Document.cpp 24 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635636637638639640641642643644645646
  1. /*
  2. * Copyright (c) 2021-2022, Matthew Olsson <mattco@serenityos.org>
  3. *
  4. * SPDX-License-Identifier: BSD-2-Clause
  5. */
  6. #include <LibPDF/CommonNames.h>
  7. #include <LibPDF/Document.h>
  8. #include <LibPDF/Parser.h>
  9. #include <LibTextCodec/Decoder.h>
  10. namespace PDF {
  11. DeprecatedString OutlineItem::to_deprecated_string(int indent) const
  12. {
  13. auto indent_str = DeprecatedString::repeated(" "sv, indent + 1);
  14. StringBuilder child_builder;
  15. child_builder.append('[');
  16. for (auto& child : children)
  17. child_builder.appendff("{}\n", child->to_deprecated_string(indent + 1));
  18. child_builder.appendff("{}]", indent_str);
  19. StringBuilder builder;
  20. builder.append("OutlineItem {{\n"sv);
  21. builder.appendff("{}title={}\n", indent_str, title);
  22. builder.appendff("{}count={}\n", indent_str, count);
  23. builder.appendff("{}dest={}\n", indent_str, dest);
  24. builder.appendff("{}color={}\n", indent_str, color);
  25. builder.appendff("{}italic={}\n", indent_str, italic);
  26. builder.appendff("{}bold={}\n", indent_str, bold);
  27. builder.appendff("{}children={}\n", indent_str, child_builder.to_deprecated_string());
  28. builder.appendff("{}}}", DeprecatedString::repeated(" "sv, indent));
  29. return builder.to_deprecated_string();
  30. }
  31. PDFErrorOr<Optional<DeprecatedString>> InfoDict::title() const
  32. {
  33. return get_text(CommonNames::Title);
  34. }
  35. PDFErrorOr<Optional<DeprecatedString>> InfoDict::author() const
  36. {
  37. return get_text(CommonNames::Author);
  38. }
  39. PDFErrorOr<Optional<DeprecatedString>> InfoDict::subject() const
  40. {
  41. return get_text(CommonNames::Subject);
  42. }
  43. PDFErrorOr<Optional<DeprecatedString>> InfoDict::keywords() const
  44. {
  45. return get_text(CommonNames::Keywords);
  46. }
  47. PDFErrorOr<Optional<DeprecatedString>> InfoDict::creator() const
  48. {
  49. return get_text(CommonNames::Creator);
  50. }
  51. PDFErrorOr<Optional<DeprecatedString>> InfoDict::producer() const
  52. {
  53. return get_text(CommonNames::Producer);
  54. }
  55. PDFErrorOr<Optional<DeprecatedString>> InfoDict::creation_date() const
  56. {
  57. return get(CommonNames::CreationDate);
  58. }
  59. PDFErrorOr<Optional<DeprecatedString>> InfoDict::modification_date() const
  60. {
  61. return get(CommonNames::ModDate);
  62. }
  63. PDFErrorOr<Optional<DeprecatedString>> InfoDict::get_text(DeprecatedFlyString const& name) const
  64. {
  65. return TRY(get(name)).map(Document::text_string_to_utf8);
  66. }
  67. DeprecatedString Document::text_string_to_utf8(DeprecatedString const& text_string)
  68. {
  69. if (text_string.bytes().starts_with(Array<u8, 2> { 0xfe, 0xff })) {
  70. // The string is encoded in UTF16-BE
  71. return TextCodec::decoder_for("utf-16be"sv)->to_utf8(text_string).release_value_but_fixme_should_propagate_errors().to_deprecated_string();
  72. }
  73. if (text_string.bytes().starts_with(Array<u8, 3> { 239, 187, 191 })) {
  74. // The string is encoded in UTF-8.
  75. return text_string.substring(3);
  76. }
  77. // FIXME: Convert from PDFDocEncoding to UTF-8.
