Go to most recent revision | Details | Last modification | View Log | RSS feed
Rev | Author | Line No. | Line |
---|---|---|---|
4680 | right-hear | 1 | /* |
2 | * PDF cleaning tool: general purpose pdf syntax washer. |
||
3 | * |
||
4 | * Rewrite PDF with pretty printed objects. |
||
5 | * Garbage collect unreachable objects. |
||
6 | * Inflate compressed streams. |
||
7 | * Create subset documents. |
||
8 | * |
||
9 | * TODO: linearize document for fast web view |
||
10 | */ |
||
11 | |||
12 | #include "fitz.h" |
||
13 | #include "mupdf.h" |
||
14 | |||
15 | static FILE *out = NULL; |
||
16 | |||
17 | static char *uselist = NULL; |
||
18 | static int *ofslist = NULL; |
||
19 | static int *genlist = NULL; |
||
20 | static int *renumbermap = NULL; |
||
21 | |||
22 | static int dogarbage = 0; |
||
23 | static int doexpand = 0; |
||
24 | static int doascii = 0; |
||
25 | |||
26 | static pdf_xref *xref = NULL; |
||
27 | |||
28 | void die(fz_error error) |
||
29 | { |
||
30 | fz_catch(error, "aborting"); |
||
31 | if (xref) |
||
32 | pdf_free_xref(xref); |
||
33 | exit(1); |
||
34 | } |
||
35 | |||
36 | static void usage(void) |
||
37 | { |
||
38 | fprintf(stderr, |
||
39 | "usage: pdfclean [options] input.pdf [output.pdf] [pages]\n" |
||
40 | "\t-p -\tpassword\n" |
||
41 | "\t-g\tgarbage collect unused objects\n" |
||
42 | "\t-gg\tin addition to -g compact xref table\n" |
||
43 | "\t-ggg\tin addition to -gg merge duplicate objects\n" |
||
44 | "\t-d\tdecompress streams\n" |
||
45 | "\t-a\tascii hex encode binary streams\n" |
||
46 | "\tpages\tcomma separated list of ranges\n"); |
||
47 | exit(1); |
||
48 | } |
||
49 | |||
50 | /* |
||
51 | * Garbage collect objects not reachable from the trailer. |
||
52 | */ |
||
53 | |||
54 | static void sweepref(fz_obj *ref); |
||
55 | |||
56 | static void sweepobj(fz_obj *obj) |
||
57 | { |
||
58 | int i; |
||
59 | |||
60 | if (fz_is_indirect(obj)) |
||
61 | sweepref(obj); |
||
62 | |||
63 | else if (fz_is_dict(obj)) |
||
64 | for (i = 0; i < fz_dict_len(obj); i++) |
||
65 | sweepobj(fz_dict_get_val(obj, i)); |
||
66 | |||
67 | else if (fz_is_array(obj)) |
||
68 | for (i = 0; i < fz_array_len(obj); i++) |
||
69 | sweepobj(fz_array_get(obj, i)); |
||
70 | } |
||
71 | |||
72 | static void sweepref(fz_obj *obj) |
||
73 | { |
||
74 | int num = fz_to_num(obj); |
||
75 | int gen = fz_to_gen(obj); |
||
76 | |||
77 | if (num < 0 || num >= xref->len) |
||
78 | return; |
||
79 | if (uselist[num]) |
||
80 | return; |
||
81 | |||
82 | uselist[num] = 1; |
||
83 | |||
84 | /* Bake in /Length in stream objects */ |
||
85 | if (pdf_is_stream(xref, num, gen)) |
||
86 | { |
||
87 | fz_obj *len = fz_dict_gets(obj, "Length"); |
||
88 | if (fz_is_indirect(len)) |
||
89 | { |
||
90 | uselist[fz_to_num(len)] = 0; |
||
91 | len = fz_resolve_indirect(len); |
||
92 | fz_dict_puts(obj, "Length", len); |
||
