added stubs for expat

htmlparser/src/nsExpatDTD.cpp

@@ -0,0 +1,470 @@
+/* -*- Mode: C++; tab-width: 2; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
+/*
+ * The contents of this file are subject to the Netscape Public License
+ * Version 1.0 (the "NPL"); you may not use this file except in
+ * compliance with the NPL.  You may obtain a copy of the NPL at
+ * http://www.mozilla.org/NPL/
+ *
+ * Software distributed under the NPL is distributed on an "AS IS" basis,
+ * WITHOUT WARRANTY OF ANY KIND, either express or implied. See the NPL
+ * for the specific language governing rights and limitations under the
+ * NPL.
+ *
+ * The Initial Developer of this code under the NPL is Netscape
+ * Communications Corporation.  Portions created by Netscape are
+ * Copyright (C) 1998 Netscape Communications Corporation.  All Rights
+ * Reserved.
+ */
+
+/**
+ * MODULE NOTES:
+ * @update  gess 4/8/98
+ * 
+ *         
+ */
+
+
+#include "nsIDTDDebug.h"
+#include "nsExpatDTD.h"
+#include "nsCRT.h"
+#include "nsParser.h"
+#include "nsScanner.h"
+#include "nsIParser.h"
+#include "nsTokenHandler.h"
+#include "nsDTDUtils.h"
+#include "nsIContentSink.h"
+#include "nsIHTMLContentSink.h"
+#include "nsExpatTokenizer.h"
+
+#include "prenv.h"  //this is here for debug reasons...
+#include "prtypes.h"  //this is here for debug reasons...
+#include "prio.h"
+#include "plstr.h"
+
+#ifdef XP_PC
+#include <direct.h> //this is here for debug reasons...
+#endif
+#include "prmem.h"
+
+
+static NS_DEFINE_IID(kISupportsIID, NS_ISUPPORTS_IID);                 
+static NS_DEFINE_IID(kIDTDIID,      NS_IDTD_IID);
+static NS_DEFINE_IID(kClassIID,     NS_EXPAT_DTD_IID); 
+
+static const char* kViewSourceCommand= "view-source";
+
+
+static CTokenRecycler gTokenRecycler;
+
+
+/**
+ *  This method gets called as part of our COM-like interfaces.
+ *  Its purpose is to create an interface to parser object
+ *  of some type.
+ *  
+ *  @update   gess 4/8/98
+ *  @param    nsIID  id of object to discover
+ *  @param    aInstancePtr ptr to newly discovered interface
+ *  @return   NS_xxx result code
+ */
+nsresult nsExpatDTD::QueryInterface(const nsIID& aIID, void** aInstancePtr)  
+{                                                                        
+  if (NULL == aInstancePtr) {                                            
+    return NS_ERROR_NULL_POINTER;                                        
+  }                                                                      
+
+  if(aIID.Equals(kISupportsIID))    {  //do IUnknown...
+    *aInstancePtr = (nsIDTD*)(this);                                        
+  }
+  else if(aIID.Equals(kIDTDIID)) {  //do IParser base class...
+    *aInstancePtr = (nsIDTD*)(this);                                        
+  }
+  else if(aIID.Equals(kClassIID)) {  //do this class...
+    *aInstancePtr = (nsExpatDTD*)(this);                                        
+  }                 
+  else {
+    *aInstancePtr=0;
+    return NS_NOINTERFACE;
+  }
+  NS_ADDREF_THIS();
+  return NS_OK;                                                        
+}
+
+/**
+ *  This method is defined in nsIParser. It is used to 
+ *  cause the COM-like construction of an nsParser.
