diff options
author | Artem Strygin <art-snake@yandex-team.ru> | 2017-08-23 15:53:20 +0300 |
---|---|---|
committer | Chromium commit bot <commit-bot@chromium.org> | 2017-08-23 13:08:19 +0000 |
commit | 81692ab7daf5389bae1bed55c4a3a8e9e68fada4 (patch) | |
tree | 5f48454525fed11d91a748c16d66bbc8dbe3dc28 /core/fpdfapi/parser | |
parent | 77d8317f2e6bb9c75d96390fdf843b86694ba6d2 (diff) | |
download | pdfium-81692ab7daf5389bae1bed55c4a3a8e9e68fada4.tar.xz |
Simplify Trailers usage in CPDF_Parser.
Do not store non main trailers in memory.
Improve readability.
Change-Id: I2eab5d31a5be056871e7e5953e4b38662b91f5ae
Reviewed-on: https://pdfium-review.googlesource.com/10750
Commit-Queue: Art Snake <art-snake@yandex-team.ru>
Reviewed-by: dsinclair <dsinclair@chromium.org>
Diffstat (limited to 'core/fpdfapi/parser')
-rw-r--r-- | core/fpdfapi/parser/cpdf_parser.cpp | 116 | ||||
-rw-r--r-- | core/fpdfapi/parser/cpdf_parser.h | 15 |
2 files changed, 78 insertions, 53 deletions
diff --git a/core/fpdfapi/parser/cpdf_parser.cpp b/core/fpdfapi/parser/cpdf_parser.cpp index 25d2ac28cf..3184d88943 100644 --- a/core/fpdfapi/parser/cpdf_parser.cpp +++ b/core/fpdfapi/parser/cpdf_parser.cpp @@ -67,14 +67,56 @@ CPDF_Parser::ObjectType GetObjectTypeFromCrossRefStreamType( } // namespace +class CPDF_Parser::TrailerData { + public: + TrailerData() {} + ~TrailerData() {} + + CPDF_Dictionary* GetMainTrailer() const { return main_trailer_.get(); } + + void SetMainTrailer(std::unique_ptr<CPDF_Dictionary> trailer) { + ASSERT(trailer); + main_trailer_ = std::move(trailer); + ApplyTrailer(main_trailer_.get()); + } + + void AppendTrailer(std::unique_ptr<CPDF_Dictionary> trailer) { + ASSERT(trailer); + ApplyTrailer(trailer.get()); + } + + void Clear() { + main_trailer_.reset(); + last_info_obj_num_ = 0; + } + + uint32_t GetInfoObjNum() const { + const CPDF_Reference* pRef = ToReference( + GetMainTrailer() ? GetMainTrailer()->GetObjectFor("Info") : nullptr); + return pRef ? pRef->GetRefObjNum() : last_info_obj_num_; + } + + private: + void ApplyTrailer(const CPDF_Dictionary* dict) { + // The most recent Info object number contained in last added trailer. + // See PDF 1.7 spec, section 3.4.5 - Incremental Updates. + const auto* pRef = ToReference(dict->GetObjectFor("Info")); + if (pRef) + last_info_obj_num_ = pRef->GetRefObjNum(); + } + + std::unique_ptr<CPDF_Dictionary> main_trailer_; + uint32_t last_info_obj_num_ = 0; +}; + CPDF_Parser::CPDF_Parser() : m_pSyntax(pdfium::MakeUnique<CPDF_SyntaxParser>()), m_bHasParsed(false), m_bXRefStream(false), m_FileVersion(0), m_pEncryptDict(nullptr), - m_TrailerPos(CPDF_Parser::kInvalidPos), - m_dwXrefStartObjNum(0) {} + m_TrailerData(pdfium::MakeUnique<TrailerData>()), + m_linearized_first_page_cross_ref_start_obj_num(0) {} CPDF_Parser::~CPDF_Parser() { ReleaseEncryptHandler(); @@ -344,8 +386,7 @@ bool CPDF_Parser::LoadAllCrossRefV4(FX_FILESIZE xrefpos) { if (!trailer) return false; - m_Trailers.push_back(std::move(trailer)); - m_TrailerPos = m_Trailers.size() - 1; + m_TrailerData->SetMainTrailer(std::move(trailer)); int32_t xrefsize = GetDirectInteger(GetTrailer(), "Size"); if (xrefsize > 0 && xrefsize <= kMaxXRefSize) ShrinkObjectMap(xrefsize); @@ -381,7 +422,7 @@ bool CPDF_Parser::LoadAllCrossRefV4(FX_FILESIZE xrefpos) { // SLOW ... XRefStreamList.insert(XRefStreamList.begin(), pDict->GetIntegerFor("XRefStm")); - m_Trailers.push_back(std::move(pDict)); + m_TrailerData->AppendTrailer(std::move(pDict)); } for (size_t i = 0; i < CrossRefList.size(); ++i) { @@ -402,8 +443,7 @@ bool CPDF_Parser::LoadLinearizedAllCrossRefV4(FX_FILESIZE xrefpos, if (!trailer) return false; - m_Trailers.push_back(std::move(trailer)); - m_TrailerPos = m_Trailers.size() - 1; + m_TrailerData->SetMainTrailer(std::move(trailer)); int32_t xrefsize = GetDirectInteger(GetTrailer(), "Size"); if (xrefsize == 0) return false; @@ -437,7 +477,7 @@ bool CPDF_Parser::LoadLinearizedAllCrossRefV4(FX_FILESIZE xrefpos, // SLOW ... XRefStreamList.insert(XRefStreamList.begin(), pDict->GetIntegerFor("XRefStm")); - m_Trailers.push_back(std::move(pDict)); + m_TrailerData->AppendTrailer(std::move(pDict)); } for (size_t i = 1; i < CrossRefList.size(); ++i) { @@ -543,8 +583,7 @@ bool CPDF_Parser::ParseAndAppendCrossRefSubsectionData( return true; } -bool CPDF_Parser::ParseCrossRefV4(std::vector<CrossRefObjData>* out_objects, - uint32_t* start_obj_num_at_last_block) { +bool CPDF_Parser::ParseCrossRefV4(std::vector<CrossRefObjData>* out_objects) { if (out_objects) out_objects->clear(); @@ -567,8 +606,6 @@ bool CPDF_Parser::ParseCrossRefV4(std::vector<CrossRefObjData>* out_objects, uint32_t start_objnum = FXSYS_atoui(word.c_str()); if (start_objnum >= kMaxObjectNumber) return false; - if (start_obj_num_at_last_block) - *start_obj_num_at_last_block = start_objnum; uint32_t count = m_pSyntax->GetDirectNum(); m_pSyntax->ToNextWord(); @@ -589,7 +626,7 @@ bool CPDF_Parser::LoadCrossRefV4(FX_FILESIZE pos, bool bSkip) { m_pSyntax->SetPos(pos); std::vector<CrossRefObjData> objects; - if (!ParseCrossRefV4(bSkip ? nullptr : &objects, &m_dwXrefStartObjNum)) + if (!ParseCrossRefV4(bSkip ? nullptr : &objects)) return false; MergeCrossRefObjectsData(objects); @@ -625,8 +662,7 @@ bool CPDF_Parser::LoadAllCrossRefV5(FX_FILESIZE xrefpos) { bool CPDF_Parser::RebuildCrossRef() { m_ObjectInfo.clear(); - m_Trailers.clear(); - m_TrailerPos = CPDF_Parser::kInvalidPos; + m_TrailerData->Clear(); ParserState state = ParserState::kDefault; int32_t inside_index = 0; @@ -786,8 +822,8 @@ bool CPDF_Parser::RebuildCrossRef() { CPDF_Object* pRoot = pDict->GetObjectFor("Root"); if (pRoot && pRoot->GetDict() && pRoot->GetDict()->GetObjectFor("Pages")) { - m_Trailers.push_back(ToDictionary(pDict->Clone())); - m_TrailerPos = m_Trailers.size() - 1; + m_TrailerData->SetMainTrailer( + ToDictionary(pDict->Clone())); } } } @@ -862,11 +898,9 @@ bool CPDF_Parser::RebuildCrossRef() { } } } else { - if (pObj->IsStream()) - m_Trailers.push_back(ToDictionary(pTrailer->Clone())); - else - m_Trailers.push_back(ToDictionary(std::move(pObj))); - m_TrailerPos = m_Trailers.size() - 1; + m_TrailerData->SetMainTrailer( + ToDictionary(pObj->IsStream() ? pTrailer->Clone() + : std::move(pObj))); FX_FILESIZE dwSavePos = m_pSyntax->GetPos(); CFX_ByteString strWord = m_pSyntax->GetKeyword(); @@ -1005,13 +1039,12 @@ bool CPDF_Parser::LoadCrossRefV5(FX_FILESIZE* pos, bool bMainXRef) { std::unique_ptr<CPDF_Dictionary> pNewTrailer = ToDictionary(pDict->Clone()); if (bMainXRef) { - m_Trailers.push_back(std::move(pNewTrailer)); - m_TrailerPos = m_Trailers.size() - 1; + m_TrailerData->SetMainTrailer(std::move(pNewTrailer)); ShrinkObjectMap(size); for (auto& it : m_ObjectInfo) it.second.type = ObjectType::kFree; } else { - m_Trailers.push_back(std::move(pNewTrailer)); + m_TrailerData->AppendTrailer(std::move(pNewTrailer)); } std::vector<std::pair<int32_t, int32_t>> arrIndex; @@ -1059,7 +1092,6 @@ bool CPDF_Parser::LoadCrossRefV5(FX_FILESIZE* pos, bool bMainXRef) { if (startnum < 0) continue; - m_dwXrefStartObjNum = pdfium::base::checked_cast<uint32_t>(startnum); uint32_t count = pdfium::base::checked_cast<uint32_t>(arrIndex[i].second); FX_SAFE_UINT32 dwCaculatedSize = segindex; dwCaculatedSize += count; @@ -1145,21 +1177,12 @@ uint32_t CPDF_Parser::GetRootObjNum() { return pRef ? pRef->GetRefObjNum() : 0; } +CPDF_Dictionary* CPDF_Parser::GetTrailer() const { + return m_TrailerData->GetMainTrailer(); +} + uint32_t CPDF_Parser::GetInfoObjNum() { - CPDF_Reference* pRef = - ToReference(GetTrailer() ? GetTrailer()->GetObjectFor("Info") : nullptr); - if (pRef) - return pRef->GetRefObjNum(); - - // Search trailers array from latest to earliest revision, as we want the - // most recent Info object number. - // See PDF 1.7 spec, section 3.4.5 - Incremental Updates. - for (auto it = m_Trailers.rbegin(); it != m_Trailers.rend(); ++it) { - pRef = ToReference(it->get()->GetObjectFor("Info")); - if (pRef) - return pRef->GetRefObjNum(); - } - return 0; + return m_TrailerData->GetInfoObjNum(); } std::unique_ptr<CPDF_Object> CPDF_Parser::ParseIndirectObject( @@ -1378,14 +1401,14 @@ CPDF_Parser::Error CPDF_Parser::StartLinearizedParse( bXRefRebuilt = true; m_LastXRefOffset = 0; } - + m_linearized_first_page_cross_ref_start_obj_num = + m_ObjectInfo.empty() ? 0 : m_ObjectInfo.begin()->first; if (bLoadV4) { std::unique_ptr<CPDF_Dictionary> trailer = LoadTrailerV4(); if (!trailer) return SUCCESS; - m_Trailers.push_back(std::move(trailer)); - m_TrailerPos = m_Trailers.size() - 1; + m_TrailerData->SetMainTrailer(std::move(trailer)); int32_t xrefsize = GetDirectInteger(GetTrailer(), "Size"); if (xrefsize > 0) ShrinkObjectMap(xrefsize); @@ -1470,7 +1493,12 @@ CPDF_Parser::Error CPDF_Parser::LoadLinearizedMainXRefTable() { m_ObjectStreamMap.clear(); m_ObjCache.clear(); - if (!LoadLinearizedAllCrossRefV4(m_LastXRefOffset, m_dwXrefStartObjNum) && + // In linearized document, the main cross ref always should start from 0 + // objnum. + // And should have count equals to first obj number of first page cross ref + // table. + if (!LoadLinearizedAllCrossRefV4( + m_LastXRefOffset, m_linearized_first_page_cross_ref_start_obj_num) && !LoadLinearizedAllCrossRefV5(m_LastXRefOffset)) { m_LastXRefOffset = 0; m_pSyntax->m_MetadataObjnum = dwSaveMetadataObjnum; diff --git a/core/fpdfapi/parser/cpdf_parser.h b/core/fpdfapi/parser/cpdf_parser.h index f636704162..437f214db6 100644 --- a/core/fpdfapi/parser/cpdf_parser.h +++ b/core/fpdfapi/parser/cpdf_parser.h @@ -61,10 +61,7 @@ class CPDF_Parser { void SetPassword(const char* password) { m_Password = password; } CFX_ByteString GetPassword() { return m_Password; } - CPDF_Dictionary* GetTrailer() const { - return m_TrailerPos == kInvalidPos ? nullptr - : m_Trailers[m_TrailerPos].get(); - } + CPDF_Dictionary* GetTrailer() const; FX_FILESIZE GetLastXRefOffset() const { return m_LastXRefOffset; } uint32_t GetPermissions() const; @@ -128,6 +125,8 @@ class CPDF_Parser { private: friend class CPDF_DataAvail; + class TrailerData; + enum class ParserState { kDefault, kComment, @@ -176,8 +175,7 @@ class CPDF_Parser { uint32_t start_objnum, uint32_t count, std::vector<CrossRefObjData>* out_objects); - bool ParseCrossRefV4(std::vector<CrossRefObjData>* out_objects, - uint32_t* start_obj_num_at_last_block); + bool ParseCrossRefV4(std::vector<CrossRefObjData>* out_objects); void MergeCrossRefObjectsData(const std::vector<CrossRefObjData>& objects); std::unique_ptr<CPDF_Object> ParseIndirectObjectAtInternal( @@ -199,10 +197,9 @@ class CPDF_Parser { FX_FILESIZE m_LastXRefOffset; std::unique_ptr<CPDF_SecurityHandler> m_pSecurityHandler; CFX_ByteString m_Password; - std::vector<std::unique_ptr<CPDF_Dictionary>> m_Trailers; - size_t m_TrailerPos; + std::unique_ptr<TrailerData> m_TrailerData; std::unique_ptr<CPDF_LinearizedHeader> m_pLinearized; - uint32_t m_dwXrefStartObjNum; + uint32_t m_linearized_first_page_cross_ref_start_obj_num; // A map of object numbers to indirect streams. std::map<uint32_t, CFX_RetainPtr<CPDF_StreamAcc>> m_ObjectStreamMap; |