  78. return text_string;
  79. }
  80. PDFErrorOr<NonnullRefPtr<Document>> Document::create(ReadonlyBytes bytes)
  81. {
  82. auto parser = adopt_ref(*new DocumentParser({}, bytes));
  83. auto document = adopt_ref(*new Document(parser));
  84. document->m_version = TRY(parser->initialize());
  85. document->m_trailer = parser->trailer();
  86. document->m_catalog = TRY(parser->trailer()->get_dict(document, CommonNames::Root));
  87. if (document->m_trailer->contains(CommonNames::Encrypt)) {
  88. auto encryption_dict = TRY(document->m_trailer->get_dict(document, CommonNames::Encrypt));
  89. document->m_security_handler = TRY(SecurityHandler::create(document, encryption_dict));
  90. // Automatically attempt to unencrypt the document with the empty string. The
  91. // result is not important; it is the caller's responsibility to ensure the
  92. // document is unencrypted before calling initialize().
  93. document->m_security_handler->try_provide_user_password(""sv);
  94. }
  95. return document;
  96. }
  97. Document::Document(NonnullRefPtr<DocumentParser> const& parser)
  98. : m_parser(parser)
  99. {
  100. m_parser->set_document(this);
  101. }
  102. PDFErrorOr<void> Document::initialize()
  103. {
  104. if (m_security_handler)
  105. VERIFY(m_security_handler->has_user_password());
  106. TRY(build_page_tree());
  107. TRY(build_outline());
  108. return {};
  109. }
  110. PDFErrorOr<Value> Document::get_or_load_value(u32 index)
  111. {
  112. auto value = get_value(index);
  113. if (!value.has<Empty>()) // FIXME: Use Optional instead?
  114. return value;
  115. auto object = TRY(m_parser->parse_object_with_index(index));
  116. m_values.set(index, object);
  117. return object;
  118. }
  119. u32 Document::get_first_page_index() const
  120. {
  121. // FIXME: A PDF can have a different default first page, which
  122. // should be fetched and returned here
  123. return 0;
  124. }
  125. u32 Document::get_page_count() const
  126. {
  127. return m_page_object_indices.size();
  128. }
  129. static ErrorOr<void> collect_referenced_indices(Value const& value, Vector<int>& referenced_indices)
  130. {
  131. TRY(value.visit(
  132. [&](Empty const&) -> ErrorOr<void> { return {}; },
  133. [&](nullptr_t const&) -> ErrorOr<void> { return {}; },
  134. [&](bool const&) -> ErrorOr<void> { return {}; },
  135. [&](int const&) -> ErrorOr<void> { return {}; },
  136. [&](float const&) -> ErrorOr<void> { return {}; },
  137. [&](Reference const& ref) -> ErrorOr<void> {
  138. TRY(referenced_indices.try_append(ref.as_ref_index()));
  139. return {};
  140. },
  141. [&](NonnullRefPtr<Object> const& object) -> ErrorOr<void> {
  142. if (object->is<ArrayObject>()) {
  143. for (auto& element : object->cast<ArrayObject>()->elements())
  144. TRY(collect_referenced_indices(element, referenced_indices));
  145. } else if (object->is<DictObject>()) {
  146. for (auto& [key, value] : object->cast<DictObject>()->map()) {
  147. if (key != CommonNames::Parent)
  148. TRY(collect_referenced_indices(value, referenced_indices));
  149. }
  150. } else if (object->is<StreamObject>()) {
  151. for (auto& [key, value] : object->cast<StreamObject>()->dict()->map()) {
  152. if (key != CommonNames::Parent)
  153. TRY(collect_referenced_indices(value, referenced_indices));
  154. }
  155. }
  156. return {};
  157. }));
  158. return {};
  159. }
  160. static PDFErrorOr<void> dump_tree(Document& document, size_t index, HashTable<int>& seen)
  161. {
  162. if (seen.contains(index))
  163. return {};
  164. seen.set(index);
  165. auto const& value = TRY(document.get_or_load_value(index));
  166. outln("{} 0 obj", index);