+ *  
+ *  @update  gess 4/8/98
+ *  @param   nsIParser** ptr to newly instantiated parser
+ *  @return  NS_xxx error result
+ */
+NS_HTMLPARS nsresult NS_New_Expat_DTD(nsIDTD** aInstancePtrResult) {
+  nsExpatDTD* it = new nsExpatDTD();
+  if (it == 0) {
+    return NS_ERROR_OUT_OF_MEMORY;
+  }
+  return it->QueryInterface(kClassIID, (void **) aInstancePtrResult);
+}
+
+
+NS_IMPL_ADDREF(nsExpatDTD)
+NS_IMPL_RELEASE(nsExpatDTD)
+
+
+/**
+ *  Default constructor
+ *  
+ *  @update  gess 4/9/98
+ *  @param   
+ *  @return  
+ */
+nsExpatDTD::nsExpatDTD() : nsIDTD() {
+  NS_INIT_REFCNT();
+  mParser=0;
+  mSink=0;
+  mFilename;
+  mLineNumber=0;
+  mTokenizer=0;
+}
+
+/**
+ *  Default destructor
+ *  
+ *  @update  gess 4/9/98
+ *  @param   
+ *  @return  
+ */
+nsExpatDTD::~nsExpatDTD(){
+  mParser=0; //just to prove we destructed...
+  if(mTokenizer)
+    delete (nsExpatTokenizer*)mTokenizer;
+}
+
+/**
+ * 
+ * @update	gess1/8/99
+ * @param 
+ * @return
+ */
+const nsIID& nsExpatDTD::GetMostDerivedIID(void) const{
+  return kClassIID;
+}
+
+/**
+ * Call this method if you want the DTD to construct a fresh 
+ * instance of itself. 
+ * @update	gess7/23/98
+ * @param 
+ * @return
+ */
+nsresult nsExpatDTD::CreateNewInstance(nsIDTD** aInstancePtrResult){
+  return NS_New_Expat_DTD(aInstancePtrResult);
+}
+
+/**
+ * This method is called to determine if the given DTD can parse
+ * a document in a given source-type. 
+ * NOTE: Parsing always assumes that the end result will involve
+ *       storing the result in the main content model.
+ * @update	gess6/24/98
+ * @param   
+ * @return  TRUE if this DTD can satisfy the request; FALSE otherwise.
+ */
+eAutoDetectResult nsExpatDTD::CanParse(nsString& aContentType, nsString& aCommand, nsString& aBuffer, PRInt32 aVersion) {
+  eAutoDetectResult result=eUnknownDetect;
+
+  if(!aCommand.Equals(kViewSourceCommand)) {
+    if(aContentType.Equals(kXMLTextContentType) ||
+       aContentType.Equals(kRDFTextContentType) ||
+       aContentType.Equals(kXULTextContentType)) {
+      result=eValidDetect;
+    }
+    else {
+      if(-1<aBuffer.Find("<?xml ")) {
+        if(0==aContentType.Length()) {
+          aContentType = kXMLTextContentType; //only reset it if it's empty
+        }
+        result=eValidDetect;
+      }
+    }
+  }
+  return result;
+}
+
+
+/**
+ * 
+ * @update	gess5/18/98
+ * @param 
+ * @return
+ */
+NS_IMETHODIMP nsExpatDTD::WillBuildModel(nsString& aFilename,PRBool aNotifySink,nsIParser* aParser,nsIContentSink* aSink){
+  nsresult result=NS_OK;
+  mFilename=aFilename;
+
+  if(aParser){
+    mSink=aParser->GetContentSink();
+    if((aNotifySink) && (mSink)) {
+      mLineNumber=0;
+      result = mSink->WillBuildModel();
+    }
+  }
+
+  return result;
+}
+
+/**
+  * The parser uses a code sandwich to wrap the parsing process. Before
+  * the process begins, WillBuildModel() is called. Afterwards the parser
+  * calls DidBuildModel(). 
+  * @update	gess  1/4/99
+  * @param	aFilename is the name of the file being parsed.
