Use string pools in some dictionaries, names, and strings.

BUG=pdfium:597

Review-Url: https://codereview.chromium.org/2345063002

5 files changed, 54 insertions, 31 deletions

diff --git a/core/fpdfapi/fpdf_page/cpdf_image.cpp b/core/fpdfapi/fpdf_page/cpdf_image.cpp
index 4112ad86ee..91f1268eae 100644
--- a/core/fpdfapi/fpdf_page/cpdf_image.cpp
+++ b/core/fpdfapi/fpdf_page/cpdf_image.cpp
@@ -85,7 +85,8 @@ CPDF_Dictionary* CPDF_Image::InitJPEG(uint8_t* pData, uint32_t size) {
     return nullptr;
   }
 
-  CPDF_Dictionary* pDict = new CPDF_Dictionary;
+  CPDF_Dictionary* pDict =
+      new CPDF_Dictionary(m_pDocument->GetByteStringPool());
   pDict->SetNameFor("Type", "XObject");
   pDict->SetNameFor("Subtype", "Image");
   pDict->SetIntegerFor("Width", width);
@@ -108,7 +109,8 @@ CPDF_Dictionary* CPDF_Image::InitJPEG(uint8_t* pData, uint32_t size) {
   pDict->SetIntegerFor("BitsPerComponent", bits);
   pDict->SetNameFor("Filter", "DCTDecode");
   if (!color_trans) {
-    CPDF_Dictionary* pParms = new CPDF_Dictionary;
+    CPDF_Dictionary* pParms =
+        new CPDF_Dictionary(m_pDocument->GetByteStringPool());
     pDict->SetFor("DecodeParms", pParms);
     pParms->SetIntegerFor("ColorTransform", 0);
   }
@@ -150,7 +152,8 @@ void CPDF_Image::SetImage(const CFX_DIBitmap* pBitmap, int32_t iCompress) {
   int32_t src_pitch = pBitmap->GetPitch();
   int32_t bpp = pBitmap->GetBPP();
 
-  CPDF_Dictionary* pDict = new CPDF_Dictionary;
+  CPDF_Dictionary* pDict =
+      new CPDF_Dictionary(m_pDocument->GetByteStringPool());
   pDict->SetNameFor("Type", "XObject");
   pDict->SetNameFor("Subtype", "Image");
   pDict->SetIntegerFor("Width", BitmapWidth);
@@ -213,8 +216,9 @@ void CPDF_Image::SetImage(const CFX_DIBitmap* pBitmap, int32_t iCompress) {
         ptr[2] = (uint8_t)argb;
         ptr += 3;
       }
-      CPDF_Stream* pCTS =
-          new CPDF_Stream(pColorTable, iPalette * 3, new CPDF_Dictionary);
+      CPDF_Stream* pCTS = new CPDF_Stream(
+          pColorTable, iPalette * 3,
+          new CPDF_Dictionary(m_pDocument->GetByteStringPool()));
       pCS->AddReference(m_pDocument, m_pDocument->AddIndirectObject(pCTS));
       pDict->SetReferenceFor("ColorSpace", m_pDocument,
                              m_pDocument->AddIndirectObject(pCS));
@@ -249,7 +253,8 @@ void CPDF_Image::SetImage(const CFX_DIBitmap* pBitmap, int32_t iCompress) {
     int32_t maskHeight = pMaskBitmap->GetHeight();
     uint8_t* mask_buf = nullptr;
     FX_STRSIZE mask_size = 0;
-    CPDF_Dictionary* pMaskDict = new CPDF_Dictionary;
+    CPDF_Dictionary* pMaskDict =
+        new CPDF_Dictionary(m_pDocument->GetByteStringPool());
     pMaskDict->SetNameFor("Type", "XObject");
     pMaskDict->SetNameFor("Subtype", "Image");
     pMaskDict->SetIntegerFor("Width", maskWidth);
diff --git a/core/fpdfapi/fpdf_page/fpdf_page_doc.cpp b/core/fpdfapi/fpdf_page/fpdf_page_doc.cpp
index c4e2a24424..192e0947bc 100644
--- a/core/fpdfapi/fpdf_page/fpdf_page_doc.cpp
+++ b/core/fpdfapi/fpdf_page/fpdf_page_doc.cpp
@@ -184,12 +184,13 @@ CPDF_Font* CPDF_DocPageData::GetStandardFont(const CFX_ByteString& fontName,
     return fontData->AddRef();
   }
 