93 | } |
||
94 | } |
||
95 | |||
96 | sweepobj(fz_resolve_indirect(obj)); |
||
97 | } |
||
98 | |||
99 | /* |
||
100 | * Scan for and remove duplicate objects (slow) |
||
101 | */ |
||
102 | |||
103 | static void removeduplicateobjs(void) |
||
104 | { |
||
105 | int num, other; |
||
106 | |||
107 | for (num = 1; num < xref->len; num++) |
||
108 | { |
||
109 | /* Only compare an object to objects preceeding it */ |
||
110 | for (other = 1; other < num; other++) |
||
111 | { |
||
112 | fz_obj *a, *b; |
||
113 | |||
114 | if (num == other || !uselist[num] || !uselist[other]) |
||
115 | continue; |
||
116 | |||
117 | /* |
||
118 | * Comparing stream objects data contents would take too long. |
||
119 | * |
||
120 | * pdf_is_stream calls pdf_cache_object and ensures |
||
121 | * that the xref table has the objects loaded. |
||
122 | */ |
||
123 | if (pdf_is_stream(xref, num, 0) || pdf_is_stream(xref, other, 0)) |
||
124 | continue; |
||
125 | |||
126 | a = xref->table[num].obj; |
||
127 | b = xref->table[other].obj; |
||
128 | |||
129 | a = fz_resolve_indirect(a); |
||
130 | b = fz_resolve_indirect(b); |
||
131 | |||
132 | if (fz_objcmp(a, b)) |
||
133 | continue; |
||
134 | |||
135 | /* Keep the lowest numbered object */ |
||
136 | renumbermap[num] = MIN(num, other); |
||
137 | renumbermap[other] = MIN(num, other); |
||
138 | uselist[MAX(num, other)] = 0; |
||
139 | |||
140 | /* One duplicate was found, do not look for another */ |
||
141 | break; |
||
142 | } |
||
143 | } |
||
144 | } |
||
145 | |||
146 | /* |
||
147 | * Renumber objects sequentially so the xref is more compact |
||
148 | */ |
||
149 | |||
150 | static void compactxref(void) |
||
151 | { |
||
152 | int num, newnum; |
||
153 | |||
154 | /* |
||
155 | * Update renumbermap in-place, clustering all used |
||
156 | * objects together at low object ids. Objects that |
||
157 | * already should be renumbered will have their new |
||
158 | * object ids be updated to reflect the compaction. |
||
159 | */ |
||
160 | |||
161 | newnum = 1; |
||
162 | for (num = 1; num < xref->len; num++) |
||
163 | { |
||
164 | if (uselist[num] && renumbermap[num] == num) |
||
165 | renumbermap[num] = newnum++; |
||
166 | else if (renumbermap[num] != num) |
||
167 | renumbermap[num] = renumbermap[renumbermap[num]]; |
||
168 | } |
||
169 | } |
||
170 | |||
171 | /* |
||
172 | * Update indirect objects according to renumbering established when |
||
173 | * removing duplicate objects and compacting the xref. |
||
174 | */ |
||
175 | |||
176 | static void renumberobj(fz_obj *obj) |
||
177 | { |
||
178 | int i; |
||
179 | |||
180 | if (fz_is_dict(obj)) |
||
181 | { |
||
182 | for (i = 0; i < fz_dict_len(obj); i++) |
||
183 | { |
||
184 | fz_obj *key = fz_dict_get_key(obj, i); |
||
185 | fz_obj *val = fz_dict_get_val(obj, i); |
||
186 | if (fz_is_indirect(val)) |
||
187 | { |
||