+  * @return	error code (almost always 0)
+  */
+NS_IMETHODIMP nsExpatDTD::BuildModel(nsIParser* aParser,nsITokenizer* aTokenizer,nsITokenObserver* anObserver,nsIContentSink* aSink) {
+  nsresult result=NS_OK;
+
+  if(aTokenizer) {
+    nsITokenizer*  oldTokenizer=mTokenizer;
+    mTokenizer=aTokenizer;
+    nsITokenRecycler* theRecycler=aTokenizer->GetTokenRecycler();
+
+    while(NS_OK==result){
+      CToken* theToken=mTokenizer->PopToken();
+      if(theToken) {
+        result=HandleToken(theToken,aParser);
+        if(NS_SUCCEEDED(result)) {
+          theRecycler->RecycleToken(theToken);
+        }
+        else if(NS_ERROR_HTMLPARSER_BLOCK!=result){
+          mTokenizer->PushTokenFront(theToken);
+        }
+        // theRootDTD->Verify(kEmptyString,aParser);
+      }
+      else break;
+    }//while
+    mTokenizer=oldTokenizer;
+  }
+  else result=NS_ERROR_HTMLPARSER_BADTOKENIZER;
+  return result;
+}
+
+
+/**
+ * 
+ * @update	gess5/18/98
+ * @param 
+ * @return
+ */
+NS_IMETHODIMP nsExpatDTD::DidBuildModel(nsresult anErrorCode,PRBool aNotifySink,nsIParser* aParser,nsIContentSink* aSink){
+  nsresult result= NS_OK;
+
+  //ADD CODE HERE TO CLOSE OPEN CONTAINERS...
+
+  if(aParser){
+    mSink=aParser->GetContentSink();
+    if((aNotifySink) && (mSink)) {
+        result = mSink->DidBuildModel(1);
+    }
+  }
+  return result;
+}
+
+/**
+ * 
+ * @update	gess8/4/98
+ * @param 
+ * @return
+ */
+nsITokenRecycler* nsExpatDTD::GetTokenRecycler(void){
+  nsITokenizer* theTokenizer=GetTokenizer();
+  return theTokenizer->GetTokenRecycler();
+}
+
+/**
+ * Retrieve the preferred tokenizer for use by this DTD.
+ * @update	gess12/28/98
+ * @param   none
+ * @return  ptr to tokenizer
+ */
+nsITokenizer* nsExpatDTD::GetTokenizer(void) {
+  if(!mTokenizer)
+    mTokenizer=new nsExpatTokenizer();
+  return mTokenizer;
+}
+
+/**
+ * 
+ * @update	gess5/18/98
+ * @param 
+ * @return
+ */
+NS_IMETHODIMP nsExpatDTD::WillResumeParse(void){
+  nsresult result = NS_OK;
+  if(mSink) {
+    result = mSink->WillResume();
+  }
+  return result;
+}
+
+/**
+ * 
+ * @update	gess5/18/98
+ * @param 
+ * @return
+ */
+NS_IMETHODIMP nsExpatDTD::WillInterruptParse(void){
+  nsresult result = NS_OK;
+  if(mSink) {
+    result = mSink->WillInterrupt();
+  }
+  return result;
+}
+
+/**
+ * Called by the parser to initiate dtd verification of the
+ * internal context stack.
+ * @update	gess 7/23/98
+ * @param 
+ * @return
+ */
+PRBool nsExpatDTD::Verify(nsString& aURLRef,nsIParser* aParser) {
+  PRBool result=PR_TRUE;
+  mParser=(nsParser*)aParser;
+  return result;
+}
+
+/**
+ * Called by the parser to enable/disable dtd verification of the
+ * internal context stack.
+ * @update	gess 7/23/98
+ * @param 
+ * @return
+ */
+void nsExpatDTD::SetVerification(PRBool aEnabled){
+}
+
+/**
+ *  This method is called to determine whether or not a tag
+ *  of one type can contain a tag of another type.
