| 1 | #include "mupdf/fitz.h" |
| 2 | #include "mupdf/pdf.h" |
| 3 | |
| 4 | #include <stdlib.h> |
| 5 | #include <string.h> |
| 6 | #include <limits.h> |
| 7 | |
| 8 | int |
| 9 | pdf_count_pages(fz_context *ctx, pdf_document *doc) |
| 10 | { |
| 11 | /* FIXME: We should reset linear_page_count to 0 when editing starts |
| 12 | * (or when linear loading ends) */ |
| 13 | if (doc->linear_page_count != 0) |
| 14 | return doc->linear_page_count; |
| 15 | return pdf_to_int(ctx, pdf_dict_getp(ctx, pdf_trailer(ctx, doc), "Root/Pages/Count" )); |
| 16 | } |
| 17 | |
| 18 | static int |
| 19 | pdf_load_page_tree_imp(fz_context *ctx, pdf_document *doc, pdf_obj *node, int idx) |
| 20 | { |
| 21 | pdf_obj *type = pdf_dict_get(ctx, node, PDF_NAME(Type)); |
| 22 | if (pdf_name_eq(ctx, type, PDF_NAME(Pages))) |
| 23 | { |
| 24 | pdf_obj *kids = pdf_dict_get(ctx, node, PDF_NAME(Kids)); |
| 25 | int i, n = pdf_array_len(ctx, kids); |
| 26 | |
| 27 | if (pdf_mark_obj(ctx, node)) |
| 28 | fz_throw(ctx, FZ_ERROR_GENERIC, "cycle in page tree" ); |
| 29 | fz_try(ctx) |
| 30 | for (i = 0; i < n; ++i) |
| 31 | idx = pdf_load_page_tree_imp(ctx, doc, pdf_array_get(ctx, kids, i), idx); |
| 32 | fz_always(ctx) |
| 33 | pdf_unmark_obj(ctx, node); |
| 34 | fz_catch(ctx) |
| 35 | fz_rethrow(ctx); |
| 36 | } |
| 37 | else if (pdf_name_eq(ctx, type, PDF_NAME(Page))) |
| 38 | { |
| 39 | if (idx >= doc->rev_page_count) |
| 40 | fz_throw(ctx, FZ_ERROR_GENERIC, "too many kids in page tree" ); |
| 41 | doc->rev_page_map[idx].page = idx; |
| 42 | doc->rev_page_map[idx].object = pdf_to_num(ctx, node); |
| 43 | ++idx; |
| 44 | } |
| 45 | else |
| 46 | { |
| 47 | fz_throw(ctx, FZ_ERROR_GENERIC, "non-page object in page tree" ); |
| 48 | } |
| 49 | return idx; |
| 50 | } |
| 51 | |
| 52 | static int |
| 53 | cmp_rev_page_map(const void *va, const void *vb) |
| 54 | { |
| 55 | const pdf_rev_page_map *a = va; |
| 56 | const pdf_rev_page_map *b = vb; |
| 57 | return a->object - b->object; |
| 58 | } |
| 59 | |
| 60 | void |
| 61 | pdf_load_page_tree(fz_context *ctx, pdf_document *doc) |
| 62 | { |
| 63 | if (!doc->rev_page_map) |
| 64 | { |
| 65 | doc->rev_page_count = pdf_count_pages(ctx, doc); |
| 66 | doc->rev_page_map = fz_malloc_array(ctx, doc->rev_page_count, pdf_rev_page_map); |
| 67 | pdf_load_page_tree_imp(ctx, doc, pdf_dict_getp(ctx, pdf_trailer(ctx, doc), "Root/Pages" ), 0); |
| 68 | qsort(doc->rev_page_map, doc->rev_page_count, sizeof *doc->rev_page_map, cmp_rev_page_map); |
| 69 | } |
| 70 | } |
| 71 | |
| 72 | void |
| 73 | pdf_drop_page_tree(fz_context *ctx, pdf_document *doc) |
| 74 | { |
| 75 | fz_free(ctx, doc->rev_page_map); |
| 76 | doc->rev_page_map = NULL; |
| 77 | doc->rev_page_count = 0; |
| 78 | } |
| 79 | |
| 80 | enum |
| 81 | { |
| 82 | LOCAL_STACK_SIZE = 16 |
| 83 | }; |
| 84 | |
| 85 | static pdf_obj * |
| 86 | pdf_lookup_page_loc_imp(fz_context *ctx, pdf_document *doc, pdf_obj *node, int *skip, pdf_obj **parentp, int *indexp) |
| 87 | { |
| 88 | pdf_obj *kids; |
| 89 | pdf_obj *hit = NULL; |
| 90 | int i, len; |
| 91 | pdf_obj *local_stack[LOCAL_STACK_SIZE]; |
| 92 | pdf_obj **stack = &local_stack[0]; |
| 93 | int stack_max = LOCAL_STACK_SIZE; |
| 94 | int stack_len = 0; |
| 95 | |
| 96 | fz_var(hit); |
| 97 | fz_var(stack); |
| 98 | fz_var(stack_len); |
| 99 | fz_var(stack_max); |
| 100 | |
| 101 | fz_try(ctx) |
| 102 | { |
| 103 | do |
| 104 | { |
| 105 | kids = pdf_dict_get(ctx, node, PDF_NAME(Kids)); |
| 106 | len = pdf_array_len(ctx, kids); |
| 107 | |
| 108 | if (len == 0) |
| 109 | fz_throw(ctx, FZ_ERROR_GENERIC, "malformed page tree" ); |
| 110 | |
| 111 | /* Every node we need to unmark goes into the stack */ |
| 112 | if (stack_len == stack_max) |
| 113 | { |
| 114 | if (stack == &local_stack[0]) |
| 115 | { |
| 116 | stack = fz_malloc_array(ctx, stack_max * 2, pdf_obj*); |
| 117 | memcpy(stack, &local_stack[0], stack_max * sizeof(*stack)); |
| 118 | } |
| 119 | else |
| 120 | { |
| 121 | stack = fz_realloc_array(ctx, stack, stack_max * 2, pdf_obj*); |
| 122 | } |
| 123 | stack_max *= 2; |
| 124 | } |
| 125 | stack[stack_len++] = node; |
| 126 | |
| 127 | if (pdf_mark_obj(ctx, node)) |
| 128 | fz_throw(ctx, FZ_ERROR_GENERIC, "cycle in page tree" ); |
| 129 | |
| 130 | for (i = 0; i < len; i++) |
| 131 | { |
| 132 | pdf_obj *kid = pdf_array_get(ctx, kids, i); |
| 133 | pdf_obj *type = pdf_dict_get(ctx, kid, PDF_NAME(Type)); |
| 134 | if (type ? pdf_name_eq(ctx, type, PDF_NAME(Pages)) : pdf_dict_get(ctx, kid, PDF_NAME(Kids)) && !pdf_dict_get(ctx, kid, PDF_NAME(MediaBox))) |
| 135 | { |
| 136 | int count = pdf_dict_get_int(ctx, kid, PDF_NAME(Count)); |
| 137 | if (*skip < count) |
| 138 | { |
| 139 | node = kid; |
| 140 | break; |
| 141 | } |
| 142 | else |
| 143 | { |
| 144 | *skip -= count; |
| 145 | } |
| 146 | } |
| 147 | else |
| 148 | { |
| 149 | if (type ? !pdf_name_eq(ctx, type, PDF_NAME(Page)) : !pdf_dict_get(ctx, kid, PDF_NAME(MediaBox))) |
| 150 | fz_warn(ctx, "non-page object in page tree (%s)" , pdf_to_name(ctx, type)); |
| 151 | if (*skip == 0) |
| 152 | { |
| 153 | if (parentp) *parentp = node; |
| 154 | if (indexp) *indexp = i; |
| 155 | hit = kid; |
| 156 | break; |
| 157 | } |
| 158 | else |
| 159 | { |
| 160 | (*skip)--; |
| 161 | } |
| 162 | } |
| 163 | } |
| 164 | } |
| 165 | /* If i < len && hit != NULL the desired page was found in the |
| 166 | Kids array, done. If i < len && hit == NULL the found page tree |
| 167 | node contains a Kids array that contains the desired page, loop |
| 168 | back to top to extract it. When i == len the Kids array has been |
| 169 | exhausted without finding the desired page, give up. |
| 170 | */ |
| 171 | while (hit == NULL && i < len); |
| 172 | } |
| 173 | fz_always(ctx) |
| 174 | { |
| 175 | for (i = stack_len; i > 0; i--) |
| 176 | pdf_unmark_obj(ctx, stack[i-1]); |
