Merge CFX_XMLParser and CFX_XMLSyntaxParser

The CFX_XMLParser was a wrapper around the CFX_XMLSyntaxParser. This CL
merges the SyntaxParser into protected/private methods if the XMLParser.

Change-Id: If1519b5de55866ed14359dffd64dc12c36ee0244
Reviewed-on: https://pdfium-review.googlesource.com/30171
Reviewed-by: Ryan Harrison <rharrison@chromium.org>
Commit-Queue: dsinclair <dsinclair@chromium.org>

1 files changed, 109 insertions, 3 deletions

diff --git a/core/fxcrt/xml/cfx_xmlparser.h b/core/fxcrt/xml/cfx_xmlparser.h
index 8c4c354699..c7587e9151 100644
--- a/core/fxcrt/xml/cfx_xmlparser.h
+++ b/core/fxcrt/xml/cfx_xmlparser.h
@@ -9,29 +9,135 @@
 
 #include <memory>
 #include <stack>
+#include <vector>
 
+#include "core/fxcrt/cfx_blockbuffer.h"
+#include "core/fxcrt/cfx_seekablestreamproxy.h"
 #include "core/fxcrt/fx_string.h"
 #include "core/fxcrt/retain_ptr.h"
-#include "core/fxcrt/xml/cfx_xmlsyntaxparser.h"
+#include "core/fxcrt/xml/cfx_xmlnode.h"
 
 class CFX_XMLElement;
 class CFX_XMLNode;
 class CFX_SeekableStreamProxy;
 
+enum class FX_XmlSyntaxResult {
+  None,
+  InstructionOpen,
+  InstructionClose,
+  ElementOpen,
+  ElementBreak,
+  ElementClose,
+  TargetName,
+  TagName,
+  AttriName,
+  AttriValue,
+  Text,
+  CData,
+  TargetData,
+  Error,
+  EndOfString
+};
+
 class CFX_XMLParser {
  public:
+  static bool IsXMLNameChar(wchar_t ch, bool bFirstChar);
+
   CFX_XMLParser(CFX_XMLNode* pParent,
                 const RetainPtr<CFX_SeekableStreamProxy>& pStream);
-  ~CFX_XMLParser();
+  virtual ~CFX_XMLParser();
 
   int32_t Parse();
 
+ protected:
+  FX_XmlSyntaxResult DoSyntaxParse();
+
+  WideString GetTagName() const {
+    return m_BlockBuffer.GetTextData(0, m_iTextDataLength);
+  }
+
+  WideString GetAttributeName() const {
+    return m_BlockBuffer.GetTextData(0, m_iTextDataLength);
+  }
+
+  WideString GetAttributeValue() const {
+    return m_BlockBuffer.GetTextData(0, m_iTextDataLength);
+  }
+
+  WideString GetTextData() const {
+    return m_BlockBuffer.GetTextData(0, m_iTextDataLength);
+  }
+
  private:
-  std::unique_ptr<CFX_XMLSyntaxParser> m_pParser;
+  enum class FDE_XmlSyntaxState {
+    Text,
+    Node,
+    Target,
+    Tag,
+    AttriName,
+    AttriEqualSign,
+    AttriQuotation,
+    AttriValue,
+    Entity,
+    EntityDecimal,
+    EntityHex,
+    CloseInstruction,
+    BreakElement,
+    CloseElement,
+    SkipDeclNode,
+    DeclCharData,
+    SkipComment,
+    SkipCommentOrDecl,
+    SkipCData,
+    TargetData
+  };
+
+  void ParseTextChar(wchar_t ch);
+
+  int32_t GetStatus() const;
+  FX_FILESIZE GetCurrentPos() const { return m_ParsedChars + m_Start; }
+  FX_FILESIZE GetCurrentBinaryPos() const;
+  int32_t GetCurrentNodeNumber() const { return m_iCurrentNodeNum; }
+  int32_t GetLastNodeNumber() const { return m_iLastNodeNum; }
+
+  WideString GetTargetName() const {
+    return m_BlockBuffer.GetTextData(0, m_iTextDataLength);
+  }
+
+  WideString GetTargetData() const {
+    return m_BlockBuffer.GetTextData(0, m_iTextDataLength);
+  }
+
   CFX_XMLNode* m_pParent;
   CFX_XMLNode* m_pChild;
   std::stack<CFX_XMLNode*> m_NodeStack;
   WideString m_ws1;
+
+  RetainPtr<CFX_SeekableStreamProxy> m_pStream;
+  size_t m_iXMLPlaneSize;
+  FX_FILESIZE m_iCurrentPos;
+  int32_t m_iCurrentNodeNum;
+  int32_t m_iLastNodeNum;
+  int32_t m_iParsedBytes;
+  FX_FILESIZE m_ParsedChars;
+  std::vector<wchar_t> m_Buffer;
+  size_t m_iBufferChars;
+  bool m_bEOS;
+  FX_FILESIZE m_Start;  // Start position in m_Buffer
+  FX_FILESIZE m_End;    // End position in m_Buffer
+  FX_XMLNODE m_CurNode;
+  std::stack<FX_XMLNODE> m_XMLNodeStack;
+  CFX_BlockBuffer m_BlockBuffer;
+  int32_t m_iAllocStep;
+  wchar_t* m_pCurrentBlock;  // Pointer into CFX_BlockBuffer
+  int32_t m_iIndexInBlock;
+  int32_t m_iTextDataLength;
+  FX_XmlSyntaxResult m_syntaxParserResult;
+  FDE_XmlSyntaxState m_syntaxParserState;
+  wchar_t m_wQuotationMark;
+  int32_t m_iEntityStart;
+  std::stack<wchar_t> m_SkipStack;
+  wchar_t m_SkipChar;
 };
 
 #endif  // CORE_FXCRT_XML_CFX_XMLPARSER_H_