188 | val = fz_new_indirect(renumbermap[fz_to_num(val)], 0, xref); |
||
189 | fz_dict_put(obj, key, val); |
||
190 | fz_drop_obj(val); |
||
191 | } |
||
192 | else |
||
193 | { |
||
194 | renumberobj(val); |
||
195 | } |
||
196 | } |
||
197 | } |
||
198 | |||
199 | else if (fz_is_array(obj)) |
||
200 | { |
||
201 | for (i = 0; i < fz_array_len(obj); i++) |
||
202 | { |
||
203 | fz_obj *val = fz_array_get(obj, i); |
||
204 | if (fz_is_indirect(val)) |
||
205 | { |
||
206 | val = fz_new_indirect(renumbermap[fz_to_num(val)], 0, xref); |
||
207 | fz_array_put(obj, i, val); |
||
208 | fz_drop_obj(val); |
||
209 | } |
||
210 | else |
||
211 | { |
||
212 | renumberobj(val); |
||
213 | } |
||
214 | } |
||
215 | } |
||
216 | } |
||
217 | |||
218 | static void renumberobjs(void) |
||
219 | { |
||
220 | pdf_xref_entry *oldxref; |
||
221 | int newlen; |
||
222 | int num; |
||
223 | |||
224 | /* Apply renumber map to indirect references in all objects in xref */ |
||
225 | renumberobj(xref->trailer); |
||
226 | for (num = 0; num < xref->len; num++) |
||
227 | { |
||
228 | fz_obj *obj = xref->table[num].obj; |
||
229 | |||
230 | if (fz_is_indirect(obj)) |
||
231 | { |
||
232 | obj = fz_new_indirect(renumbermap[fz_to_num(obj)], 0, xref); |
||
233 | pdf_update_object(xref, num, 0, obj); |
||
234 | fz_drop_obj(obj); |
||
235 | } |
||
236 | else |
||
237 | { |
||
238 | renumberobj(obj); |
||
239 | } |
||
240 | } |
||
241 | |||
242 | /* Create new table for the reordered, compacted xref */ |
||
243 | oldxref = xref->table; |
||
244 | xref->table = fz_calloc(xref->len, sizeof(pdf_xref_entry)); |
||
245 | xref->table[0] = oldxref[0]; |
||
246 | |||
247 | /* Move used objects into the new compacted xref */ |
||
248 | newlen = 0; |
||
249 | for (num = 1; num < xref->len; num++) |
||
250 | { |
||
251 | if (uselist[num]) |
||
252 | { |
||
253 | if (newlen < renumbermap[num]) |
||
254 | newlen = renumbermap[num]; |
||
255 | xref->table[renumbermap[num]] = oldxref[num]; |
||
256 | } |
||
257 | else |
||
258 | { |
||
259 | if (oldxref[num].obj) |
||
260 | fz_drop_obj(oldxref[num].obj); |
||
261 | } |
||
262 | } |
||
263 | |||
264 | fz_free(oldxref); |
||
265 | |||
266 | /* Update the used objects count in compacted xref */ |
||
267 | xref->len = newlen + 1; |
||
268 | |||
269 | /* Update list of used objects to fit with compacted xref */ |
||
270 | for (num = 1; num < xref->len; num++) |
||
271 | uselist[num] = 1; |
||
272 | } |
||
273 | |||
274 | /* |
||
275 | * Recreate page tree to only retain specified pages. |
||
276 | */ |
||
277 | |||
278 | static void retainpages(int argc, char **argv) |
||
279 | { |
||
280 | fz_error error; |
||
281 | fz_obj *oldroot, *root, *pages, *kids, *countobj, *parent; |
||
282 | |||
283 | /* Load the old page tree */ |
||
284 | error = pdf_load_page_tree(xref); |
||
285 | if (error) |
||
