Remove CPDF_Parser::GetSyntax()
Move some code into CPDF_Parser so that it need not expose an unowned
reference to its syntax parser. This code probably belonged in
CPDF_Parser in the first place, since it is mutating an object
owned by the parser.
Change-Id: I7354059a537660fdd7ad57ef5452f46d83153d4b
Reviewed-on: https://pdfium-review.googlesource.com/c/pdfium/+/99311
Reviewed-by: Lei Zhang <thestig@chromium.org>
Commit-Queue: Tom Sepez <tsepez@chromium.org>
diff --git a/core/fpdfapi/edit/cpdf_creator.cpp b/core/fpdfapi/edit/cpdf_creator.cpp
index ab1dfeb..5bd35fe 100644
--- a/core/fpdfapi/edit/cpdf_creator.cpp
+++ b/core/fpdfapi/edit/cpdf_creator.cpp
@@ -21,7 +21,6 @@
#include "core/fpdfapi/parser/cpdf_parser.h"
#include "core/fpdfapi/parser/cpdf_security_handler.h"
#include "core/fpdfapi/parser/cpdf_string.h"
-#include "core/fpdfapi/parser/cpdf_syntax_parser.h"
#include "core/fpdfapi/parser/fpdf_parser_utility.h"
#include "core/fxcrt/data_vector.h"
#include "core/fxcrt/fx_extension.h"
@@ -235,26 +234,14 @@
}
m_iStage = Stage::kInitWriteObjs20;
} else {
- m_SavedOffset = m_pParser->GetSyntax()->GetDocumentSize();
+ m_SavedOffset = m_pParser->GetDocumentSize();
m_iStage = Stage::kWriteIncremental15;
}
}
if (m_iStage == Stage::kWriteIncremental15) {
if (m_IsOriginal && m_SavedOffset > 0) {
- static constexpr FX_FILESIZE kBufferSize = 4096;
- DataVector<uint8_t> buffer(kBufferSize);
- FX_FILESIZE src_size = m_SavedOffset;
- m_pParser->GetSyntax()->SetPos(0);
- while (src_size) {
- const uint32_t block_size =
- static_cast<uint32_t>(std::min(kBufferSize, src_size));
- if (!m_pParser->GetSyntax()->ReadBlock(buffer.data(), block_size))
- return Stage::kInvalid;
- if (!m_Archive->WriteBlock(buffer.data(), block_size))
- return Stage::kInvalid;
-
- src_size -= block_size;
- }
+ if (!m_pParser->WriteToArchive(m_Archive.get(), m_SavedOffset))
+ return Stage::kInvalid;
}
if (m_IsOriginal && m_pParser->GetLastXRefOffset() == 0) {
for (uint32_t num = 0; num <= m_pParser->GetLastObjNum(); ++num) {
diff --git a/core/fpdfapi/edit/cpdf_creator.h b/core/fpdfapi/edit/cpdf_creator.h
index 60c4b6b..5cef585 100644
--- a/core/fpdfapi/edit/cpdf_creator.h
+++ b/core/fpdfapi/edit/cpdf_creator.h
@@ -73,7 +73,7 @@
CPDF_CryptoHandler* GetCryptoHandler();
UnownedPtr<CPDF_Document> const m_pDocument;
- UnownedPtr<const CPDF_Parser> const m_pParser;
+ UnownedPtr<CPDF_Parser> const m_pParser;
RetainPtr<const CPDF_Dictionary> m_pEncryptDict;
RetainPtr<CPDF_Dictionary> m_pNewEncryptDict;
RetainPtr<CPDF_SecurityHandler> m_pSecurityHandler;
diff --git a/core/fpdfapi/parser/cpdf_parser.cpp b/core/fpdfapi/parser/cpdf_parser.cpp
index 64a718f..336c0b0 100644
--- a/core/fpdfapi/parser/cpdf_parser.cpp
+++ b/core/fpdfapi/parser/cpdf_parser.cpp
@@ -995,6 +995,10 @@
return result;
}
+FX_FILESIZE CPDF_Parser::GetDocumentSize() const {
+ return m_pSyntax->GetDocumentSize();
+}
+
uint32_t CPDF_Parser::GetFirstPageNo() const {
return m_pLinearized ? m_pLinearized->GetFirstPageNo() : 0;
}
@@ -1144,3 +1148,64 @@
std::unique_ptr<CPDF_SyntaxParser> parser) {
m_pSyntax = std::move(parser);
}
+
+std::vector<unsigned int> CPDF_Parser::GetTrailerEnds() {
+ std::vector<unsigned int> trailer_ends;
+ m_pSyntax->SetTrailerEnds(&trailer_ends);
+
+ // Traverse the document.
+ m_pSyntax->SetPos(0);
+ while (true) {
+ CPDF_SyntaxParser::WordResult word_result = m_pSyntax->GetNextWord();
+ if (word_result.is_number) {
+ // The object number was read. Read the generation number.