  167. outln("{}", value.to_deprecated_string(0));
  168. outln("endobj");
  169. Vector<int> referenced_indices;
  170. TRY(collect_referenced_indices(value, referenced_indices));
  171. for (auto index : referenced_indices)
  172. TRY(dump_tree(document, index, seen));
  173. return {};
  174. }
  175. PDFErrorOr<void> Document::dump_page(u32 index)
  176. {
  177. VERIFY(index < m_page_object_indices.size());
  178. auto page_object_index = m_page_object_indices[index];
  179. HashTable<int> seen;
  180. TRY(dump_tree(*this, page_object_index, seen));
  181. return {};
  182. }
  183. PDFErrorOr<Page> Document::get_page(u32 index)
  184. {
  185. VERIFY(index < m_page_object_indices.size());
  186. auto cached_page = m_pages.get(index);
  187. if (cached_page.has_value())
  188. return cached_page.value();
  189. auto page_object_index = m_page_object_indices[index];
  190. auto page_object = TRY(get_or_load_value(page_object_index));
  191. auto raw_page_object = TRY(resolve_to<DictObject>(page_object));
  192. RefPtr<DictObject> resources;
  193. auto maybe_resources_object = TRY(get_inheritable_object(CommonNames::Resources, raw_page_object));
  194. if (maybe_resources_object.has_value())
  195. resources = maybe_resources_object.value()->cast<DictObject>();
  196. else
  197. resources = adopt_ref(*new DictObject({}));
  198. RefPtr<Object> contents;
  199. if (raw_page_object->contains(CommonNames::Contents))
  200. contents = TRY(raw_page_object->get_object(this, CommonNames::Contents));
  201. Rectangle media_box;
  202. auto maybe_media_box_object = TRY(get_inheritable_object(CommonNames::MediaBox, raw_page_object));
  203. if (maybe_media_box_object.has_value()) {
  204. auto media_box_array = maybe_media_box_object.value()->cast<ArrayObject>();
  205. media_box = Rectangle {
  206. media_box_array->at(0).to_float(),
  207. media_box_array->at(1).to_float(),
  208. media_box_array->at(2).to_float(),
  209. media_box_array->at(3).to_float(),
  210. };
  211. } else {
  212. // As most other libraries seem to do, we default to the standard
  213. // US letter size of 8.5" x 11" (612 x 792 Postscript units).
  214. media_box = Rectangle {
  215. 0, 0,
  216. 612, 792
  217. };
  218. }
  219. Rectangle crop_box;
  220. auto maybe_crop_box_object = TRY(get_inheritable_object(CommonNames::CropBox, raw_page_object));
  221. if (maybe_crop_box_object.has_value()) {
  222. auto crop_box_array = maybe_crop_box_object.value()->cast<ArrayObject>();
  223. crop_box = Rectangle {
  224. crop_box_array->at(0).to_float(),
  225. crop_box_array->at(1).to_float(),
  226. crop_box_array->at(2).to_float(),
  227. crop_box_array->at(3).to_float(),
  228. };
  229. } else {
  230. crop_box = media_box;
  231. }
  232. float user_unit = 1.0f;
  233. if (raw_page_object->contains(CommonNames::UserUnit))
  234. user_unit = raw_page_object->get_value(CommonNames::UserUnit).to_float();
  235. int rotate = 0;
  236. auto maybe_rotate = TRY(get_inheritable_value(CommonNames::Rotate, raw_page_object));
  237. if (maybe_rotate.has_value()) {
  238. rotate = TRY(resolve_to<int>(maybe_rotate.value()));
  239. VERIFY(rotate % 90 == 0);
  240. }
  241. Page page { resources.release_nonnull(), move(contents), media_box, crop_box, user_unit, rotate };
  242. m_pages.set(index, page);
  243. return page;
  244. }
  245. PDFErrorOr<Value> Document::resolve(Value const& value)
  246. {
  247. if (value.has<Reference>()) {
  248. // FIXME: Surely indirect PDF objects can't contain another indirect PDF object,
  249. // right? Unsure from the spec, but if they can, these return values would have
  250. // to be wrapped with another resolve() call.