286 | die(fz_rethrow(error, "cannot load page tree")); |
||
287 | |||
288 | /* Keep only pages/type entry to avoid references to unretained pages */ |
||
289 | oldroot = fz_dict_gets(xref->trailer, "Root"); |
||
290 | pages = fz_dict_gets(oldroot, "Pages"); |
||
291 | |||
292 | root = fz_new_dict(2); |
||
293 | fz_dict_puts(root, "Type", fz_dict_gets(oldroot, "Type")); |
||
294 | fz_dict_puts(root, "Pages", fz_dict_gets(oldroot, "Pages")); |
||
295 | |||
296 | pdf_update_object(xref, fz_to_num(oldroot), fz_to_gen(oldroot), root); |
||
297 | |||
298 | fz_drop_obj(root); |
||
299 | |||
300 | /* Create a new kids array with only the pages we want to keep */ |
||
301 | parent = fz_new_indirect(fz_to_num(pages), fz_to_gen(pages), xref); |
||
302 | kids = fz_new_array(1); |
||
303 | |||
304 | /* Retain pages specified */ |
||
305 | while (argc - fz_optind) |
||
306 | { |
||
307 | int page, spage, epage; |
||
308 | char *spec, *dash; |
||
309 | char *pagelist = argv[fz_optind]; |
||
310 | |||
311 | spec = fz_strsep(&pagelist, ","); |
||
312 | while (spec) |
||
313 | { |
||
314 | dash = strchr(spec, '-'); |
||
315 | |||
316 | if (dash == spec) |
||
317 | spage = epage = pdf_count_pages(xref); |
||
318 | else |
||
319 | spage = epage = atoi(spec); |
||
320 | |||
321 | if (dash) |
||
322 | { |
||
323 | if (strlen(dash) > 1) |
||
324 | epage = atoi(dash + 1); |
||
325 | else |
||
326 | epage = pdf_count_pages(xref); |
||
327 | } |
||
328 | |||
329 | if (spage > epage) |
||
330 | page = spage, spage = epage, epage = page; |
||
331 | |||
332 | if (spage < 1) |
||
333 | spage = 1; |
||
334 | if (epage > pdf_count_pages(xref)) |
||
335 | epage = pdf_count_pages(xref); |
||
336 | |||
337 | for (page = spage; page <= epage; page++) |
||
338 | { |
||
339 | fz_obj *pageobj = xref->page_objs[page-1]; |
||
340 | fz_obj *pageref = xref->page_refs[page-1]; |
||
341 | |||
342 | fz_dict_puts(pageobj, "Parent", parent); |
||
343 | |||
344 | /* Store page object in new kids array */ |
||
345 | fz_array_push(kids, pageref); |
||
346 | } |
||
347 | |||
348 | spec = fz_strsep(&pagelist, ","); |
||
349 | } |
||
350 | |||
351 | fz_optind++; |
||
352 | } |
||
353 | |||
354 | fz_drop_obj(parent); |
||
355 | |||
356 | /* Update page count and kids array */ |
||
357 | countobj = fz_new_int(fz_array_len(kids)); |
||
358 | fz_dict_puts(pages, "Count", countobj); |
||
359 | fz_drop_obj(countobj); |
||
360 | fz_dict_puts(pages, "Kids", kids); |
||
361 | fz_drop_obj(kids); |
||
362 | } |
||
363 | |||
364 | /* |
||
365 | * Make sure we have loaded objects from object streams. |
||
366 | */ |
||
367 | |||
368 | static void preloadobjstms(void) |
||
369 | { |
||
370 | fz_error error; |
||
371 | fz_obj *obj; |
||
372 | int num; |
||
373 | |||
374 | for (num = 0; num < xref->len; num++) |
||
375 | { |
||
376 | if (xref->table[num].type == 'o') |
||
377 | { |
||