+ *  
+ *  @update  gess 3/25/98
+ *  @param   aParent -- int tag of parent container
+ *  @param   aChild -- int tag of child container
+ *  @return  PR_TRUE if parent can contain child
+ */
+PRBool nsExpatDTD::CanContain(PRInt32 aParent,PRInt32 aChild) const{
+  PRBool result=PR_TRUE;
+  return result;
+}
+
+/**
+ *  This method gets called to determine whether a given 
+ *  tag is itself a container
+ *  
+ *  @update  gess 3/25/98
+ *  @param   aTag -- tag to test for containership
+ *  @return  PR_TRUE if given tag can contain other tags
+ */
+PRBool nsExpatDTD::IsContainer(PRInt32 aTag) const{
+  PRBool result=PR_TRUE;
+  return result;
+}
+
+/**
+ *  
+ *  @update  vidur 11/12/98
+ *  @param   aToken -- token object to be put into content model
+ *  @return  0 if all is well; non-zero is an error
+ */
+NS_IMETHODIMP nsExpatDTD::HandleToken(CToken* aToken,nsIParser* aParser) {
+  nsresult        result=NS_OK;
+  CHTMLToken*     theToken= (CHTMLToken*)(aToken);
+  eHTMLTokenTypes theType= (eHTMLTokenTypes)theToken->GetTokenType();
+
+  mParser=(nsParser*)aParser;
+  mSink=aParser->GetContentSink();
+
+  nsCParserNode theNode(theToken,mLineNumber);
+  switch(theType) {
+
+    case eToken_newline:
+      mLineNumber++; //now fall through
+    case eToken_entity:
+    case eToken_whitespace:
+    case eToken_text:
+    case eToken_cdatasection:
+      result=mSink->AddLeaf(theNode); 
+      break;
+
+    case eToken_comment:
+      result=mSink->AddComment(theNode); 
+      break;
+    
+    case eToken_instruction:
+      result=mSink->AddProcessingInstruction(theNode); 
+      break;
+
+    case eToken_start:
+      {
+        PRInt16 attrCount=aToken->GetAttributeCount();
+
+        if(0<attrCount){ //go collect the attributes...
+          int attr=0;
+          for(attr=0;attr<attrCount;attr++){
+            CToken* theToken=mTokenizer->PeekToken();
+            if(theToken)  {
+              eHTMLTokenTypes theType=eHTMLTokenTypes(theToken->GetTokenType());
+              if(eToken_attribute==theType){
+                mTokenizer->PopToken(); //pop it for real...
+                theNode.AddAttribute(theToken);
+              } 
+            }
+            else return kEOF;
+          }
+        }
+        if(NS_OK==result){
+          result=mSink->OpenContainer(theNode); 
+          if(((CStartToken*)aToken)->IsEmpty()){
+            result=mSink->CloseContainer(theNode); 
+          }
+        }
+      }
+      break;
+
+    case eToken_end:
+      result=mSink->CloseContainer(theNode); 
+      break;
+
+    case eToken_style:
+    case eToken_skippedcontent:
+    default:
+      result=NS_OK;
+  }//switch
+  return result;
+}
+
+
+/**
+ *  This method causes all tokens to be dispatched to the given tag handler.
+ *
+ *  @update  gess 3/25/98
+ *  @param   aHandler -- object to receive subsequent tokens...
+ *  @return	 error code (usually 0)
+ */
+nsresult nsExpatDTD::CaptureTokenPump(nsITagHandler* aHandler) {
+  nsresult result=NS_OK;
+  return result;
+}
+
+/**
+ *  This method releases the token-pump capture obtained in CaptureTokenPump()
+ *
+ *  @update  gess 3/25/98
+ *  @param   aHandler -- object that received tokens...
+ *  @return	 error code (usually 0)
+ */
+nsresult nsExpatDTD::ReleaseTokenPump(nsITagHandler* aHandler){
+  nsresult result=NS_OK;
+  return result;
+}

htmlparser/src/nsExpatDTD.h

@@ -0,0 +1,238 @@
+
+/* -*- Mode: C++; tab-width: 2; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
+/*
+ * The contents of this file are subject to the Netscape Public License
+ * Version 1.0 (the "NPL"); you may not use this file except in
+ * compliance with the NPL.  You may obtain a copy of the NPL at
+ * http://www.mozilla.org/NPL/
+ *
+ * Software distributed under the NPL is distributed on an "AS IS" basis,
+ * WITHOUT WARRANTY OF ANY KIND, either express or implied. See the NPL
+ * for the specific language governing rights and limitations under the
+ * NPL.