  251. return get_or_load_value(value.as_ref_index());
  252. }
  253. if (!value.has<NonnullRefPtr<Object>>())
  254. return value;
  255. auto& obj = value.get<NonnullRefPtr<Object>>();
  256. if (obj->is<IndirectValue>())
  257. return static_ptr_cast<IndirectValue>(obj)->value();
  258. return value;
  259. }
  260. PDFErrorOr<Optional<InfoDict>> Document::info_dict()
  261. {
  262. if (!trailer()->contains(CommonNames::Info))
  263. return OptionalNone {};
  264. return InfoDict(this, TRY(trailer()->get_dict(this, CommonNames::Info)));
  265. }
  266. PDFErrorOr<Vector<DeprecatedFlyString>> Document::read_filters(NonnullRefPtr<DictObject> dict)
  267. {
  268. Vector<DeprecatedFlyString> filters;
  269. // We may either get a single filter or an array of cascading filters
  270. auto filter_object = TRY(dict->get_object(this, CommonNames::Filter));
  271. if (filter_object->is<ArrayObject>()) {
  272. auto filter_array = filter_object->cast<ArrayObject>();
  273. for (size_t i = 0; i < filter_array->size(); ++i)
  274. filters.append(TRY(filter_array->get_name_at(this, i))->name());
  275. } else {
  276. filters.append(filter_object->cast<NameObject>()->name());
  277. }
  278. return filters;
  279. }
  280. PDFErrorOr<void> Document::build_page_tree()
  281. {
  282. auto page_tree = TRY(m_catalog->get_dict(this, CommonNames::Pages));
  283. return add_page_tree_node_to_page_tree(page_tree);
  284. }
  285. PDFErrorOr<void> Document::add_page_tree_node_to_page_tree(NonnullRefPtr<DictObject> const& page_tree)
  286. {
  287. auto kids_array = TRY(page_tree->get_array(this, CommonNames::Kids));
  288. auto page_count = page_tree->get(CommonNames::Count).value().get<int>();
  289. if (static_cast<size_t>(page_count) != kids_array->elements().size()) {
  290. // This page tree contains child page trees, so we recursively add
  291. // these pages to the overall page tree
  292. for (auto& value : *kids_array) {
  293. auto reference_index = value.as_ref_index();
  294. auto maybe_page_tree_node = TRY(m_parser->conditionally_parse_page_tree_node(reference_index));
  295. if (maybe_page_tree_node) {
  296. TRY(add_page_tree_node_to_page_tree(maybe_page_tree_node.release_nonnull()));
  297. } else {
  298. m_page_object_indices.append(reference_index);
  299. }
  300. }
  301. } else {
  302. // We know all of the kids are leaf nodes
  303. for (auto& value : *kids_array)
  304. m_page_object_indices.append(value.as_ref_index());
  305. }
  306. return {};
  307. }
  308. PDFErrorOr<NonnullRefPtr<Object>> Document::find_in_name_tree(NonnullRefPtr<DictObject> tree, DeprecatedFlyString name)
  309. {
  310. if (tree->contains(CommonNames::Kids)) {
  311. return find_in_name_tree_nodes(tree->get_array(CommonNames::Kids), name);
  312. }
  313. if (!tree->contains(CommonNames::Names))
  314. return Error { Error::Type::MalformedPDF, "name tree has neither Kids nor Names" };
  315. auto key_value_names_array = TRY(tree->get_array(this, CommonNames::Names));
  316. return find_in_key_value_array(key_value_names_array, name);
  317. }
  318. PDFErrorOr<NonnullRefPtr<Object>> Document::find_in_name_tree_nodes(NonnullRefPtr<ArrayObject> siblings, DeprecatedFlyString name)
  319. {
  320. for (size_t i = 0; i < siblings->size(); i++) {
  321. auto sibling = TRY(resolve_to<DictObject>(siblings->at(i)));
  322. auto limits = sibling->get_array(CommonNames::Limits);
  323. if (limits->size() != 2)
  324. return Error { Error::Type::MalformedPDF, "Expected 2-element Limits array" };
  325. auto start = limits->get_string_at(0);
  326. auto end = limits->get_string_at(1);