378 | error = pdf_load_object(&obj, xref, num, 0); |
||
379 | if (error) |
||
380 | die(error); |
||
381 | fz_drop_obj(obj); |
||
382 | } |
||
383 | } |
||
384 | } |
||
385 | |||
386 | /* |
||
387 | * Save streams and objects to the output |
||
388 | */ |
||
389 | |||
390 | static inline int isbinary(int c) |
||
391 | { |
||
392 | if (c == '\n' || c == '\r' || c == '\t') |
||
393 | return 0; |
||
394 | return c < 32 || c > 127; |
||
395 | } |
||
396 | |||
397 | static int isbinarystream(fz_buffer *buf) |
||
398 | { |
||
399 | int i; |
||
400 | for (i = 0; i < buf->len; i++) |
||
401 | if (isbinary(buf->data[i])) |
||
402 | return 1; |
||
403 | return 0; |
||
404 | } |
||
405 | |||
406 | static fz_buffer *hexbuf(unsigned char *p, int n) |
||
407 | { |
||
408 | static const char hex[16] = "0123456789abcdef"; |
||
409 | fz_buffer *buf; |
||
410 | int x = 0; |
||
411 | |||
412 | buf = fz_new_buffer(n * 2 + (n / 32) + 2); |
||
413 | |||
414 | while (n--) |
||
415 | { |
||
416 | buf->data[buf->len++] = hex[*p >> 4]; |
||
417 | buf->data[buf->len++] = hex[*p & 15]; |
||
418 | if (++x == 32) |
||
419 | { |
||
420 | buf->data[buf->len++] = '\n'; |
||
421 | x = 0; |
||
422 | } |
||
423 | p++; |
||
424 | } |
||
425 | |||
426 | buf->data[buf->len++] = '>'; |
||
427 | buf->data[buf->len++] = '\n'; |
||
428 | |||
429 | return buf; |
||
430 | } |
||
431 | |||
432 | static void addhexfilter(fz_obj *dict) |
||
433 | { |
||
434 | fz_obj *f, *dp, *newf, *newdp; |
||
435 | fz_obj *ahx, *nullobj; |
||
436 | |||
437 | ahx = fz_new_name("ASCIIHexDecode"); |
||
438 | nullobj = fz_new_null(); |
||
439 | newf = newdp = NULL; |
||
440 | |||
441 | f = fz_dict_gets(dict, "Filter"); |
||
442 | dp = fz_dict_gets(dict, "DecodeParms"); |
||
443 | |||
444 | if (fz_is_name(f)) |
||
445 | { |
||
446 | newf = fz_new_array(2); |
||
447 | fz_array_push(newf, ahx); |
||
448 | fz_array_push(newf, f); |
||
449 | f = newf; |
||
450 | if (fz_is_dict(dp)) |
||
451 | { |
||
452 | newdp = fz_new_array(2); |
||
453 | fz_array_push(newdp, nullobj); |
||
454 | fz_array_push(newdp, dp); |
||
455 | dp = newdp; |
||
456 | } |
||
457 | } |
||
458 | else if (fz_is_array(f)) |
||
459 | { |
||
460 | fz_array_insert(f, ahx); |
||
461 | if (fz_is_array(dp)) |
||
462 | fz_array_insert(dp, nullobj); |
||
463 | } |
||
464 | else |
||
465 | f = ahx; |
||
466 | |||
467 | fz_dict_puts(dict, "Filter", f); |
||
468 | if (dp) |
||
469 | fz_dict_puts(dict, "DecodeParms", dp); |
||
470 | |||
471 | fz_drop_obj(ahx); |
||
472 | fz_drop_obj(nullobj); |
||
473 | if (newf) |
||
474 | fz_drop_obj(newf); |
||
475 | if (newdp) |
||
476 | fz_drop_obj(newdp); |
||
477 | } |
||
478 | |||
479 | static void copystream(fz_obj *obj, int num, int gen) |
||
480 | { |
||
481 | fz_error error; |
||
482 | fz_buffer *buf, *tmp; |
||
483 | fz_obj *newlen; |
||
484 | |||
485 | error = pdf_load_raw_stream(&buf, xref, num, gen); |