+ word_result = m_pSyntax->GetNextWord();
+ if (!word_result.is_number)
+ break;
+
+ word_result = m_pSyntax->GetNextWord();
+ if (word_result.word != "obj")
+ break;
+
+ m_pSyntax->GetObjectBody(nullptr);
+
+ word_result = m_pSyntax->GetNextWord();
+ if (word_result.word != "endobj")
+ break;
+ } else if (word_result.word == "trailer") {
+ m_pSyntax->GetObjectBody(nullptr);
+ } else if (word_result.word == "startxref") {
+ m_pSyntax->GetNextWord();
+ } else if (word_result.word == "xref") {
+ while (true) {
+ word_result = m_pSyntax->GetNextWord();
+ if (word_result.word.IsEmpty() || word_result.word == "startxref")
+ break;
+ }
+ m_pSyntax->GetNextWord();
+ } else {
+ break;
+ }
+ }
+
+ // Stop recording trailer ends.
+ m_pSyntax->SetTrailerEnds(nullptr);
+ return trailer_ends;
+}
+
+bool CPDF_Parser::WriteToArchive(IFX_ArchiveStream* archive,
+ FX_FILESIZE src_size) {
+ static constexpr FX_FILESIZE kBufferSize = 4096;
+ DataVector<uint8_t> buffer(kBufferSize);
+ m_pSyntax->SetPos(0);
+ while (src_size) {
+ const uint32_t block_size =
+ static_cast<uint32_t>(std::min(kBufferSize, src_size));
+ if (!m_pSyntax->ReadBlock(buffer.data(), block_size))
+ return false;
+ if (!archive->WriteBlock(buffer.data(), block_size))
+ return false;
+ src_size -= block_size;
+ }
+ return true;
+}
diff --git a/core/fpdfapi/parser/cpdf_parser.h b/core/fpdfapi/parser/cpdf_parser.h
index 0d87606..4d2f7ab 100644
--- a/core/fpdfapi/parser/cpdf_parser.h
+++ b/core/fpdfapi/parser/cpdf_parser.h
@@ -31,6 +31,7 @@
class CPDF_ReadValidator;
class CPDF_SecurityHandler;
class CPDF_SyntaxParser;
+class IFX_ArchiveStream;
class IFX_SeekableReadStream;
class CPDF_Parser {
@@ -109,6 +110,7 @@
RetainPtr<CPDF_Object> ParseIndirectObjectAt(FX_FILESIZE pos,
uint32_t objnum);
+ FX_FILESIZE GetDocumentSize() const;
uint32_t GetFirstPageNo() const;
const CPDF_LinearizedHeader* GetLinearizedHeader() const {
return m_pLinearized.get();
@@ -120,7 +122,8 @@
bool xref_table_rebuilt() const { return m_bXRefTableRebuilt; }
- CPDF_SyntaxParser* GetSyntax() const { return m_pSyntax.get(); }
+ std::vector<unsigned int> GetTrailerEnds();
+ bool WriteToArchive(IFX_ArchiveStream* archive, FX_FILESIZE src_size);
void SetLinearizedHeaderForTesting(
std::unique_ptr<CPDF_LinearizedHeader> pLinearized);
diff --git a/fpdfsdk/fpdf_view.cpp b/fpdfsdk/fpdf_view.cpp
index 1627575..e365b67 100644
--- a/fpdfsdk/fpdf_view.cpp
+++ b/fpdfsdk/fpdf_view.cpp
@@ -22,7 +22,6 @@
#include "core/fpdfapi/parser/cpdf_parser.h"
#include "core/fpdfapi/parser/cpdf_stream.h"
#include "core/fpdfapi/parser/cpdf_string.h"
-#include "core/fpdfapi/parser/cpdf_syntax_parser.h"
#include "core/fpdfapi/parser/fpdf_parser_decode.h"
#include "core/fpdfapi/render/cpdf_docrenderdata.h"
#include "core/fpdfapi/render/cpdf_pagerendercache.h"
@@ -1278,48 +1277,7 @@
// Start recording trailer ends.
auto* parser = doc->GetParser();
- CPDF_SyntaxParser* syntax = parser->GetSyntax();
- std::vector<unsigned int> trailer_ends;
- syntax->SetTrailerEnds(&trailer_ends);
-
- // Traverse the document.
- syntax->SetPos(0);
- while (true) {
- CPDF_SyntaxParser::WordResult word_result = syntax->GetNextWord();
- if (word_result.is_number) {
- // The object number was read. Read the generation number.
- word_result = syntax->GetNextWord();
- if (!word_result.is_number)
- break;
-
- word_result = syntax->GetNextWord();
- if (word_result.word != "obj")
- break;
-
- syntax->GetObjectBody(nullptr);
-
- word_result = syntax->GetNextWord();
- if (word_result.word != "endobj")
- break;
- } else if (word_result.word == "trailer") {
- syntax->GetObjectBody(nullptr);
- } else if (word_result.word == "startxref") {
- syntax->GetNextWord();
- } else if (word_result.word == "xref") {
- while (true) {
- word_result = syntax->GetNextWord();
- if (word_result.word.IsEmpty() || word_result.word == "startxref")
- break;
- }
- syntax->GetNextWord();
- } else {
- break;
- }
- }
-
- // Stop recording trailer ends.
- syntax->SetTrailerEnds(nullptr);
-
+ std::vector<unsigned int> trailer_ends = parser->GetTrailerEnds();
const unsigned long trailer_ends_len =
fxcrt::CollectionSize<unsigned long>(trailer_ends);
if (buffer && length >= trailer_ends_len) {
