isIdent = gTrue; |
244 | 0 | wMode = wModeA; |
245 | 0 | vector = NULL; |
246 | 0 | refCnt = 1; |
247 | 0 | } |
248 | | |
249 | 0 | void CMap::useCMap(CMapCache *cache, char *useName, GHash *usedCMaps) { |
250 | 0 | GString *useNameStr; |
251 | 0 | CMap *subCMap; |
252 | | |
253 | | // check for a loop |
254 | 0 | if (usedCMaps) { |
255 | 0 | if (usedCMaps->lookupInt(useName)) { |
256 | 0 | error(errSyntaxError, -1, "Loop in usecmap"); |
257 | 0 | return; |
258 | 0 | } |
259 | 0 | usedCMaps->add(new GString(useName), 1); |
260 | 0 | } |
261 | | |
262 | 0 | useNameStr = new GString(useName); |
263 | | // if cache is non-NULL, we already have a lock, and we can use |
264 | | // CMapCache::getCMap() directly; otherwise, we need to use |
265 | | // GlobalParams::getCMap() in order to acqure the lock |
266 | 0 | if (cache) { |
267 | 0 | subCMap = cache->getCMap(collection, useNameStr, usedCMaps); |
268 | 0 | } else { |
269 | 0 | subCMap = globalParams->getCMap(collection, useNameStr); |
270 | 0 | } |
271 | 0 | delete useNameStr; |
272 | 0 | if (!subCMap) { |
273 | 0 | return; |
274 | 0 | } |
275 | 0 | isIdent = subCMap->isIdent; |
276 | 0 | if (subCMap->vector) { |
277 | 0 | copyVector(vector, subCMap->vector); |
278 | 0 | } |
279 | 0 | subCMap->decRefCnt(); |
280 | 0 | } |
281 | | |
282 | 0 | void CMap::useCMap(CMapCache *cache, Object *obj, GHash *usedCMaps) { |
283 | 0 | CMap *subCMap; |
284 | |
|
285 | 0 | subCMap = CMap::parse(cache, collection, obj, usedCMaps); |
286 | 0 | if (!subCMap) { |
287 | 0 | return; |
288 | 0 | } |
289 | 0 | isIdent = subCMap->isIdent; |
290 | 0 | if (subCMap->vector) { |
291 | 0 | copyVector(vector, subCMap->vector); |
292 | 0 | } |
293 | 0 | subCMap->decRefCnt(); |
294 | 0 | } |
295 | | |
296 | 0 | void CMap::copyVector(CMapVectorEntry *dest, CMapVectorEntry *src) { |
297 | 0 | int i, j; |
298 | |
|
299 | 0 | for (i = 0; i < 256; ++i) { |
300 | 0 | if (src[i].isVector) { |
301 | 0 | if (!dest[i].isVector) { |
302 | 0 | dest[i].isVector = gTrue; |
303 | 0 | dest[i].vector = |
304 | 0 | (CMapVectorEntry *)gmallocn(256, sizeof(CMapVectorEntry)); |
305 | 0 | for (j = 0; j < 256; ++j) { |
306 | 0 | dest[i].vector[j].isVector = gFalse; |
307 | 0 | dest[i].vector[j].cid = 0; |
308 | 0 | } |
309 | 0 | } |
310 | 0 | copyVector(dest[i].vector, src[i].vector); |
311 | 0 | } else { |
312 | 0 | if (dest[i].isVector) { |
313 | 0 | error(errSyntaxError, -1, "Collision in usecmap"); |
314 | 0 | } else { |
315 | 0 | dest[i].cid = src[i].cid; |
316 | 0 | } |
317 | 0 | } |
318 | 0 | } |
319 | 0 | } |
320 | | |
321 | 0 | void CMap::addCIDs(Guint start, Guint end, Guint nBytes, CID firstCID) { |
322 | 0 | CMapVectorEntry *vec; |
323 | 0 | int byte, byte0, byte1; |
324 | 0 | Guint start1, end1, i, j, k; |
325 | |
|
326 | 0 | start1 = start & 0xffffff00; |
327 | 0 | end1 = end & 0xffffff00; |
328 | 0 | for (i = start1; i <= end1; i += 0x100) { |
329 | 0 | vec = vector; |
330 | 0 | for (j = nBytes - 1; j >= 1; --j) { |
331 | 0 | byte = (i >> (8 * j)) & 0xff; |
332 | 0 | if (!vec[byte].isVector) { |
333 | 0 | vec[byte].isVector = gTrue; |
334 | 0 | vec[byte].vector = |
335 | 0 | (CMapVectorEntry *)gmallocn(256, sizeof(CMapVectorEntry)); |
336 | 0 | for (k = 0; k < 256; ++k) { |
337 | 0 | vec[byte].vector[k].isVector = gFalse; |
338 | 0 | vec[byte].vector[k].cid = 0; |
339 | 0 | } |
340 | 0 | } |
341 | 0 | vec = vec[byte].vector; |
342 | 0 | } |
343 | 0 | byte0 = (i < start) ? (start & 0xff) : 0; |
344 | 0 | byte1 = (i + 0xff > end) ? (end & 0xff) : 0xff; |
345 | 0 | for (byte = byte0; byte <= byte1; ++byte) { |
346 | 0 | if (vec[byte].isVector) { |
347 | 0 | error(errSyntaxError, -1, "Invalid CID ({0:x} [{1:d} bytes]) in CMap", |
348 | 0 | i, nBytes); |
349 | 0 | } else { |