-  CPDF_Dictionary* pDict = new CPDF_Dictionary;
+  CPDF_Dictionary* pDict = new CPDF_Dictionary(m_pPDFDoc->GetByteStringPool());
   pDict->SetNameFor("Type", "Font");
   pDict->SetNameFor("Subtype", "Type1");
   pDict->SetNameFor("BaseFont", fontName);
   if (pEncoding) {
-    pDict->SetFor("Encoding", pEncoding->Realize());
+    pDict->SetFor("Encoding",
+                  pEncoding->Realize(m_pPDFDoc->GetByteStringPool()));
   }
   m_pPDFDoc->AddIndirectObject(pDict);
   std::unique_ptr<CPDF_Font> pFont = CPDF_Font::Create(m_pPDFDoc, pDict);
diff --git a/core/fpdfapi/fpdf_page/fpdf_page_parser.cpp b/core/fpdfapi/fpdf_page/fpdf_page_parser.cpp
index 46ab067b3e..881f9eed25 100644
--- a/core/fpdfapi/fpdf_page/fpdf_page_parser.cpp
+++ b/core/fpdfapi/fpdf_page/fpdf_page_parser.cpp
@@ -281,7 +281,8 @@ void CPDF_StreamContentParser::AddNameParam(const FX_CHAR* name, int len) {
   ContentParam& param = m_ParamBuf[GetNextParamPos()];
   if (len > 32) {
     param.m_Type = ContentParam::OBJECT;
-    param.m_pObject = new CPDF_Name(PDF_NameDecode(bsName));
+    param.m_pObject = new CPDF_Name(
+        m_pDocument->GetByteStringPool()->Intern(PDF_NameDecode(bsName)));
   } else {
     param.m_Type = ContentParam::NAME;
     if (bsName.Find('#') == -1) {
@@ -343,8 +344,8 @@ CPDF_Object* CPDF_StreamContentParser::GetObject(uint32_t index) {
     return pNumber;
   }
   if (param.m_Type == ContentParam::NAME) {
-    CPDF_Name* pName = new CPDF_Name(
-        CFX_ByteString(param.m_Name.m_Buffer, param.m_Name.m_Len));
+    CPDF_Name* pName = new CPDF_Name(m_pDocument->GetByteStringPool()->Intern(
+        CFX_ByteString(param.m_Name.m_Buffer, param.m_Name.m_Len)));
     param.m_Type = ContentParam::OBJECT;
     param.m_pObject = pName;
     return pName;
@@ -589,7 +590,8 @@ void CPDF_StreamContentParser::Handle_BeginMarkedContent_Dictionary() {
 
 void CPDF_StreamContentParser::Handle_BeginImage() {
   FX_FILESIZE savePos = m_pSyntax->GetPos();
-  CPDF_Dictionary* pDict = new CPDF_Dictionary;
+  CPDF_Dictionary* pDict =
+      new CPDF_Dictionary(m_pDocument->GetByteStringPool());
   while (1) {
     CPDF_StreamParser::SyntaxType type = m_pSyntax->ParseNextElement();
     if (type == CPDF_StreamParser::Keyword) {
@@ -1499,7 +1501,7 @@ uint32_t CPDF_StreamContentParser::Parse(const uint8_t* pData,
     return dwSize;
   }
   uint32_t InitObjCount = m_pObjectHolder->GetPageObjectList()->size();
-  CPDF_StreamParser syntax(pData, dwSize);
+  CPDF_StreamParser syntax(pData, dwSize, m_pDocument->GetByteStringPool());
   CPDF_StreamParserAutoClearer auto_clearer(&m_pSyntax, &syntax);
   while (1) {
     uint32_t cost = m_pObjectHolder->GetPageObjectList()->size() - InitObjCount;
diff --git a/core/fpdfapi/fpdf_page/fpdf_page_parser_old.cpp b/core/fpdfapi/fpdf_page/fpdf_page_parser_old.cpp
index eab0ee595b..0e9cd3b55d 100644
--- a/core/fpdfapi/fpdf_page/fpdf_page_parser_old.cpp
+++ b/core/fpdfapi/fpdf_page/fpdf_page_parser_old.cpp
@@ -41,12 +41,22 @@ const FX_STRSIZE kMaxStringLength = 32767;
 