||
486 | if (error) |
||
487 | die(error); |
||
488 | |||
489 | if (doascii && isbinarystream(buf)) |
||
490 | { |
||
491 | tmp = hexbuf(buf->data, buf->len); |
||
492 | fz_drop_buffer(buf); |
||
493 | buf = tmp; |
||
494 | |||
495 | addhexfilter(obj); |
||
496 | |||
497 | newlen = fz_new_int(buf->len); |
||
498 | fz_dict_puts(obj, "Length", newlen); |
||
499 | fz_drop_obj(newlen); |
||
500 | } |
||
501 | |||
502 | fprintf(out, "%d %d obj\n", num, gen); |
||
503 | fz_fprint_obj(out, obj, !doexpand); |
||
504 | fprintf(out, "stream\n"); |
||
505 | fwrite(buf->data, 1, buf->len, out); |
||
506 | fprintf(out, "endstream\nendobj\n\n"); |
||
507 | |||
508 | fz_drop_buffer(buf); |
||
509 | } |
||
510 | |||
511 | static void expandstream(fz_obj *obj, int num, int gen) |
||
512 | { |
||
513 | fz_error error; |
||
514 | fz_buffer *buf, *tmp; |
||
515 | fz_obj *newlen; |
||
516 | |||
517 | error = pdf_load_stream(&buf, xref, num, gen); |
||
518 | if (error) |
||
519 | die(error); |
||
520 | |||
521 | fz_dict_dels(obj, "Filter"); |
||
522 | fz_dict_dels(obj, "DecodeParms"); |
||
523 | |||
524 | if (doascii && isbinarystream(buf)) |
||
525 | { |
||
526 | tmp = hexbuf(buf->data, buf->len); |
||
527 | fz_drop_buffer(buf); |
||
528 | buf = tmp; |
||
529 | |||
530 | addhexfilter(obj); |
||
531 | } |
||
532 | |||
533 | newlen = fz_new_int(buf->len); |
||
534 | fz_dict_puts(obj, "Length", newlen); |
||
535 | fz_drop_obj(newlen); |
||
536 | |||
537 | fprintf(out, "%d %d obj\n", num, gen); |
||
538 | fz_fprint_obj(out, obj, !doexpand); |
||
539 | fprintf(out, "stream\n"); |
||
540 | fwrite(buf->data, 1, buf->len, out); |
||
541 | fprintf(out, "endstream\nendobj\n\n"); |
||
542 | |||
543 | fz_drop_buffer(buf); |
||
544 | } |
||
545 | |||
546 | static void writeobject(int num, int gen) |
||
547 | { |
||
548 | fz_error error; |
||
549 | fz_obj *obj; |
||
550 | fz_obj *type; |
||
551 | |||
552 | error = pdf_load_object(&obj, xref, num, gen); |
||
553 | if (error) |
||
554 | die(error); |
||
555 | |||
556 | /* skip ObjStm and XRef objects */ |
||
557 | if (fz_is_dict(obj)) |
||
558 | { |
||
559 | type = fz_dict_gets(obj, "Type"); |
||
560 | if (fz_is_name(type) && !strcmp(fz_to_name(type), "ObjStm")) |
||
561 | { |
||
562 | uselist[num] = 0; |
||
563 | fz_drop_obj(obj); |
||
564 | return; |
||
565 | } |
||
566 | if (fz_is_name(type) && !strcmp(fz_to_name(type), "XRef")) |
||
567 | { |
||
568 | uselist[num] = 0; |
||
569 | fz_drop_obj(obj); |
||
570 | return; |
||
571 | } |
||
572 | } |
||
573 | |||
574 | if (!pdf_is_stream(xref, num, gen)) |
||
575 | { |
||
576 | fprintf(out, "%d %d obj\n", num, gen); |
||
577 | fz_fprint_obj(out, obj, !doexpand); |
||
578 | fprintf(out, "endobj\n\n"); |
||
579 | } |
||
580 | else |
||
581 | { |
||
582 | if (doexpand && !pdf_is_jpx_image(obj)) |
||
583 | expandstream(obj, num, gen); |
||
584 | else |
