summaryrefslogtreecommitdiff
path: root/core/fpdfapi/parser
diff options
context:
space:
mode:
authorLei Zhang <thestig@chromium.org>2017-08-02 17:41:22 -0700
committerChromium commit bot <commit-bot@chromium.org>2017-08-03 01:05:51 +0000
commitd1a8458e6390103e123e9d265040b3d02c16955b (patch)
tree3b169237c62b6b6210722ed403314b52aea34d8f /core/fpdfapi/parser
parentf44812cf6d7559dd2a0bbc25a684e54675f4794d (diff)
downloadpdfium-d1a8458e6390103e123e9d265040b3d02c16955b.tar.xz
Avoid a redundant header offset check in CPDF_Parser.
CPDF_Parser::StartLinearizedParse() calls StartParse(), but already knows the PDF header offset. Refactor StartParse() so it does not have to look for the header again. Change-Id: Id8cc39301ae72da868dafc53921622d5b28ce26e Reviewed-on: https://pdfium-review.googlesource.com/9830 Commit-Queue: Lei Zhang <thestig@chromium.org> Reviewed-by: Art Snake <art-snake@yandex-team.ru>
Diffstat (limited to 'core/fpdfapi/parser')
-rw-r--r--core/fpdfapi/parser/cpdf_data_avail.cpp2
-rw-r--r--core/fpdfapi/parser/cpdf_parser.cpp26
-rw-r--r--core/fpdfapi/parser/cpdf_parser.h3
-rw-r--r--core/fpdfapi/parser/fpdf_parser_utility.cpp4
-rw-r--r--core/fpdfapi/parser/fpdf_parser_utility.h6
5 files changed, 31 insertions, 10 deletions
diff --git a/core/fpdfapi/parser/cpdf_data_avail.cpp b/core/fpdfapi/parser/cpdf_data_avail.cpp
index 88755a9eb7..46c71a5ab9 100644
--- a/core/fpdfapi/parser/cpdf_data_avail.cpp
+++ b/core/fpdfapi/parser/cpdf_data_avail.cpp
@@ -754,7 +754,7 @@ bool CPDF_DataAvail::IsLinearizedFile(uint8_t* pData, uint32_t dwLen) {
auto file = pdfium::MakeRetain<CFX_MemoryStream>(
pData, static_cast<size_t>(dwLen), false);
int32_t offset = GetHeaderOffset(file);
- if (offset == -1) {
+ if (offset == kInvalidHeaderOffset) {
m_docStatus = PDF_DATAAVAIL_ERROR;
return false;
}
diff --git a/core/fpdfapi/parser/cpdf_parser.cpp b/core/fpdfapi/parser/cpdf_parser.cpp
index e33cec0165..d669f24674 100644
--- a/core/fpdfapi/parser/cpdf_parser.cpp
+++ b/core/fpdfapi/parser/cpdf_parser.cpp
@@ -148,14 +148,26 @@ void CPDF_Parser::ShrinkObjectMap(uint32_t objnum) {
CPDF_Parser::Error CPDF_Parser::StartParse(
const CFX_RetainPtr<IFX_SeekableReadStream>& pFileAccess,
CPDF_Document* pDocument) {
+ return StartParseInternal(pFileAccess, pDocument, kInvalidHeaderOffset);
+}
+
+CPDF_Parser::Error CPDF_Parser::StartParseInternal(
+ const CFX_RetainPtr<IFX_SeekableReadStream>& pFileAccess,
+ CPDF_Document* pDocument,
+ int32_t iHeaderOffset) {
ASSERT(!m_bHasParsed);
m_bHasParsed = true;
m_bXRefStream = false;
m_LastXRefOffset = 0;
- int32_t offset = GetHeaderOffset(pFileAccess);
- if (offset == -1)
- return FORMAT_ERROR;
+ int32_t offset;
+ if (iHeaderOffset == kInvalidHeaderOffset) {
+ offset = GetHeaderOffset(pFileAccess);
+ if (offset == kInvalidHeaderOffset)
+ return FORMAT_ERROR;
+ } else {
+ offset = iHeaderOffset;
+ }
m_pSyntax->InitParser(pFileAccess, offset);
@@ -1521,12 +1533,12 @@ CPDF_Parser::Error CPDF_Parser::StartLinearizedParse(
m_LastXRefOffset = 0;
int32_t offset = GetHeaderOffset(pFileAccess);
- if (offset == -1)
+ if (offset == kInvalidHeaderOffset)
return FORMAT_ERROR;
- if (!IsLinearizedFile(pFileAccess, offset)) {
- return StartParse(pFileAccess, std::move(pDocument));
- }
+ if (!IsLinearizedFile(pFileAccess, offset))
+ return StartParseInternal(pFileAccess, std::move(pDocument), offset);
+
m_bHasParsed = true;
m_pDocument = pDocument;
diff --git a/core/fpdfapi/parser/cpdf_parser.h b/core/fpdfapi/parser/cpdf_parser.h
index 759d042360..189de168f4 100644
--- a/core/fpdfapi/parser/cpdf_parser.h
+++ b/core/fpdfapi/parser/cpdf_parser.h
@@ -153,6 +153,9 @@ class CPDF_Parser {
ObjectInfo info;
};
+ Error StartParseInternal(const CFX_RetainPtr<IFX_SeekableReadStream>& pFile,
+ CPDF_Document* pDocument,
+ int32_t iHeaderOffset);
CPDF_Object* ParseDirect(CPDF_Object* pObj);
bool LoadAllCrossRefV4(FX_FILESIZE pos);
bool LoadAllCrossRefV5(FX_FILESIZE pos);
diff --git a/core/fpdfapi/parser/fpdf_parser_utility.cpp b/core/fpdfapi/parser/fpdf_parser_utility.cpp
index 8323426e74..0c0ca6644f 100644
--- a/core/fpdfapi/parser/fpdf_parser_utility.cpp
+++ b/core/fpdfapi/parser/fpdf_parser_utility.cpp
@@ -75,12 +75,12 @@ int32_t GetHeaderOffset(const CFX_RetainPtr<IFX_SeekableReadStream>& pFile) {
uint8_t buf[kBufSize];
for (int32_t offset = 0; offset <= 1024; ++offset) {
if (!pFile->ReadBlock(buf, offset, kBufSize))
- return -1;
+ return kInvalidHeaderOffset;
if (memcmp(buf, "%PDF", 4) == 0)
return offset;
}
- return -1;
+ return kInvalidHeaderOffset;
}
int32_t GetDirectInteger(CPDF_Dictionary* pDict, const CFX_ByteString& key) {
diff --git a/core/fpdfapi/parser/fpdf_parser_utility.h b/core/fpdfapi/parser/fpdf_parser_utility.h
index eb8442ac68..5e764ad07b 100644
--- a/core/fpdfapi/parser/fpdf_parser_utility.h
+++ b/core/fpdfapi/parser/fpdf_parser_utility.h
@@ -34,7 +34,13 @@ inline bool PDFCharIsLineEnding(uint8_t c) {
return c == '\r' || c == '\n';
}
+constexpr int32_t kInvalidHeaderOffset = -1;
+
+// On success, return a positive offset value to the PDF header.. If the header
+// cannot be found, or if there is an error reading from |pFile|, then return
+// |kInvalidHeaderOffset|.
int32_t GetHeaderOffset(const CFX_RetainPtr<IFX_SeekableReadStream>& pFile);
+
int32_t GetDirectInteger(CPDF_Dictionary* pDict, const CFX_ByteString& key);
CFX_ByteTextBuf& operator<<(CFX_ByteTextBuf& buf, const CPDF_Object* pObj);