  327. if (start->string() <= name && end->string() >= name) {
  328. return find_in_name_tree(sibling, name);
  329. }
  330. }
  331. return Error { Error::Type::MalformedPDF, DeprecatedString::formatted("Didn't find node in name tree containing name {}", name) };
  332. }
  333. PDFErrorOr<NonnullRefPtr<Object>> Document::find_in_key_value_array(NonnullRefPtr<ArrayObject> key_value_array, DeprecatedFlyString name)
  334. {
  335. if (key_value_array->size() % 2 == 1)
  336. return Error { Error::Type::MalformedPDF, "key/value array has dangling key" };
  337. for (size_t i = 0; i < key_value_array->size() / 2; i++) {
  338. auto key = key_value_array->get_string_at(2 * i);
  339. if (key->string() == name) {
  340. return key_value_array->get_object_at(this, 2 * i + 1);
  341. }
  342. }
  343. return Error { Error::Type::MalformedPDF, DeprecatedString::formatted("Didn't find expected name {} in key/value array", name) };
  344. }
  345. PDFErrorOr<void> Document::build_outline()
  346. {
  347. if (!m_catalog->contains(CommonNames::Outlines))
  348. return {};
  349. auto outline_dict = TRY(m_catalog->get_dict(this, CommonNames::Outlines));
  350. if (!outline_dict->contains(CommonNames::First))
  351. return {};
  352. if (!outline_dict->contains(CommonNames::Last))
  353. return {};
  354. HashMap<u32, u32> page_number_by_index_ref;
  355. for (u32 page_number = 0; page_number < m_page_object_indices.size(); ++page_number) {
  356. page_number_by_index_ref.set(m_page_object_indices[page_number], page_number);
  357. }
  358. auto first_ref = outline_dict->get_value(CommonNames::First);
  359. auto children = TRY(build_outline_item_chain(first_ref, page_number_by_index_ref));
  360. m_outline = adopt_ref(*new OutlineDict());
  361. m_outline->children = move(children);
  362. if (outline_dict->contains(CommonNames::Count))
  363. m_outline->count = outline_dict->get_value(CommonNames::Count).get<int>();
  364. return {};
  365. }
  366. PDFErrorOr<Destination> Document::create_destination_from_parameters(NonnullRefPtr<ArrayObject> array, HashMap<u32, u32> const& page_number_by_index_ref)
  367. {
  368. auto page_ref = array->at(0);
  369. if (page_ref.has<nullptr_t>())
  370. return Destination { Destination::Type::XYZ, {}, {} };
  371. auto type_name = TRY(array->get_name_at(this, 1))->name();
  372. Vector<Optional<float>> parameters;
  373. TRY(parameters.try_ensure_capacity(array->size() - 2));
  374. for (size_t i = 2; i < array->size(); i++) {
  375. auto& param = array->at(i);
  376. if (param.has<nullptr_t>())
  377. parameters.unchecked_append({});
  378. else
  379. parameters.append(param.to_float());
  380. }
  381. Destination::Type type;
  382. if (type_name == CommonNames::XYZ) {
  383. type = Destination::Type::XYZ;
  384. } else if (type_name == CommonNames::Fit) {
  385. type = Destination::Type::Fit;
  386. } else if (type_name == CommonNames::FitH) {
  387. type = Destination::Type::FitH;
  388. } else if (type_name == CommonNames::FitV) {
  389. type = Destination::Type::FitV;
  390. } else if (type_name == CommonNames::FitR) {
  391. type = Destination::Type::FitR;
  392. } else if (type_name == CommonNames::FitB) {
  393. type = Destination::Type::FitB;
  394. } else if (type_name == CommonNames::FitBH) {
  395. type = Destination::Type::FitBH;
  396. } else if (type_name == CommonNames::FitBV) {
  397. type = Destination::Type::FitBV;
  398. } else {
  399. VERIFY_NOT_REACHED();
  400. }
  401. // The spec requires page_ref to be an indirect reference to a page object,
  402. // but in practice it's sometimes a page index.