 }  // namespace
 
-CPDF_StreamParser::CPDF_StreamParser(const uint8_t* pData, uint32_t dwSize) {
-  m_pBuf = pData;
-  m_Size = dwSize;
-  m_Pos = 0;
-  m_pLastObj = nullptr;
-}
+CPDF_StreamParser::CPDF_StreamParser(const uint8_t* pData, uint32_t dwSize)
+    : m_pBuf(pData),
+      m_Size(dwSize),
+      m_Pos(0),
+      m_pLastObj(nullptr),
+      m_pPool(nullptr) {}
+
+CPDF_StreamParser::CPDF_StreamParser(
+    const uint8_t* pData,
+    uint32_t dwSize,
+    const CFX_WeakPtr<CFX_ByteStringPool>& pPool)
+    : m_pBuf(pData),
+      m_Size(dwSize),
+      m_Pos(0),
+      m_pLastObj(nullptr),
+      m_pPool(pPool) {}
 
 CPDF_StreamParser::~CPDF_StreamParser() {
   if (m_pLastObj) {
@@ -336,18 +346,21 @@ CPDF_Object* CPDF_StreamParser::ReadNextObject(bool bAllowNestedArray,
 
   int first_char = m_WordBuffer[0];
   if (first_char == '/') {
-    return new CPDF_Name(
-        PDF_NameDecode(CFX_ByteStringC(m_WordBuffer + 1, m_WordSize - 1)));
+    CFX_ByteString name =
+        PDF_NameDecode(CFX_ByteStringC(m_WordBuffer + 1, m_WordSize - 1));
+    return new CPDF_Name(m_pPool ? m_pPool->Intern(name) : name);
   }
 
-  if (first_char == '(')
-    return new CPDF_String(ReadString(), FALSE);
+  if (first_char == '(') {
+    CFX_ByteString str = ReadString();
+    return new CPDF_String(m_pPool ? m_pPool->Intern(str) : str, FALSE);
+  }
 
   if (first_char == '<') {
     if (m_WordSize == 1)
       return new CPDF_String(ReadHexString(), TRUE);
 
-    CPDF_Dictionary* pDict = new CPDF_Dictionary;
+    CPDF_Dictionary* pDict = new CPDF_Dictionary(m_pPool);
     while (1) {
       GetNextWord(bIsNumber);
       if (m_WordSize == 2 && m_WordBuffer[0] == '>')
diff --git a/core/fpdfapi/fpdf_page/pageint.h b/core/fpdfapi/fpdf_page/pageint.h
index 427a363e8e..c5b80aa550 100644
--- a/core/fpdfapi/fpdf_page/pageint.h
+++ b/core/fpdfapi/fpdf_page/pageint.h
@@ -16,6 +16,8 @@
 #include "core/fpdfapi/fpdf_page/cpdf_contentmark.h"
 #include "core/fpdfapi/fpdf_page/cpdf_countedobject.h"
 #include "core/fpdfapi/fpdf_page/include/cpdf_pageobjectholder.h"
+#include "core/fxcrt/include/cfx_string_pool_template.h"
+#include "core/fxcrt/include/cfx_weak_ptr.h"
 #include "core/fxge/include/cfx_pathdata.h"
 #include "core/fxge/include/cfx_renderdevice.h"
 
@@ -43,6 +45,9 @@ class CPDF_StreamParser {
   enum SyntaxType { EndOfData, Number, Keyword, Name, Others };
 
   CPDF_StreamParser(const uint8_t* pData, uint32_t dwSize);
+  CPDF_StreamParser(const uint8_t* pData,
+                    uint32_t dwSize,
+                    const CFX_WeakPtr<CFX_ByteStringPool>& pPool);
   ~CPDF_StreamParser();
 
   CPDF_Stream* ReadInlineStream(CPDF_Document* pDoc,
@@ -66,17 +71,14 @@ class CPDF_StreamParser {
   void GetNextWord(FX_BOOL& bIsNumber);
   CFX_ByteString ReadString();
   CFX_ByteString ReadHexString();
-  const uint8_t* m_pBuf;
-
-  // Length in bytes of m_pBuf.
-  uint32_t m_Size;
-
-  // Current byte position within m_pBuf.
-  uint32_t m_Pos;
 
+  const uint8_t* m_pBuf;
+  uint32_t m_Size;  // Length in bytes of m_pBuf.
+  uint32_t m_Pos;   // Current byte position within m_pBuf.
   uint8_t m_WordBuffer[256];
   uint32_t m_WordSize;
   CPDF_Object* m_pLastObj;
+  CFX_WeakPtr<CFX_ByteStringPool> m_pPool;
 
  private:
   bool PositionIsInBounds() const;