| 177 | if (stack != &local_stack[0]) |
| 178 | fz_free(ctx, stack); |
| 179 | } |
| 180 | fz_catch(ctx) |
| 181 | { |
| 182 | fz_rethrow(ctx); |
| 183 | } |
| 184 | |
| 185 | return hit; |
| 186 | } |
| 187 | |
| 188 | pdf_obj * |
| 189 | pdf_lookup_page_loc(fz_context *ctx, pdf_document *doc, int needle, pdf_obj **parentp, int *indexp) |
| 190 | { |
| 191 | pdf_obj *root = pdf_dict_get(ctx, pdf_trailer(ctx, doc), PDF_NAME(Root)); |
| 192 | pdf_obj *node = pdf_dict_get(ctx, root, PDF_NAME(Pages)); |
| 193 | int skip = needle; |
| 194 | pdf_obj *hit; |
| 195 | |
| 196 | if (!node) |
| 197 | fz_throw(ctx, FZ_ERROR_GENERIC, "cannot find page tree" ); |
| 198 | |
| 199 | hit = pdf_lookup_page_loc_imp(ctx, doc, node, &skip, parentp, indexp); |
| 200 | if (!hit) |
| 201 | fz_throw(ctx, FZ_ERROR_GENERIC, "cannot find page %d in page tree" , needle+1); |
| 202 | return hit; |
| 203 | } |
| 204 | |
| 205 | pdf_obj * |
| 206 | pdf_lookup_page_obj(fz_context *ctx, pdf_document *doc, int needle) |
| 207 | { |
| 208 | return pdf_lookup_page_loc(ctx, doc, needle, NULL, NULL); |
| 209 | } |
| 210 | |
| 211 | static int |
| 212 | pdf_count_pages_before_kid(fz_context *ctx, pdf_document *doc, pdf_obj *parent, int kid_num) |
| 213 | { |
| 214 | pdf_obj *kids = pdf_dict_get(ctx, parent, PDF_NAME(Kids)); |
| 215 | int i, total = 0, len = pdf_array_len(ctx, kids); |
| 216 | for (i = 0; i < len; i++) |
| 217 | { |
| 218 | pdf_obj *kid = pdf_array_get(ctx, kids, i); |
| 219 | if (pdf_to_num(ctx, kid) == kid_num) |
| 220 | return total; |
| 221 | if (pdf_name_eq(ctx, pdf_dict_get(ctx, kid, PDF_NAME(Type)), PDF_NAME(Pages))) |
| 222 | { |
| 223 | pdf_obj *count = pdf_dict_get(ctx, kid, PDF_NAME(Count)); |
| 224 | int n = pdf_to_int(ctx, count); |
| 225 | if (!pdf_is_int(ctx, count) || n < 0) |
| 226 | fz_throw(ctx, FZ_ERROR_GENERIC, "illegal or missing count in pages tree" ); |
| 227 | total += n; |
| 228 | } |
| 229 | else |
| 230 | total++; |
| 231 | } |
| 232 | fz_throw(ctx, FZ_ERROR_GENERIC, "kid not found in parent's kids array" ); |
| 233 | } |
| 234 | |
| 235 | static int |
| 236 | pdf_lookup_page_number_slow(fz_context *ctx, pdf_document *doc, pdf_obj *node) |
| 237 | { |
| 238 | int needle = pdf_to_num(ctx, node); |
| 239 | int total = 0; |
| 240 | pdf_obj *parent, *parent2; |
| 241 | |
| 242 | if (!pdf_name_eq(ctx, pdf_dict_get(ctx, node, PDF_NAME(Type)), PDF_NAME(Page))) |
| 243 | fz_throw(ctx, FZ_ERROR_GENERIC, "invalid page object" ); |
| 244 | |
| 245 | parent2 = parent = pdf_dict_get(ctx, node, PDF_NAME(Parent)); |
| 246 | fz_var(parent); |
| 247 | fz_try(ctx) |
| 248 | { |
| 249 | while (pdf_is_dict(ctx, parent)) |
| 250 | { |
| 251 | if (pdf_mark_obj(ctx, parent)) |
| 252 | fz_throw(ctx, FZ_ERROR_GENERIC, "cycle in page tree (parents)" ); |
| 253 | total += pdf_count_pages_before_kid(ctx, doc, parent, needle); |
| 254 | needle = pdf_to_num(ctx, parent); |
| 255 | parent = pdf_dict_get(ctx, parent, PDF_NAME(Parent)); |
| 256 | } |
| 257 | } |
| 258 | fz_always(ctx) |
| 259 | { |
| 260 | /* Run back and unmark */ |
| 261 | while (parent2) |
| 262 | { |
| 263 | pdf_unmark_obj(ctx, parent2); |
| 264 | if (parent2 == parent) |
| 265 | break; |
| 266 | parent2 = pdf_dict_get(ctx, parent2, PDF_NAME(Parent)); |
| 267 | } |
| 268 | } |
| 269 | fz_catch(ctx) |
| 270 | { |
| 271 | fz_rethrow(ctx); |
| 272 | } |
| 273 | |
| 274 | return total; |
| 275 | } |
| 276 | |
| 277 | static int |
| 278 | pdf_lookup_page_number_fast(fz_context *ctx, pdf_document *doc, int needle) |
| 279 | { |
| 280 | int l = 0; |
| 281 | int r = doc->rev_page_count - 1; |
| 282 | while (l <= r) |
| 283 | { |
| 284 | int m = (l + r) >> 1; |
| 285 | int c = needle - doc->rev_page_map[m].object; |
| 286 | if (c < 0) |
| 287 | r = m - 1; |
| 288 | else if (c > 0) |
| 289 | l = m + 1; |
| 290 | else |
| 291 | return doc->rev_page_map[m].page; |
| 292 | } |
| 293 | return -1; |
| 294 | } |
| 295 | |
| 296 | int |
| 297 | pdf_lookup_page_number(fz_context *ctx, pdf_document *doc, pdf_obj *page) |
| 298 | { |
| 299 | if (doc->rev_page_map) |
| 300 | return pdf_lookup_page_number_fast(ctx, doc, pdf_to_num(ctx, page)); |
| 301 | else |
| 302 | return pdf_lookup_page_number_slow(ctx, doc, page); |
| 303 | } |
| 304 | |
| 305 | /* |
| 306 | Find the page number of a named destination. |
| 307 | |
| 308 | For use with looking up the destination page of a fragment |
| 309 | identifier in hyperlinks: foo.pdf#bar or foo.pdf#page=5. |
| 310 | */ |
| 311 | int |
| 312 | pdf_lookup_anchor(fz_context *ctx, pdf_document *doc, const char *name, float *xp, float *yp) |
| 313 | { |
| 314 | pdf_obj *needle, *dest = NULL; |
| 315 | char *uri; |
| 316 | |
| 317 | if (xp) *xp = 0; |
| 318 | if (yp) *yp = 0; |
| 319 | |
| 320 | needle = pdf_new_string(ctx, name, strlen(name)); |
| 321 | fz_try(ctx) |
| 322 | dest = pdf_lookup_dest(ctx, doc, needle); |
| 323 | fz_always(ctx) |
| 324 | pdf_drop_obj(ctx, needle); |
| 325 | fz_catch(ctx) |
| 326 | fz_rethrow(ctx); |
| 327 | |
| 328 | if (dest) |
| 329 | { |
| 330 | uri = pdf_parse_link_dest(ctx, doc, dest); |
| 331 | return pdf_resolve_link(ctx, doc, uri, xp, yp); |
| 332 | } |
| 333 | |
| 334 | if (!strncmp(name, "page=" , 5)) |
| 335 | return fz_atoi(name + 5) - 1; |
| 336 | |
| 337 | return fz_atoi(name) - 1; |
| 338 | } |
| 339 | |
| 340 | static void |
| 341 | pdf_flatten_inheritable_page_item(fz_context *ctx, pdf_obj *page, pdf_obj *key) |
| 342 | { |
| 343 | pdf_obj *val = pdf_dict_get_inheritable(ctx, page, key); |
| 344 | if (val) |
| 345 | pdf_dict_put(ctx, page, key, val); |
| 346 | } |
| 347 | |
| 348 | /* |
| 349 | Make page self sufficient. |
| 350 | |
| 351 | Copy any inheritable page keys into the actual page object, removing |
| 352 | any dependencies on the page tree parents. |
| 353 | */ |
| 354 | void |