350 | 0 | vec[byte].cid = firstCID + ((i + byte) - start); |
351 | 0 | } |
352 | 0 | } |
353 | 0 | } |
354 | 0 | } |
355 | | |
356 | 0 | CMap::~CMap() { |
357 | 0 | delete collection; |
358 | 0 | if (cMapName) { |
359 | 0 | delete cMapName; |
360 | 0 | } |
361 | 0 | if (vector) { |
362 | 0 | freeCMapVector(vector); |
363 | 0 | } |
364 | 0 | } |
365 | | |
366 | 0 | void CMap::freeCMapVector(CMapVectorEntry *vec) { |
367 | 0 | int i; |
368 | |
|
369 | 0 | for (i = 0; i < 256; ++i) { |
370 | 0 | if (vec[i].isVector) { |
371 | 0 | freeCMapVector(vec[i].vector); |
372 | 0 | } |
373 | 0 | } |
374 | 0 | gfree(vec); |
375 | 0 | } |
376 | | |
377 | 0 | void CMap::incRefCnt() { |
378 | 0 | #if MULTITHREADED |
379 | 0 | gAtomicIncrement(&refCnt); |
380 | | #else |
381 | | ++refCnt; |
382 | | #endif |
383 | 0 | } |
384 | | |
385 | 0 | void CMap::decRefCnt() { |
386 | 0 | GBool done; |
387 | |
|
388 | 0 | #if MULTITHREADED |
389 | 0 | done = gAtomicDecrement(&refCnt) == 0; |
390 | | #else |
391 | | done = --refCnt == 0; |
392 | | #endif |
393 | 0 | if (done) { |
394 | 0 | delete this; |
395 | 0 | } |
396 | 0 | } |
397 | | |
398 | 0 | GBool CMap::match(GString *collectionA, GString *cMapNameA) { |
399 | 0 | return !collection->cmp(collectionA) && !cMapName->cmp(cMapNameA); |
400 | 0 | } |
401 | | |
402 | 0 | CID CMap::getCID(char *s, int len, CharCode *c, int *nUsed) { |
403 | 0 | CMapVectorEntry *vec; |
404 | 0 | CharCode cc; |
405 | 0 | int n, i; |
406 | |
|
407 | 0 | vec = vector; |
408 | 0 | cc = 0; |
409 | 0 | n = 0; |
410 | 0 | while (vec && n < len) { |
411 | 0 | i = s[n++] & 0xff; |
412 | 0 | cc = (cc << 8) | i; |
413 | 0 | if (!vec[i].isVector) { |
414 | 0 | *c = cc; |
415 | 0 | *nUsed = n; |
416 | 0 | return vec[i].cid; |
417 | 0 | } |
418 | 0 | vec = vec[i].vector; |
419 | 0 | } |
420 | 0 | if (isIdent && len >= 2) { |
421 | | // identity CMap |
422 | 0 | *nUsed = 2; |
423 | 0 | *c = cc = ((s[0] & 0xff) << 8) + (s[1] & 0xff); |
424 | 0 | return cc; |
425 | 0 | } |
426 | 0 | *nUsed = 1; |
427 | 0 | *c = s[0] & 0xff; |
428 | 0 | return 0; |
429 | 0 | } |
430 | | |
431 | | //------------------------------------------------------------------------ |
432 | | |
433 | 10.5k | CMapCache::CMapCache() { |
434 | 10.5k | int i; |
435 | | |
436 | 52.9k | for (i = 0; i < cMapCacheSize; ++i) { |
437 | 42.3k | cache[i] = NULL; |
438 | 42.3k | } |
439 | 10.5k | } |
440 | | |
441 | 10.5k | CMapCache::~CMapCache() { |
442 | 10.5k | int i; |
443 | | |
444 | 52.9k | for (i = 0; i < cMapCacheSize; ++i) { |
445 | 42.3k | if (cache[i]) { |
446 | 0 | cache[i]->decRefCnt(); |
447 | 0 | } |
448 | 42.3k | } |
449 | 10.5k | } |
450 | | |
451 | | CMap *CMapCache::getCMap(GString *collection, GString *cMapName, |
452 | 0 | GHash *usedCMaps) { |
453 | 0 | CMap *cmap; |
454 | 0 | int i, j; |
455 | |
|
456 | 0 | if (cache[0] && cache[0]->match(collection, cMapName)) { |
457 | 0 | cache[0]->incRefCnt(); |
458 | 0 | return cache[0]; |
459 | 0 | } |
460 | 0 | for (i = 1; i < cMapCacheSize; ++i) { |
461 | 0 | if (cache[i] && cache[i]->match(collection, cMapName)) { |
462 | 0 | cmap = cache[i]; |
463 | 0 | for (j = i; j >= 1; --j) { |
464 | 0 | cache[j] = cache[j - 1]; |
465 | 0 | } |
466 | 0 | cache[0] = cmap; |
467 | 0 | cmap->incRefCnt(); |
468 | 0 | return cmap; |
469 | 0 | } |
470 | 0 | } |
471 | 0 | if ((cmap = CMap::parse(this, collection, cMapName, usedCMaps))) { |
472 | 0 | if (cache[cMapCacheSize - 1]) { |
473 | 0 | cache[cMapCacheSize - 1]->decRefCnt(); |
474 | 0 | } |
475 | 0 | for (j = cMapCacheSize - 1; j >= 1; --j) { |
476 | 0 | cache[j] = cache[j - 1]; |
477 | 0 | } |
478 | 0 | cache[0] = cmap; |
479 | 0 | cmap->incRefCnt(); |
480 | 0 | return cmap; |
481 | 0 | } |
482 | 0 | return NULL; |
483 | 0 | } |