||
585 | copystream(obj, num, gen); |
||
586 | } |
||
587 | |||
588 | fz_drop_obj(obj); |
||
589 | } |
||
590 | |||
591 | static void writexref(void) |
||
592 | { |
||
593 | fz_obj *trailer; |
||
594 | fz_obj *obj; |
||
595 | int startxref; |
||
596 | int num; |
||
597 | |||
598 | startxref = ftell(out); |
||
599 | |||
600 | fprintf(out, "xref\n0 %d\n", xref->len); |
||
601 | for (num = 0; num < xref->len; num++) |
||
602 | { |
||
603 | if (uselist[num]) |
||
604 | fprintf(out, "%010d %05d n \n", ofslist[num], genlist[num]); |
||
605 | else |
||
606 | fprintf(out, "%010d %05d f \n", ofslist[num], genlist[num]); |
||
607 | } |
||
608 | fprintf(out, "\n"); |
||
609 | |||
610 | trailer = fz_new_dict(5); |
||
611 | |||
612 | obj = fz_new_int(xref->len); |
||
613 | fz_dict_puts(trailer, "Size", obj); |
||
614 | fz_drop_obj(obj); |
||
615 | |||
616 | obj = fz_dict_gets(xref->trailer, "Info"); |
||
617 | if (obj) |
||
618 | fz_dict_puts(trailer, "Info", obj); |
||
619 | |||
620 | obj = fz_dict_gets(xref->trailer, "Root"); |
||
621 | if (obj) |
||
622 | fz_dict_puts(trailer, "Root", obj); |
||
623 | |||
624 | obj = fz_dict_gets(xref->trailer, "ID"); |
||
625 | if (obj) |
||
626 | fz_dict_puts(trailer, "ID", obj); |
||
627 | |||
628 | fprintf(out, "trailer\n"); |
||
629 | fz_fprint_obj(out, trailer, !doexpand); |
||
630 | fprintf(out, "\n"); |
||
631 | |||
632 | fz_drop_obj(trailer); |
||
633 | |||
634 | fprintf(out, "startxref\n%d\n%%%%EOF\n", startxref); |
||
635 | } |
||
636 | |||
637 | static void writepdf(void) |
||
638 | { |
||
639 | int lastfree; |
||
640 | int num; |
||
641 | |||
642 | for (num = 0; num < xref->len; num++) |
||
643 | { |
||
644 | if (xref->table[num].type == 'f') |
||
645 | genlist[num] = xref->table[num].gen; |
||
646 | if (xref->table[num].type == 'n') |
||
647 | genlist[num] = xref->table[num].gen; |
||
648 | if (xref->table[num].type == 'o') |
||
649 | genlist[num] = 0; |
||
650 | |||
651 | if (dogarbage && !uselist[num]) |
||
652 | continue; |
||
653 | |||
654 | if (xref->table[num].type == 'n' || xref->table[num].type == 'o') |
||
655 | { |
||
656 | uselist[num] = 1; |
||
657 | ofslist[num] = ftell(out); |
||
658 | writeobject(num, genlist[num]); |
||
659 | } |
||
660 | } |
||
661 | |||
662 | /* Construct linked list of free object slots */ |
||
663 | lastfree = 0; |
||
664 | for (num = 0; num < xref->len; num++) |
||
665 | { |
||
666 | if (!uselist[num]) |
||
667 | { |
||
668 | genlist[num]++; |
||
669 | ofslist[lastfree] = num; |
||
670 | lastfree = num; |
||
671 | } |
||
672 | } |
||
673 | |||
674 | writexref(); |
||
675 | } |
||
676 | |||
677 | int main(int argc, char **argv) |
||
678 | { |
||
679 | fz_error error; |
||
680 | char *infile; |
||
681 | char *outfile = "out.pdf"; |
||
682 | char *password = ""; |
||
683 | int c, num; |
||
684 | int subset; |
||
685 | |||