  403. Optional<u32> page_number;
  404. if (page_ref.has<int>())
  405. page_number = page_ref.get<int>();
  406. else
  407. page_number = page_number_by_index_ref.get(page_ref.as_ref_index());
  408. return Destination { type, page_number, parameters };
  409. }
  410. PDFErrorOr<Optional<NonnullRefPtr<Object>>> Document::get_inheritable_object(DeprecatedFlyString const& name, NonnullRefPtr<DictObject> object)
  411. {
  412. if (!object->contains(name)) {
  413. if (!object->contains(CommonNames::Parent))
  414. return { OptionalNone() };
  415. auto parent = TRY(object->get_dict(this, CommonNames::Parent));
  416. return get_inheritable_object(name, parent);
  417. }
  418. return TRY(object->get_object(this, name));
  419. }
  420. PDFErrorOr<Optional<Value>> Document::get_inheritable_value(DeprecatedFlyString const& name, NonnullRefPtr<DictObject> object)
  421. {
  422. if (!object->contains(name)) {
  423. if (!object->contains(CommonNames::Parent))
  424. return { OptionalNone() };
  425. auto parent = TRY(object->get_dict(this, CommonNames::Parent));
  426. return get_inheritable_value(name, parent);
  427. }
  428. return object->get(name);
  429. }
  430. PDFErrorOr<Destination> Document::create_destination_from_dictionary_entry(NonnullRefPtr<Object> const& entry, HashMap<u32, u32> const& page_number_by_index_ref)
  431. {
  432. if (entry->is<ArrayObject>()) {
  433. auto entry_array = entry->cast<ArrayObject>();
  434. return create_destination_from_parameters(entry_array, page_number_by_index_ref);
  435. }
  436. auto entry_dictionary = entry->cast<DictObject>();
  437. auto d_array = MUST(entry_dictionary->get_array(this, CommonNames::D));
  438. return create_destination_from_parameters(d_array, page_number_by_index_ref);
  439. }
  440. PDFErrorOr<Destination> Document::create_destination_from_object(NonnullRefPtr<Object> const& dest_obj, HashMap<u32, u32> const& page_number_by_index_ref)
  441. {
  442. // PDF 1.7 spec, "8.2.1 Destinations"
  443. if (dest_obj->is<ArrayObject>()) {
  444. auto dest_arr = dest_obj->cast<ArrayObject>();
  445. return TRY(create_destination_from_parameters(dest_arr, page_number_by_index_ref));
  446. }
  447. if (dest_obj->is<NameObject>() || dest_obj->is<StringObject>()) {
  448. DeprecatedFlyString dest_name;
  449. if (dest_obj->is<NameObject>())
  450. dest_name = dest_obj->cast<NameObject>()->name();
  451. else
  452. dest_name = dest_obj->cast<StringObject>()->string();
  453. if (auto dests_value = m_catalog->get(CommonNames::Dests); dests_value.has_value()) {
  454. auto dests = TRY(resolve_to<DictObject>(dests_value.value()));
  455. auto entry = MUST(dests->get_object(this, dest_name));
  456. return TRY(create_destination_from_dictionary_entry(entry, page_number_by_index_ref));
  457. }
  458. if (auto names_value = m_catalog->get(CommonNames::Names); names_value.has_value()) {
  459. auto names = TRY(resolve_to<DictObject>(names_value.release_value()));
  460. if (!names->contains(CommonNames::Dests))
  461. return Error { Error::Type::MalformedPDF, "Missing Dests key in document catalogue's Names dictionary" };
  462. auto dest_obj = TRY(find_in_name_tree(TRY(names->get_dict(this, CommonNames::Dests)), dest_name));
  463. return TRY(create_destination_from_dictionary_entry(dest_obj, page_number_by_index_ref));
  464. }
  465. }
  466. return Error { Error::Type::MalformedPDF, "Malformed outline destination" };
  467. }
  468. PDFErrorOr<NonnullRefPtr<OutlineItem>> Document::build_outline_item(NonnullRefPtr<DictObject> const& outline_item_dict, HashMap<u32, u32> const& page_number_by_index_ref)
  469. {
  470. auto outline_item = adopt_ref(*new OutlineItem {});
  471. if (outline_item_dict->contains(CommonNames::First)) {
  472. VERIFY(outline_item_dict->contains(CommonNames::Last));
  473. auto first_ref = outline_item_dict->get_value(CommonNames::First);