| 355 | pdf_flatten_inheritable_page_items(fz_context *ctx, pdf_obj *page) |
| 356 | { |
| 357 | pdf_flatten_inheritable_page_item(ctx, page, PDF_NAME(MediaBox)); |
| 358 | pdf_flatten_inheritable_page_item(ctx, page, PDF_NAME(CropBox)); |
| 359 | pdf_flatten_inheritable_page_item(ctx, page, PDF_NAME(Rotate)); |
| 360 | pdf_flatten_inheritable_page_item(ctx, page, PDF_NAME(Resources)); |
| 361 | } |
| 362 | |
| 363 | /* We need to know whether to install a page-level transparency group */ |
| 364 | |
| 365 | /* |
| 366 | * Object memo flags - allows us to secretly remember "a memo" (a bool) in an |
| 367 | * object, and to read back whether there was a memo, and if so, what it was. |
| 368 | */ |
| 369 | enum |
| 370 | { |
| 371 | PDF_FLAGS_MEMO_BM = 0, |
| 372 | PDF_FLAGS_MEMO_OP = 1 |
| 373 | }; |
| 374 | |
| 375 | static int pdf_resources_use_blending(fz_context *ctx, pdf_obj *rdb); |
| 376 | |
| 377 | static int |
| 378 | pdf_extgstate_uses_blending(fz_context *ctx, pdf_obj *dict) |
| 379 | { |
| 380 | pdf_obj *obj = pdf_dict_get(ctx, dict, PDF_NAME(BM)); |
| 381 | if (obj && !pdf_name_eq(ctx, obj, PDF_NAME(Normal))) |
| 382 | return 1; |
| 383 | return 0; |
| 384 | } |
| 385 | |
| 386 | static int |
| 387 | pdf_pattern_uses_blending(fz_context *ctx, pdf_obj *dict) |
| 388 | { |
| 389 | pdf_obj *obj; |
| 390 | obj = pdf_dict_get(ctx, dict, PDF_NAME(Resources)); |
| 391 | if (pdf_resources_use_blending(ctx, obj)) |
| 392 | return 1; |
| 393 | obj = pdf_dict_get(ctx, dict, PDF_NAME(ExtGState)); |
| 394 | return pdf_extgstate_uses_blending(ctx, obj); |
| 395 | } |
| 396 | |
| 397 | static int |
| 398 | pdf_xobject_uses_blending(fz_context *ctx, pdf_obj *dict) |
| 399 | { |
| 400 | pdf_obj *obj = pdf_dict_get(ctx, dict, PDF_NAME(Resources)); |
| 401 | if (pdf_name_eq(ctx, pdf_dict_getp(ctx, dict, "Group/S" ), PDF_NAME(Transparency))) |
| 402 | return 1; |
| 403 | return pdf_resources_use_blending(ctx, obj); |
| 404 | } |
| 405 | |
| 406 | static int |
| 407 | pdf_resources_use_blending(fz_context *ctx, pdf_obj *rdb) |
| 408 | { |
| 409 | pdf_obj *obj; |
| 410 | int i, n, useBM = 0; |
| 411 | |
| 412 | if (!rdb) |
| 413 | return 0; |
| 414 | |
| 415 | /* Have we been here before and remembered an answer? */ |
| 416 | if (pdf_obj_memo(ctx, rdb, PDF_FLAGS_MEMO_BM, &useBM)) |
| 417 | return useBM; |
| 418 | |
| 419 | /* stop on cyclic resource dependencies */ |
| 420 | if (pdf_mark_obj(ctx, rdb)) |
| 421 | return 0; |
| 422 | |
| 423 | fz_try(ctx) |
| 424 | { |
| 425 | obj = pdf_dict_get(ctx, rdb, PDF_NAME(ExtGState)); |
| 426 | n = pdf_dict_len(ctx, obj); |
| 427 | for (i = 0; i < n; i++) |
| 428 | if (pdf_extgstate_uses_blending(ctx, pdf_dict_get_val(ctx, obj, i))) |
| 429 | goto found; |
| 430 | |
| 431 | obj = pdf_dict_get(ctx, rdb, PDF_NAME(Pattern)); |
| 432 | n = pdf_dict_len(ctx, obj); |
| 433 | for (i = 0; i < n; i++) |
| 434 | if (pdf_pattern_uses_blending(ctx, pdf_dict_get_val(ctx, obj, i))) |
| 435 | goto found; |
| 436 | |
| 437 | obj = pdf_dict_get(ctx, rdb, PDF_NAME(XObject)); |
| 438 | n = pdf_dict_len(ctx, obj); |
| 439 | for (i = 0; i < n; i++) |
| 440 | if (pdf_xobject_uses_blending(ctx, pdf_dict_get_val(ctx, obj, i))) |
| 441 | goto found; |
| 442 | if (0) |
| 443 | { |
| 444 | found: |
| 445 | useBM = 1; |
| 446 | } |
| 447 | } |
| 448 | fz_always(ctx) |
| 449 | { |
| 450 | pdf_unmark_obj(ctx, rdb); |
| 451 | } |
| 452 | fz_catch(ctx) |
| 453 | { |
| 454 | fz_rethrow(ctx); |
| 455 | } |
| 456 | |
| 457 | pdf_set_obj_memo(ctx, rdb, PDF_FLAGS_MEMO_BM, useBM); |
| 458 | return useBM; |
| 459 | } |
| 460 | |
| 461 | static int pdf_resources_use_overprint(fz_context *ctx, pdf_obj *rdb); |
| 462 | |
| 463 | static int |
| 464 | pdf_extgstate_uses_overprint(fz_context *ctx, pdf_obj *dict) |
| 465 | { |
| 466 | pdf_obj *obj = pdf_dict_get(ctx, dict, PDF_NAME(OP)); |
| 467 | if (obj && pdf_to_bool(ctx, obj)) |
| 468 | return 1; |
| 469 | return 0; |
| 470 | } |
| 471 | |
| 472 | static int |
| 473 | pdf_pattern_uses_overprint(fz_context *ctx, pdf_obj *dict) |
| 474 | { |
| 475 | pdf_obj *obj; |
| 476 | obj = pdf_dict_get(ctx, dict, PDF_NAME(Resources)); |
| 477 | if (pdf_resources_use_overprint(ctx, obj)) |
| 478 | return 1; |
| 479 | obj = pdf_dict_get(ctx, dict, PDF_NAME(ExtGState)); |
| 480 | return pdf_extgstate_uses_overprint(ctx, obj); |
| 481 | } |
| 482 | |
| 483 | static int |
| 484 | pdf_xobject_uses_overprint(fz_context *ctx, pdf_obj *dict) |
| 485 | { |
| 486 | pdf_obj *obj = pdf_dict_get(ctx, dict, PDF_NAME(Resources)); |
| 487 | return pdf_resources_use_overprint(ctx, obj); |
| 488 | } |
| 489 | |
| 490 | static int |
| 491 | pdf_resources_use_overprint(fz_context *ctx, pdf_obj *rdb) |
| 492 | { |
| 493 | pdf_obj *obj; |
| 494 | int i, n, useOP = 0; |
| 495 | |
| 496 | if (!rdb) |
| 497 | return 0; |
| 498 | |
| 499 | /* Have we been here before and remembered an answer? */ |
| 500 | if (pdf_obj_memo(ctx, rdb, PDF_FLAGS_MEMO_OP, &useOP)) |
| 501 | return useOP; |
| 502 | |
| 503 | /* stop on cyclic resource dependencies */ |
| 504 | if (pdf_mark_obj(ctx, rdb)) |
| 505 | return 0; |
| 506 | |
| 507 | fz_try(ctx) |
| 508 | { |
| 509 | obj = pdf_dict_get(ctx, rdb, PDF_NAME(ExtGState)); |
| 510 | n = pdf_dict_len(ctx, obj); |
| 511 | for (i = 0; i < n; i++) |
| 512 | if (pdf_extgstate_uses_overprint(ctx, pdf_dict_get_val(ctx, obj, i))) |
| 513 | goto found; |
| 514 | |
| 515 | obj = pdf_dict_get(ctx, rdb, PDF_NAME(Pattern)); |
| 516 | n = pdf_dict_len(ctx, obj); |
| 517 | for (i = 0; i < n; i++) |
| 518 | if (pdf_pattern_uses_overprint(ctx, pdf_dict_get_val(ctx, obj, i))) |
| 519 | goto found; |
| 520 | |
| 521 | obj = pdf_dict_get(ctx, rdb, PDF_NAME(XObject)); |
| 522 | n = pdf_dict_len(ctx, obj); |
| 523 | for (i = 0; i < n; i++) |
| 524 | if (pdf_xobject_uses_overprint(ctx, pdf_dict_get_val(ctx, obj, i))) |