686 | while ((c = fz_getopt(argc, argv, "adgp:")) != -1) |
||
687 | { |
||
688 | switch (c) |
||
689 | { |
||
690 | case 'p': password = fz_optarg; break; |
||
691 | case 'g': dogarbage ++; break; |
||
692 | case 'd': doexpand ++; break; |
||
693 | case 'a': doascii ++; break; |
||
694 | default: usage(); break; |
||
695 | } |
||
696 | } |
||
697 | |||
698 | if (argc - fz_optind < 1) |
||
699 | usage(); |
||
700 | |||
701 | infile = argv[fz_optind++]; |
||
702 | |||
703 | if (argc - fz_optind > 0 && |
||
704 | (strstr(argv[fz_optind], ".pdf") || strstr(argv[fz_optind], ".PDF"))) |
||
705 | { |
||
706 | outfile = argv[fz_optind++]; |
||
707 | } |
||
708 | |||
709 | subset = 0; |
||
710 | if (argc - fz_optind > 0) |
||
711 | subset = 1; |
||
712 | |||
713 | error = pdf_open_xref(&xref, infile, password); |
||
714 | if (error) |
||
715 | die(fz_rethrow(error, "cannot open input file '%s'", infile)); |
||
716 | |||
717 | out = fopen(outfile, "wb"); |
||
718 | if (!out) |
||
719 | die(fz_throw("cannot open output file '%s'", outfile)); |
||
720 | |||
721 | fprintf(out, "%%PDF-%d.%d\n", xref->version / 10, xref->version % 10); |
||
722 | fprintf(out, "%%\316\274\341\277\246\n\n"); |
||
723 | |||
724 | uselist = fz_calloc(xref->len + 1, sizeof(char)); |
||
725 | ofslist = fz_calloc(xref->len + 1, sizeof(int)); |
||
726 | genlist = fz_calloc(xref->len + 1, sizeof(int)); |
||
727 | renumbermap = fz_calloc(xref->len + 1, sizeof(int)); |
||
728 | |||
729 | for (num = 0; num < xref->len; num++) |
||
730 | { |
||
731 | uselist[num] = 0; |
||
732 | ofslist[num] = 0; |
||
733 | genlist[num] = 0; |
||
734 | renumbermap[num] = num; |
||
735 | } |
||
736 | |||
737 | /* Make sure any objects hidden in compressed streams have been loaded */ |
||
738 | preloadobjstms(); |
||
739 | |||
740 | /* Only retain the specified subset of the pages */ |
||
741 | if (subset) |
||
742 | retainpages(argc, argv); |
||
743 | |||
744 | /* Sweep & mark objects from the trailer */ |
||
745 | if (dogarbage >= 1) |
||
746 | sweepobj(xref->trailer); |
||
747 | |||
748 | /* Coalesce and renumber duplicate objects */ |
||
749 | if (dogarbage >= 3) |
||
750 | removeduplicateobjs(); |
||
751 | |||
752 | /* Compact xref by renumbering and removing unused objects */ |
||
753 | if (dogarbage >= 2) |
||
754 | compactxref(); |
||
755 | |||
756 | /* Make renumbering affect all indirect references and update xref */ |
||
757 | if (dogarbage >= 2) |
||
758 | renumberobjs(); |
||
759 | |||
760 | writepdf(); |
||
761 | |||
762 | if (fclose(out)) |
||
763 | die(fz_throw("cannot close output file '%s'", outfile)); |
||
764 | |||
765 | fz_free(uselist); |
||
766 | fz_free(ofslist); |
||
767 | fz_free(genlist); |
||
768 | fz_free(renumbermap); |
||
769 | |||
770 | pdf_free_xref(xref); |
||
771 | |||
772 | fz_flush_warnings(); |
||
773 | |||
774 | return 0; |
||
775 | }>>>>>>>>=>>>>>>>>>>>>>> |