  474. auto children = TRY(build_outline_item_chain(first_ref, page_number_by_index_ref));
  475. for (auto& child : children) {
  476. child->parent = outline_item;
  477. }
  478. outline_item->children = move(children);
  479. }
  480. outline_item->title = text_string_to_utf8(TRY(outline_item_dict->get_string(this, CommonNames::Title))->string());
  481. if (outline_item_dict->contains(CommonNames::Count))
  482. outline_item->count = outline_item_dict->get_value(CommonNames::Count).get<int>();
  483. if (outline_item_dict->contains(CommonNames::Dest)) {
  484. auto dest_obj = TRY(outline_item_dict->get_object(this, CommonNames::Dest));
  485. outline_item->dest = TRY(create_destination_from_object(dest_obj, page_number_by_index_ref));
  486. } else if (outline_item_dict->contains(CommonNames::A)) {
  487. // PDF 1.7 spec, "8.5 Actions"
  488. auto action_dict = TRY(outline_item_dict->get_dict(this, CommonNames::A));
  489. if (action_dict->contains(CommonNames::S)) {
  490. // PDF 1.7 spec, "TABLE 8.48 Action types"
  491. auto action_type = TRY(action_dict->get_name(this, CommonNames::S))->name();
  492. if (action_type == "GoTo") {
  493. // PDF 1.7 spec, "Go-To Actions"
  494. if (action_dict->contains(CommonNames::D)) {
  495. auto dest_obj = TRY(action_dict->get_object(this, CommonNames::D));
  496. outline_item->dest = TRY(create_destination_from_object(dest_obj, page_number_by_index_ref));
  497. }
  498. } else {
  499. dbgln("Unhandled action type {}", action_type);
  500. }
  501. }
  502. }
  503. if (outline_item_dict->contains(CommonNames::C)) {
  504. auto color_array = TRY(outline_item_dict->get_array(this, CommonNames::C));
  505. auto r = static_cast<int>(255.0f * color_array->at(0).to_float());
  506. auto g = static_cast<int>(255.0f * color_array->at(1).to_float());
  507. auto b = static_cast<int>(255.0f * color_array->at(2).to_float());
  508. outline_item->color = Color(r, g, b);
  509. }
  510. if (outline_item_dict->contains(CommonNames::F)) {
  511. auto bitfield = outline_item_dict->get_value(CommonNames::F).get<int>();
  512. outline_item->italic = bitfield & 0x1;
  513. outline_item->bold = bitfield & 0x2;
  514. }
  515. return outline_item;
  516. }
  517. PDFErrorOr<Vector<NonnullRefPtr<OutlineItem>>> Document::build_outline_item_chain(Value const& first_ref, HashMap<u32, u32> const& page_number_by_index_ref)
  518. {
  519. // We used to receive a last_ref parameter, which was what the parent of this chain
  520. // thought was this chain's last child. There are documents out there in the wild
  521. // where this cross-references don't match though, and it seems like simply following
  522. // the /First and /Next links is the way to go to construct the whole Outline
  523. // (we already ignore the /Parent attribute too, which can also be out of sync).
  524. VERIFY(first_ref.has<Reference>());
  525. Vector<NonnullRefPtr<OutlineItem>> children;
  526. auto first_value = TRY(get_or_load_value(first_ref.as_ref_index())).get<NonnullRefPtr<Object>>();
  527. auto first_dict = first_value->cast<DictObject>();
  528. auto first = TRY(build_outline_item(first_dict, page_number_by_index_ref));
  529. children.append(first);
  530. auto current_child_dict = first_dict;
  531. u32 current_child_index = first_ref.as_ref_index();
  532. while (current_child_dict->contains(CommonNames::Next)) {
  533. auto next_child_dict_ref = current_child_dict->get_value(CommonNames::Next);
  534. current_child_index = next_child_dict_ref.as_ref_index();
  535. auto next_child_value = TRY(get_or_load_value(current_child_index)).get<NonnullRefPtr<Object>>();
  536. auto next_child_dict = next_child_value->cast<DictObject>();
  537. auto next_child = TRY(build_outline_item(next_child_dict, page_number_by_index_ref));
  538. children.append(next_child);
  539. current_child_dict = move(next_child_dict);
  540. }
  541. return children;
  542. }
  543. }