| 525 | goto found; |
| 526 | if (0) |
| 527 | { |
| 528 | found: |
| 529 | useOP = 1; |
| 530 | } |
| 531 | } |
| 532 | fz_always(ctx) |
| 533 | { |
| 534 | pdf_unmark_obj(ctx, rdb); |
| 535 | } |
| 536 | fz_catch(ctx) |
| 537 | { |
| 538 | fz_rethrow(ctx); |
| 539 | } |
| 540 | |
| 541 | pdf_set_obj_memo(ctx, rdb, PDF_FLAGS_MEMO_OP, useOP); |
| 542 | return useOP; |
| 543 | } |
| 544 | |
| 545 | fz_transition * |
| 546 | pdf_page_presentation(fz_context *ctx, pdf_page *page, fz_transition *transition, float *duration) |
| 547 | { |
| 548 | pdf_obj *obj, *transdict; |
| 549 | |
| 550 | *duration = pdf_dict_get_real(ctx, page->obj, PDF_NAME(Dur)); |
| 551 | |
| 552 | transdict = pdf_dict_get(ctx, page->obj, PDF_NAME(Trans)); |
| 553 | if (!transdict) |
| 554 | return NULL; |
| 555 | |
| 556 | obj = pdf_dict_get(ctx, transdict, PDF_NAME(D)); |
| 557 | |
| 558 | transition->duration = (obj ? pdf_to_real(ctx, obj) : 1); |
| 559 | |
| 560 | transition->vertical = !pdf_name_eq(ctx, pdf_dict_get(ctx, transdict, PDF_NAME(Dm)), PDF_NAME(H)); |
| 561 | transition->outwards = !pdf_name_eq(ctx, pdf_dict_get(ctx, transdict, PDF_NAME(M)), PDF_NAME(I)); |
| 562 | /* FIXME: If 'Di' is None, it should be handled differently, but |
| 563 | * this only affects Fly, and we don't implement that currently. */ |
| 564 | transition->direction = (pdf_dict_get_int(ctx, transdict, PDF_NAME(Di))); |
| 565 | /* FIXME: Read SS for Fly when we implement it */ |
| 566 | /* FIXME: Read B for Fly when we implement it */ |
| 567 | |
| 568 | obj = pdf_dict_get(ctx, transdict, PDF_NAME(S)); |
| 569 | if (pdf_name_eq(ctx, obj, PDF_NAME(Split))) |
| 570 | transition->type = FZ_TRANSITION_SPLIT; |
| 571 | else if (pdf_name_eq(ctx, obj, PDF_NAME(Blinds))) |
| 572 | transition->type = FZ_TRANSITION_BLINDS; |
| 573 | else if (pdf_name_eq(ctx, obj, PDF_NAME(Box))) |
| 574 | transition->type = FZ_TRANSITION_BOX; |
| 575 | else if (pdf_name_eq(ctx, obj, PDF_NAME(Wipe))) |
| 576 | transition->type = FZ_TRANSITION_WIPE; |
| 577 | else if (pdf_name_eq(ctx, obj, PDF_NAME(Dissolve))) |
| 578 | transition->type = FZ_TRANSITION_DISSOLVE; |
| 579 | else if (pdf_name_eq(ctx, obj, PDF_NAME(Glitter))) |
| 580 | transition->type = FZ_TRANSITION_GLITTER; |
| 581 | else if (pdf_name_eq(ctx, obj, PDF_NAME(Fly))) |
| 582 | transition->type = FZ_TRANSITION_FLY; |
| 583 | else if (pdf_name_eq(ctx, obj, PDF_NAME(Push))) |
| 584 | transition->type = FZ_TRANSITION_PUSH; |
| 585 | else if (pdf_name_eq(ctx, obj, PDF_NAME(Cover))) |
| 586 | transition->type = FZ_TRANSITION_COVER; |
| 587 | else if (pdf_name_eq(ctx, obj, PDF_NAME(Uncover))) |
| 588 | transition->type = FZ_TRANSITION_UNCOVER; |
| 589 | else if (pdf_name_eq(ctx, obj, PDF_NAME(Fade))) |
| 590 | transition->type = FZ_TRANSITION_FADE; |
| 591 | else |
| 592 | transition->type = FZ_TRANSITION_NONE; |
| 593 | |
| 594 | return transition; |
| 595 | } |
| 596 | |
| 597 | /* |
| 598 | Determine the size of a page. |
| 599 | |
| 600 | Determine the page size in user space units, taking page rotation |
| 601 | into account. The page size is taken to be the crop box if it |
| 602 | exists (visible area after cropping), otherwise the media box will |
| 603 | be used (possibly including printing marks). |
| 604 | */ |
| 605 | fz_rect |
| 606 | pdf_bound_page(fz_context *ctx, pdf_page *page) |
| 607 | { |
| 608 | fz_matrix page_ctm; |
| 609 | fz_rect mediabox; |
| 610 | pdf_page_transform(ctx, page, &mediabox, &page_ctm); |
| 611 | return fz_transform_rect(mediabox, page_ctm); |
| 612 | } |
| 613 | |
| 614 | fz_link * |
| 615 | pdf_load_links(fz_context *ctx, pdf_page *page) |
| 616 | { |
| 617 | return fz_keep_link(ctx, page->links); |
| 618 | } |
| 619 | |
| 620 | pdf_obj * |
| 621 | pdf_page_resources(fz_context *ctx, pdf_page *page) |
| 622 | { |
| 623 | return pdf_dict_get_inheritable(ctx, page->obj, PDF_NAME(Resources)); |
| 624 | } |
| 625 | |
| 626 | pdf_obj * |
| 627 | pdf_page_contents(fz_context *ctx, pdf_page *page) |
| 628 | { |
| 629 | return pdf_dict_get(ctx, page->obj, PDF_NAME(Contents)); |
| 630 | } |
| 631 | |
| 632 | pdf_obj * |
| 633 | pdf_page_group(fz_context *ctx, pdf_page *page) |
| 634 | { |
| 635 | return pdf_dict_get(ctx, page->obj, PDF_NAME(Group)); |
| 636 | } |
| 637 | |
| 638 | void |
| 639 | pdf_page_obj_transform(fz_context *ctx, pdf_obj *pageobj, fz_rect *page_mediabox, fz_matrix *page_ctm) |
| 640 | { |
| 641 | pdf_obj *obj; |
| 642 | fz_rect mediabox, cropbox, realbox, pagebox; |
| 643 | float userunit = 1; |
| 644 | int rotate; |
| 645 | |
| 646 | if (!page_mediabox) |
| 647 | page_mediabox = &pagebox; |
| 648 | |
| 649 | obj = pdf_dict_get(ctx, pageobj, PDF_NAME(UserUnit)); |
| 650 | if (pdf_is_real(ctx, obj)) |
| 651 | userunit = pdf_to_real(ctx, obj); |
| 652 | |
| 653 | mediabox = pdf_to_rect(ctx, pdf_dict_get_inheritable(ctx, pageobj, PDF_NAME(MediaBox))); |
| 654 | if (fz_is_empty_rect(mediabox)) |
| 655 | { |
| 656 | mediabox.x0 = 0; |
| 657 | mediabox.y0 = 0; |
| 658 | mediabox.x1 = 612; |
| 659 | mediabox.y1 = 792; |
| 660 | } |
| 661 | |
| 662 | cropbox = pdf_to_rect(ctx, pdf_dict_get_inheritable(ctx, pageobj, PDF_NAME(CropBox))); |
| 663 | if (!fz_is_empty_rect(cropbox)) |
| 664 | mediabox = fz_intersect_rect(mediabox, cropbox); |
| 665 | |
| 666 | page_mediabox->x0 = fz_min(mediabox.x0, mediabox.x1); |
| 667 | page_mediabox->y0 = fz_min(mediabox.y0, mediabox.y1); |
| 668 | page_mediabox->x1 = fz_max(mediabox.x0, mediabox.x1); |
| 669 | page_mediabox->y1 = fz_max(mediabox.y0, mediabox.y1); |
| 670 | |
| 671 | if (page_mediabox->x1 - page_mediabox->x0 < 1 || page_mediabox->y1 - page_mediabox->y0 < 1) |
| 672 | *page_mediabox = fz_unit_rect; |
| 673 | |
| 674 | rotate = pdf_to_int(ctx, pdf_dict_get_inheritable(ctx, pageobj, PDF_NAME(Rotate))); |
| 675 | |
| 676 | /* Snap page rotation to 0, 90, 180 or 270 */ |
| 677 | if (rotate < 0) |
| 678 | rotate = 360 - ((-rotate) % 360); |
| 679 | if (rotate >= 360) |
| 680 | rotate = rotate % 360; |
| 681 | rotate = 90*((rotate + 45)/90); |
| 682 | if (rotate >= 360) |
| 683 | rotate = 0; |
| 684 | |
| 685 | /* Compute transform from fitz' page space (upper left page origin, y descending, 72 dpi) |
| 686 | * to PDF user space (arbitrary page origin, y ascending, UserUnit dpi). */ |
| 687 | |
| 688 | /* Make left-handed and scale by UserUnit */ |
| 689 | *page_ctm = fz_scale(userunit, -userunit); |
| 690 | |
| 691 | /* Rotate */ |
| 692 | *page_ctm = fz_pre_rotate(*page_ctm, -rotate); |
| 693 | |
| 694 | /* Translate page origin to 0,0 */ |
| 695 | realbox = fz_transform_rect(*page_mediabox, *page_ctm); |
| 696 | *page_ctm = fz_concat(*page_ctm, fz_translate(-realbox.x0, -realbox.y0)); |
| 697 | } |
| 698 | |
| 699 | void |
| 700 | pdf_page_transform(fz_context *ctx, pdf_page *page, fz_rect *page_mediabox, fz_matrix *page_ctm) |
| 701 | { |
| 702 | pdf_page_obj_transform(ctx, page->obj, page_mediabox, page_ctm); |
| 703 | } |
| 704 | |
| 705 | static void |
| 706 | find_seps(fz_context *ctx, fz_separations **seps, pdf_obj *obj, pdf_obj *clearme) |
| 707 | { |
| 708 | int i, n; |
| 709 | pdf_obj *nameobj; |
| 710 | |
| 711 | /* Indexed and DeviceN may have cyclic references */ |
| 712 | if (pdf_is_indirect(ctx, obj)) |
| 713 | { |
| 714 | if (pdf_mark_obj(ctx, obj)) |
| 715 | return; /* already been here */ |
| 716 | /* remember to clear this colorspace dictionary at the end */ |
| 717 | pdf_array_push(ctx, clearme, obj); |
| 718 | } |
| 719 | |
| 720 | nameobj = pdf_array_get(ctx, obj, 0); |
| 721 | if (pdf_name_eq(ctx, nameobj, PDF_NAME(Separation))) |
| 722 | { |
| 723 | fz_colorspace *cs; |
| 724 | const char *name = pdf_to_name(ctx, pdf_array_get(ctx, obj, 1)); |
| 725 | |
| 726 | /* Skip 'special' colorants. */ |
| 727 | if (!strcmp(name, "Black" ) || |
| 728 | !strcmp(name, "Cyan" ) || |
| 729 | !strcmp(name, "Magenta" ) || |
| 730 | !strcmp(name, "Yellow" ) || |
| 731 | !strcmp(name, "All" ) || |
| 732 | !strcmp(name, "None" )) |
| 733 | return; |
| 734 | |
| 735 | n = fz_count_separations(ctx, *seps); |
| 736 | for (i = 0; i < n; i++) |
| 737 | { |
| 738 | if (!strcmp(name, fz_separation_name(ctx, *seps, i))) |
| 739 | return; /* Got that one already */ |
| 740 | } |
| 741 | |
| 742 | fz_try(ctx) |
| 743 | cs = pdf_load_colorspace(ctx, obj); |
| 744 | fz_catch(ctx) |
| 745 | { |
| 746 | fz_rethrow_if(ctx, FZ_ERROR_TRYLATER); |
| 747 | return; /* ignore broken colorspace */ |
| 748 | } |
| 749 | fz_try(ctx) |
| 750 | { |
| 751 | if (!*seps) |
| 752 | *seps = fz_new_separations(ctx, 0); |
| 753 | fz_add_separation(ctx, *seps, name, cs, 0); |
| 754 | } |
| 755 | fz_always(ctx) |
| 756 | fz_drop_colorspace(ctx, cs); |
| 757 | fz_catch(ctx) |
| 758 | fz_rethrow(ctx); |
| 759 | } |
| 760 | else if (pdf_name_eq(ctx, nameobj, PDF_NAME(Indexed))) |
| 761 | { |
| 762 | find_seps(ctx, seps, pdf_array_get(ctx, obj, 1), clearme); |
| 763 | } |
| 764 | else if (pdf_name_eq(ctx, nameobj, PDF_NAME(DeviceN))) |
| 765 | { |
| 766 | /* If the separation colorants exists for this DeviceN color space |
| 767 | * add those prior to our search for DeviceN color */ |
| 768 | pdf_obj *cols = pdf_dict_get(ctx, pdf_array_get(ctx, obj, 4), PDF_NAME(Colorants)); |
| 769 | n = pdf_dict_len(ctx, cols); |
| 770 | for (i = 0; i < n; i++) |
| 771 | find_seps(ctx, seps, pdf_dict_get_val(ctx, cols, i), clearme); |
| 772 | } |
| 773 | } |
| 774 | |
| 775 | static void |
| 776 | find_devn(fz_context *ctx, fz_separations **seps, pdf_obj *obj, pdf_obj *clearme) |
| 777 | { |
| 778 | int i, j, n, m; |
| 779 | pdf_obj *arr; |
| 780 | pdf_obj *nameobj = pdf_array_get(ctx, obj, 0); |
| 781 | |
| 782 | if (!pdf_name_eq(ctx, nameobj, PDF_NAME(DeviceN))) |
| 783 | return; |
| 784 | |
| 785 | arr = pdf_array_get(ctx, obj, 1); |
| 786 | m = pdf_array_len(ctx, arr); |
| 787 | for (j = 0; j < m; j++) |
| 788 | { |
| 789 | fz_colorspace *cs; |
| 790 | const char *name = pdf_to_name(ctx, pdf_array_get(ctx, arr, j)); |
| 791 | |
| 792 | /* Skip 'special' colorants. */ |
| 793 | if (!strcmp(name, "Black" ) || |
| 794 | !strcmp(name, "Cyan" ) || |
| 795 | !strcmp(name, "Magenta" ) || |
| 796 | !strcmp(name, "Yellow" ) || |
| 797 | !strcmp(name, "All" ) || |
| 798 | !strcmp(name, "None" )) |
| 799 | continue; |
| 800 | |
| 801 | n = fz_count_separations(ctx, *seps); |
| 802 | for (i = 0; i < n; i++) |
| 803 | { |
| 804 | if (!strcmp(name, fz_separation_name(ctx, *seps, i))) |
| 805 | break; /* Got that one already */ |
| 806 | } |
| 807 | |
| 808 | if (i == n) |
| 809 | { |
| 810 | fz_try(ctx) |
| 811 | cs = pdf_load_colorspace(ctx, obj); |
| 812 | fz_catch(ctx) |
| 813 | { |
| 814 | fz_rethrow_if(ctx, FZ_ERROR_TRYLATER); |
| 815 | continue; /* ignore broken colorspace */ |
| 816 | } |
| 817 | fz_try(ctx) |
| 818 | { |
| 819 | if (!*seps) |
| 820 | *seps = fz_new_separations(ctx, 0); |
| 821 | fz_add_separation(ctx, *seps, name, cs, j); |
| 822 | } |
| 823 | fz_always(ctx) |
| 824 | fz_drop_colorspace(ctx, cs); |
| 825 | fz_catch(ctx) |
| 826 | fz_rethrow(ctx); |
| 827 | } |
| 828 | } |
| 829 | } |
| 830 | |
| 831 | typedef void (res_finder_fn)(fz_context *ctx, fz_separations **seps, pdf_obj *obj, pdf_obj *clearme); |
| 832 | |
| 833 | static void |
| 834 | scan_page_seps(fz_context *ctx, pdf_obj *res, fz_separations **seps, res_finder_fn *fn, pdf_obj *clearme) |
| 835 | { |
| 836 | pdf_obj *dict; |
| 837 | pdf_obj *obj; |
| 838 | int i, n; |
| 839 | |
| 840 | if (pdf_mark_obj(ctx, res)) |
| 841 | return; /* already been here */ |
| 842 | |
| 843 | /* remember to clear this resource dictionary at the end */ |
| 844 | pdf_array_push(ctx, clearme, res); |
| 845 | |
| 846 | dict = pdf_dict_get(ctx, res, PDF_NAME(ColorSpace)); |
| 847 | n = pdf_dict_len(ctx, dict); |
| 848 | for (i = 0; i < n; i++) |
| 849 | { |
| 850 | obj = pdf_dict_get_val(ctx, dict, i); |
| 851 | fn(ctx, seps, obj, clearme); |
| 852 | } |
| 853 | |
| 854 | dict = pdf_dict_get(ctx, res, PDF_NAME(Shading)); |
| 855 | n = pdf_dict_len(ctx, dict); |
| 856 | for (i = 0; i < n; i++) |
| 857 | { |
| 858 | obj = pdf_dict_get_val(ctx, dict, i); |
| 859 | fn(ctx, seps, pdf_dict_get(ctx, obj, PDF_NAME(ColorSpace)), clearme); |
| 860 | } |
| 861 | |
| 862 | dict = pdf_dict_get(ctx, res, PDF_NAME(XObject)); |
| 863 | n = pdf_dict_len(ctx, dict); |
| 864 | for (i = 0; i < n; i++) |
| 865 | { |
| 866 | obj = pdf_dict_get_val(ctx, dict, i); |
| 867 | fn(ctx, seps, pdf_dict_get(ctx, obj, PDF_NAME(ColorSpace)), clearme); |
| 868 | /* Recurse on XObject forms. */ |
| 869 | scan_page_seps(ctx, pdf_dict_get(ctx, obj, PDF_NAME(Resources)), seps, fn, clearme); |
| 870 | } |
| 871 | } |
| 872 | |
| 873 | /* |
| 874 | Get the separation details for a page. |
| 875 | */ |
| 876 | fz_separations * |
| 877 | pdf_page_separations(fz_context *ctx, pdf_page *page) |
| 878 | { |
| 879 | pdf_obj *res = pdf_page_resources(ctx, page); |
| 880 | pdf_obj *clearme = NULL; |
| 881 | fz_separations *seps = NULL; |
| 882 | |
| 883 | clearme = pdf_new_array(ctx, page->doc, 100); |
| 884 | fz_try(ctx) |
| 885 | { |
| 886 | /* Run through and look for separations first. This is |
| 887 | * because separations are simplest to deal with, and |
| 888 | * because DeviceN may be implemented on top of separations. |
| 889 | */ |
| 890 | scan_page_seps(ctx, res, &seps, find_seps, clearme); |
| 891 | } |
| 892 | fz_always(ctx) |
| 893 | { |
| 894 | int i, n = pdf_array_len(ctx, clearme); |
| 895 | for (i = 0; i < n; ++i) |
| 896 | pdf_unmark_obj(ctx, pdf_array_get(ctx, clearme, i)); |
| 897 | pdf_drop_obj(ctx, clearme); |
| 898 | } |
| 899 | fz_catch(ctx) |
| 900 | { |
| 901 | fz_drop_separations(ctx, seps); |
| 902 | fz_rethrow(ctx); |
| 903 | } |
| 904 | |
| 905 | clearme = pdf_new_array(ctx, page->doc, 100); |
| 906 | fz_try(ctx) |
| 907 | { |
| 908 | /* Now run through again, and look for DeviceNs. These may |
| 909 | * have spot colors in that aren't defined in terms of |
| 910 | * separations. */ |
| 911 | scan_page_seps(ctx, res, &seps, find_devn, clearme); |
| 912 | } |
| 913 | fz_always(ctx) |
| 914 | { |
| 915 | int i, n = pdf_array_len(ctx, clearme); |
| 916 | for (i = 0; i < n; ++i) |
| 917 | pdf_unmark_obj(ctx, pdf_array_get(ctx, clearme, i)); |
| 918 | pdf_drop_obj(ctx, clearme); |
| 919 | } |
| 920 | fz_catch(ctx) |
| 921 | { |
| 922 | fz_drop_separations(ctx, seps); |
| 923 | fz_rethrow(ctx); |
| 924 | } |
| 925 | |
| 926 | return seps; |
| 927 | } |
| 928 | |
| 929 | int |
| 930 | pdf_page_uses_overprint(fz_context *ctx, pdf_page *page) |
| 931 | { |
| 932 | return page ? page->overprint : 0; |
| 933 | } |
| 934 | |
| 935 | static void |
| 936 | pdf_drop_page_imp(fz_context *ctx, pdf_page *page) |
| 937 | { |
| 938 | fz_drop_link(ctx, page->links); |
| 939 | pdf_drop_annots(ctx, page->annots); |
| 940 | pdf_drop_widgets(ctx, page->widgets); |
| 941 | |
| 942 | pdf_drop_obj(ctx, page->obj); |
| 943 | |
| 944 | fz_drop_document(ctx, &page->doc->super); |
| 945 | } |
| 946 | |
| 947 | static pdf_page * |
| 948 | pdf_new_page(fz_context *ctx, pdf_document *doc) |
| 949 | { |
| 950 | pdf_page *page = fz_new_derived_page(ctx, pdf_page); |
| 951 | |
| 952 | page->doc = (pdf_document*) fz_keep_document(ctx, &doc->super); |
| 953 | |
| 954 | page->super.drop_page = (fz_page_drop_page_fn*)pdf_drop_page_imp; |
| 955 | page->super.load_links = (fz_page_load_links_fn*)pdf_load_links; |
| 956 | page->super.bound_page = (fz_page_bound_page_fn*)pdf_bound_page; |
| 957 | page->super.run_page_contents = (fz_page_run_page_fn*)pdf_run_page_contents; |
| 958 | page->super.run_page_annots = (fz_page_run_page_fn*)pdf_run_page_annots; |
| 959 | page->super.run_page_widgets = (fz_page_run_page_fn*)pdf_run_page_widgets; |
| 960 | page->super.page_presentation = (fz_page_page_presentation_fn*)pdf_page_presentation; |
| 961 | page->super.separations = (fz_page_separations_fn *)pdf_page_separations; |
| 962 | page->super.overprint = (fz_page_uses_overprint_fn *)pdf_page_uses_overprint; |
| 963 | |
| 964 | page->obj = NULL; |
| 965 | |
| 966 | page->transparency = 0; |
| 967 | page->links = NULL; |
| 968 | page->annots = NULL; |
| 969 | page->annot_tailp = &page->annots; |
| 970 | page->widgets = NULL; |
| 971 | page->widget_tailp = &page->widgets; |
| 972 | |
| 973 | return page; |
| 974 | } |
| 975 | |
| 976 | static void |
| 977 | pdf_load_default_colorspaces_imp(fz_context *ctx, fz_default_colorspaces *default_cs, pdf_obj *obj) |
| 978 | { |
| 979 | pdf_obj *cs_obj; |
| 980 | |
| 981 | /* The spec says to ignore any colors we can't understand */ |
| 982 | |
| 983 | cs_obj = pdf_dict_get(ctx, obj, PDF_NAME(DefaultGray)); |
| 984 | if (cs_obj) |
| 985 | { |
| 986 | fz_try(ctx) |
| 987 | { |
| 988 | fz_colorspace *cs = pdf_load_colorspace(ctx, cs_obj); |
| 989 | fz_set_default_gray(ctx, default_cs, cs); |
| 990 | fz_drop_colorspace(ctx, cs); |
| 991 | } |
| 992 | fz_catch(ctx) |
| 993 | fz_rethrow_if(ctx, FZ_ERROR_TRYLATER); |
| 994 | } |
| 995 | |
| 996 | cs_obj = pdf_dict_get(ctx, obj, PDF_NAME(DefaultRGB)); |
| 997 | if (cs_obj) |
| 998 | { |
| 999 | fz_try(ctx) |
| 1000 | { |
| 1001 | fz_colorspace *cs = pdf_load_colorspace(ctx, cs_obj); |
| 1002 | fz_set_default_rgb(ctx, default_cs, cs); |
| 1003 | fz_drop_colorspace(ctx, cs); |
| 1004 | } |
| 1005 | fz_catch(ctx) |
| 1006 | fz_rethrow_if(ctx, FZ_ERROR_TRYLATER); |
| 1007 | } |
| 1008 | |
| 1009 | cs_obj = pdf_dict_get(ctx, obj, PDF_NAME(DefaultCMYK)); |
| 1010 | if (cs_obj) |
| 1011 | { |
| 1012 | fz_try(ctx) |
| 1013 | { |
| 1014 | fz_colorspace *cs = pdf_load_colorspace(ctx, cs_obj); |
| 1015 | fz_set_default_cmyk(ctx, default_cs, cs); |
| 1016 | fz_drop_colorspace(ctx, cs); |
| 1017 | } |
| 1018 | fz_catch(ctx) |
| 1019 | fz_rethrow_if(ctx, FZ_ERROR_TRYLATER); |
| 1020 | } |
| 1021 | } |
| 1022 | |
| 1023 | fz_default_colorspaces * |
| 1024 | pdf_load_default_colorspaces(fz_context *ctx, pdf_document *doc, pdf_page *page) |
| 1025 | { |
| 1026 | pdf_obj *res; |
| 1027 | pdf_obj *obj; |
| 1028 | fz_default_colorspaces *default_cs; |
| 1029 | fz_colorspace *oi; |
| 1030 | |
| 1031 | default_cs = fz_new_default_colorspaces(ctx); |
| 1032 | |
| 1033 | fz_try(ctx) |
| 1034 | { |
| 1035 | res = pdf_page_resources(ctx, page); |
| 1036 | obj = pdf_dict_get(ctx, res, PDF_NAME(ColorSpace)); |
| 1037 | if (obj) |
| 1038 | pdf_load_default_colorspaces_imp(ctx, default_cs, obj); |
| 1039 | |
| 1040 | oi = pdf_document_output_intent(ctx, doc); |
| 1041 | if (oi) |
| 1042 | fz_set_default_output_intent(ctx, default_cs, oi); |
| 1043 | } |
| 1044 | fz_catch(ctx) |
| 1045 | { |
| 1046 | if (fz_caught(ctx) != FZ_ERROR_TRYLATER) |
| 1047 | { |
| 1048 | fz_drop_default_colorspaces(ctx, default_cs); |
| 1049 | fz_rethrow(ctx); |
| 1050 | } |
| 1051 | page->super.incomplete = 1; |
| 1052 | } |
| 1053 | |
| 1054 | return default_cs; |
| 1055 | } |
| 1056 | |
| 1057 | /* |
| 1058 | Update default colorspaces for an xobject. |
| 1059 | */ |
| 1060 | fz_default_colorspaces * |
| 1061 | pdf_update_default_colorspaces(fz_context *ctx, fz_default_colorspaces *old_cs, pdf_obj *res) |
| 1062 | { |
| 1063 | pdf_obj *obj; |
| 1064 | fz_default_colorspaces *new_cs; |
| 1065 | |
| 1066 | obj = pdf_dict_get(ctx, res, PDF_NAME(ColorSpace)); |
| 1067 | if (!obj) |
| 1068 | return fz_keep_default_colorspaces(ctx, old_cs); |
| 1069 | |
| 1070 | new_cs = fz_clone_default_colorspaces(ctx, old_cs); |
| 1071 | fz_try(ctx) |
| 1072 | pdf_load_default_colorspaces_imp(ctx, new_cs, obj); |
| 1073 | fz_catch(ctx) |
| 1074 | { |
| 1075 | fz_drop_default_colorspaces(ctx, new_cs); |
| 1076 | fz_rethrow(ctx); |
| 1077 | } |
| 1078 | |
| 1079 | return new_cs; |
| 1080 | } |
| 1081 | |
| 1082 | /* |
| 1083 | Load a page and its resources. |
| 1084 | |
| 1085 | Locates the page in the PDF document and loads the page and its |
| 1086 | resources. After pdf_load_page is it possible to retrieve the size |
| 1087 | of the page using pdf_bound_page, or to render the page using |
| 1088 | pdf_run_page_*. |
| 1089 | |
| 1090 | number: page number, where 0 is the first page of the document. |
| 1091 | */ |
| 1092 | pdf_page * |
| 1093 | pdf_load_page(fz_context *ctx, pdf_document *doc, int number) |
| 1094 | { |
| 1095 | pdf_page *page; |
| 1096 | pdf_annot *annot; |
| 1097 | pdf_obj *pageobj, *obj; |
| 1098 | |
| 1099 | if (doc->file_reading_linearly) |
| 1100 | { |
| 1101 | pageobj = pdf_progressive_advance(ctx, doc, number); |
| 1102 | if (pageobj == NULL) |
| 1103 | fz_throw(ctx, FZ_ERROR_TRYLATER, "page %d not available yet" , number); |
| 1104 | } |
| 1105 | else |
| 1106 | pageobj = pdf_lookup_page_obj(ctx, doc, number); |
| 1107 | |
| 1108 | page = pdf_new_page(ctx, doc); |
| 1109 | page->obj = pdf_keep_obj(ctx, pageobj); |
| 1110 | |
| 1111 | /* Pre-load annotations and links */ |
| 1112 | fz_try(ctx) |
| 1113 | { |
| 1114 | obj = pdf_dict_get(ctx, pageobj, PDF_NAME(Annots)); |
| 1115 | if (obj) |
| 1116 | { |
| 1117 | fz_rect page_mediabox; |
| 1118 | fz_matrix page_ctm; |
| 1119 | pdf_page_transform(ctx, page, &page_mediabox, &page_ctm); |
| 1120 | page->links = pdf_load_link_annots(ctx, doc, obj, number, page_ctm); |
| 1121 | pdf_load_annots(ctx, page, obj); |
| 1122 | } |
| 1123 | } |
| 1124 | fz_catch(ctx) |
| 1125 | { |
| 1126 | if (fz_caught(ctx) != FZ_ERROR_TRYLATER) |
| 1127 | { |
| 1128 | fz_drop_page(ctx, &page->super); |
| 1129 | fz_rethrow(ctx); |
| 1130 | } |
| 1131 | page->super.incomplete = 1; |
| 1132 | fz_drop_link(ctx, page->links); |
| 1133 | page->links = NULL; |
| 1134 | } |
| 1135 | |
| 1136 | /* Scan for transparency and overprint */ |
| 1137 | fz_try(ctx) |
| 1138 | { |
| 1139 | pdf_obj *resources = pdf_page_resources(ctx, page); |
| 1140 | if (pdf_name_eq(ctx, pdf_dict_getp(ctx, pageobj, "Group/S" ), PDF_NAME(Transparency))) |
| 1141 | page->transparency = 1; |
| 1142 | else if (pdf_resources_use_blending(ctx, resources)) |
| 1143 | page->transparency = 1; |
| 1144 | for (annot = page->annots; annot && !page->transparency; annot = annot->next) |
| 1145 | if (annot->ap && pdf_resources_use_blending(ctx, pdf_xobject_resources(ctx, annot->ap))) |
| 1146 | page->transparency = 1; |
| 1147 | if (pdf_resources_use_overprint(ctx, resources)) |
| 1148 | page->overprint = 1; |
| 1149 | for (annot = page->annots; annot && !page->overprint; annot = annot->next) |
| 1150 | if (annot->ap && pdf_resources_use_overprint(ctx, pdf_xobject_resources(ctx, annot->ap))) |
| 1151 | page->overprint = 1; |
| 1152 | } |
| 1153 | fz_catch(ctx) |
| 1154 | { |
| 1155 | if (fz_caught(ctx) != FZ_ERROR_TRYLATER) |
| 1156 | { |
| 1157 | fz_drop_page(ctx, &page->super); |
| 1158 | fz_rethrow(ctx); |
| 1159 | } |
| 1160 | page->super.incomplete = 1; |
| 1161 | } |
| 1162 | |
| 1163 | return page; |
| 1164 | } |
| 1165 | |
| 1166 | /* |
| 1167 | Delete a page from the page tree of |
| 1168 | a document. This does not remove the page contents |
| 1169 | or resources from the file. |
| 1170 | |
| 1171 | doc: The document to operate on. |
| 1172 | |
| 1173 | number: The page to remove (numbered from 0) |
| 1174 | */ |
| 1175 | void |
| 1176 | pdf_delete_page(fz_context *ctx, pdf_document *doc, int at) |
| 1177 | { |
| 1178 | pdf_obj *parent, *kids; |
| 1179 | int i; |
| 1180 | |
| 1181 | pdf_lookup_page_loc(ctx, doc, at, &parent, &i); |
| 1182 | kids = pdf_dict_get(ctx, parent, PDF_NAME(Kids)); |
| 1183 | pdf_array_delete(ctx, kids, i); |
| 1184 | |
| 1185 | while (parent) |
| 1186 | { |
| 1187 | int count = pdf_dict_get_int(ctx, parent, PDF_NAME(Count)); |
| 1188 | pdf_dict_put_int(ctx, parent, PDF_NAME(Count), count - 1); |
| 1189 | parent = pdf_dict_get(ctx, parent, PDF_NAME(Parent)); |
| 1190 | } |
| 1191 | } |
| 1192 | |
| 1193 | /* |
| 1194 | Delete a range of pages from the |
| 1195 | page tree of a document. This does not remove the page |
| 1196 | contents or resources from the file. |
| 1197 | |
| 1198 | doc: The document to operate on. |
| 1199 | |
| 1200 | start, end: The range of pages (numbered from 0) |
| 1201 | (inclusive, exclusive) to remove. If end is negative or |
| 1202 | greater than the number of pages in the document, it |
| 1203 | will be taken to be the end of the document. |
| 1204 | */ |
| 1205 | void |
| 1206 | pdf_delete_page_range(fz_context *ctx, pdf_document *doc, int start, int end) |
| 1207 | { |
| 1208 | int count = pdf_count_pages(ctx, doc); |
| 1209 | |
| 1210 | if (end < 0 || end > count) |
| 1211 | end = count+1; |
| 1212 | if (start < 0) |
| 1213 | start = 0; |
| 1214 | while (start < end) |
| 1215 | { |
| 1216 | pdf_delete_page(ctx, doc, start); |
| 1217 | end--; |
| 1218 | } |
| 1219 | } |
| 1220 | |
| 1221 | /* |
| 1222 | Create a pdf_obj within a document that |
| 1223 | represents a page, from a previously created resources |
| 1224 | dictionary and page content stream. This should then be |
| 1225 | inserted into the document using pdf_insert_page. |
| 1226 | |
| 1227 | After this call the page exists within the document |
| 1228 | structure, but is not actually ever displayed as it is |
| 1229 | not linked into the PDF page tree. |
| 1230 | |
| 1231 | doc: The document to which to add the page. |
| 1232 | |
| 1233 | mediabox: The mediabox for the page (should be identical |
| 1234 | to that used when creating the resources/contents). |
| 1235 | |
| 1236 | rotate: 0, 90, 180 or 270. The rotation to use for the |
| 1237 | page. |
| 1238 | |
| 1239 | resources: The resources dictionary for the new page |
| 1240 | (typically created by pdf_page_write). |
| 1241 | |
| 1242 | contents: The page contents for the new page (typically |
| 1243 | create by pdf_page_write). |
| 1244 | */ |
| 1245 | pdf_obj * |
| 1246 | pdf_add_page(fz_context *ctx, pdf_document *doc, fz_rect mediabox, int rotate, pdf_obj *resources, fz_buffer *contents) |
| 1247 | { |
| 1248 | pdf_obj *page_obj = pdf_new_dict(ctx, doc, 5); |
| 1249 | fz_try(ctx) |
| 1250 | { |
| 1251 | pdf_dict_put(ctx, page_obj, PDF_NAME(Type), PDF_NAME(Page)); |
| 1252 | pdf_dict_put_rect(ctx, page_obj, PDF_NAME(MediaBox), mediabox); |
| 1253 | pdf_dict_put_int(ctx, page_obj, PDF_NAME(Rotate), rotate); |
| 1254 | |
| 1255 | if (pdf_is_indirect(ctx, resources)) |
| 1256 | pdf_dict_put(ctx, page_obj, PDF_NAME(Resources), resources); |
| 1257 | else if (pdf_is_dict(ctx, resources)) |
| 1258 | pdf_dict_put_drop(ctx, page_obj, PDF_NAME(Resources), pdf_add_object(ctx, doc, resources)); |
| 1259 | else |
| 1260 | pdf_dict_put_dict(ctx, page_obj, PDF_NAME(Resources), 1); |
| 1261 | |
| 1262 | if (contents) |
| 1263 | pdf_dict_put_drop(ctx, page_obj, PDF_NAME(Contents), pdf_add_stream(ctx, doc, contents, NULL, 0)); |
| 1264 | } |
| 1265 | fz_catch(ctx) |
| 1266 | { |
| 1267 | pdf_drop_obj(ctx, page_obj); |
| 1268 | fz_rethrow(ctx); |
| 1269 | } |
| 1270 | return pdf_add_object_drop(ctx, doc, page_obj); |
| 1271 | } |
| 1272 | |
| 1273 | /* |
| 1274 | Insert a page previously created by |
| 1275 | pdf_add_page into the pages tree of the document. |
| 1276 | |
| 1277 | doc: The document to insert into. |
| 1278 | |
| 1279 | at: The page number to insert at. 0 inserts at the start. |
| 1280 | negative numbers, or INT_MAX insert at the end. Otherwise |
| 1281 | n inserts after page n. |
| 1282 | |
| 1283 | page: The page to insert. |
| 1284 | */ |
| 1285 | void |
| 1286 | pdf_insert_page(fz_context *ctx, pdf_document *doc, int at, pdf_obj *page_ref) |
| 1287 | { |
| 1288 | int count = pdf_count_pages(ctx, doc); |
| 1289 | pdf_obj *parent, *kids; |
| 1290 | int i; |
| 1291 | |
| 1292 | if (at < 0) |
| 1293 | at = count; |
| 1294 | if (at == INT_MAX) |
| 1295 | at = count; |
| 1296 | if (at > count) |
| 1297 | fz_throw(ctx, FZ_ERROR_GENERIC, "cannot insert page beyond end of page tree" ); |
| 1298 | |
| 1299 | if (count == 0) |
| 1300 | { |
| 1301 | pdf_obj *root = pdf_dict_get(ctx, pdf_trailer(ctx, doc), PDF_NAME(Root)); |
| 1302 | parent = pdf_dict_get(ctx, root, PDF_NAME(Pages)); |
| 1303 | if (!parent) |
| 1304 | fz_throw(ctx, FZ_ERROR_GENERIC, "cannot find page tree" ); |
| 1305 | kids = pdf_dict_get(ctx, parent, PDF_NAME(Kids)); |
| 1306 | if (!kids) |
| 1307 | fz_throw(ctx, FZ_ERROR_GENERIC, "malformed page tree" ); |
| 1308 | pdf_array_insert(ctx, kids, page_ref, 0); |
| 1309 | } |
| 1310 | else if (at == count) |
| 1311 | { |
| 1312 | /* append after last page */ |
| 1313 | pdf_lookup_page_loc(ctx, doc, count - 1, &parent, &i); |
| 1314 | kids = pdf_dict_get(ctx, parent, PDF_NAME(Kids)); |
| 1315 | pdf_array_insert(ctx, kids, page_ref, i + 1); |
| 1316 | } |
| 1317 | else |
| 1318 | { |
| 1319 | /* insert before found page */ |
| 1320 | pdf_lookup_page_loc(ctx, doc, at, &parent, &i); |
| 1321 | kids = pdf_dict_get(ctx, parent, PDF_NAME(Kids)); |
| 1322 | pdf_array_insert(ctx, kids, page_ref, i); |
| 1323 | } |
| 1324 | |
| 1325 | pdf_dict_put(ctx, page_ref, PDF_NAME(Parent), parent); |
| 1326 | |
| 1327 | /* Adjust page counts */ |
| 1328 | while (parent) |
| 1329 | { |
| 1330 | count = pdf_dict_get_int(ctx, parent, PDF_NAME(Count)); |
| 1331 | pdf_dict_put_int(ctx, parent, PDF_NAME(Count), count + 1); |
| 1332 | parent = pdf_dict_get(ctx, parent, PDF_NAME(Parent)); |
| 1333 | } |
| 1334 | } |
| 1335 | |