+ *
+ * The Initial Developer of this code under the NPL is Netscape
+ * Communications Corporation.  Portions created by Netscape are
+ * Copyright (C) 1998 Netscape Communications Corporation.  All Rights
+ * Reserved.
+ */
+
+/**
+ * MODULE NOTES:
+ * @update  gess 4/8/98
+ * 
+ *         
+ */
+
+#ifndef __NS_EXPAT_DTD
+#define __NS_EXPAT_DTD
+
+#include "nsIDTD.h"
+#include "nsISupports.h"
+#include "nsHTMLTokens.h"
+#include "nshtmlpars.h"
+#include "nsVoidArray.h"
+#include "nsDeque.h"
+#include "nsIContentSink.h"
+
+
+#define NS_EXPAT_DTD_IID      \
+  {0x5ad4b431, 0xcabb, 0x11d2, \
+  {0xab, 0xcb, 0x0, 0x10, 0x4b, 0x98, 0x3f, 0xd4}}
+
+
+class nsIDTDDebug;
+class nsIParserNode;
+class CITokenHandler;
+class nsParser;
+
+
+
+class nsExpatDTD : public nsIDTD {
+            
+  public:
+
+    NS_DECL_ISUPPORTS
+
+
+    /**
+     *  
+     *  
+     *  @update  gess 4/9/98
+     *  @param   
+     *  @return  
+     */
+    nsExpatDTD();
+
+    /**
+     *  
+     *  
+     *  @update  gess 4/9/98
+     *  @param   
+     *  @return  
+     */
+    virtual ~nsExpatDTD();
+
+    virtual const nsIID&  GetMostDerivedIID(void) const;
+
+    /**
+     * Call this method if you want the DTD to construct a clone of itself.
+     * @update	gess7/23/98
+     * @param 
+     * @return
+     */
+    virtual nsresult CreateNewInstance(nsIDTD** aInstancePtrResult);
+
+    /**
+     * This method is called to determine if the given DTD can parse
+     * a document in a given source-type. 
+     * NOTE: Parsing always assumes that the end result will involve
+     *       storing the result in the main content model.
+     * @update	gess6/24/98
+     * @param   
+     * @return  TRUE if this DTD can satisfy the request; FALSE otherwise.
+     */
+    virtual eAutoDetectResult CanParse(nsString& aContentType, nsString& aCommand, nsString& aBuffer, PRInt32 aVersion);
+
+    /**
+      * The parser uses a code sandwich to wrap the parsing process. Before
+      * the process begins, WillBuildModel() is called. Afterwards the parser
+      * calls DidBuildModel(). 
+      * @update	gess5/18/98
+      * @param	aFilename is the name of the file being parsed.
+      * @return	error code (almost always 0)
+      */
+    NS_IMETHOD WillBuildModel(nsString& aFilename,PRBool aNotifySink,nsIParser* aParser,nsIContentSink* aSink=0);
+
+    /**
+      * The parser uses a code sandwich to wrap the parsing process. Before
+      * the process begins, WillBuildModel() is called. Afterwards the parser
+      * calls DidBuildModel(). 
+      * @update	gess5/18/98
+      * @param	aFilename is the name of the file being parsed.
+      * @return	error code (almost always 0)
+      */
+    NS_IMETHOD BuildModel(nsIParser* aParser,nsITokenizer* aTokenizer,nsITokenObserver* anObserver=0,nsIContentSink* aSink=0);
+
+   /**
+     * The parser uses a code sandwich to wrap the parsing process. Before
+     * the process begins, WillBuildModel() is called. Afterwards the parser
+     * calls DidBuildModel(). 
+     * @update	gess5/18/98
+     * @param	anErrorCode contans the last error that occured
+     * @return	error code
+     */
+    NS_IMETHOD DidBuildModel(nsresult anErrorCode,PRBool aNotifySink,nsIParser* aParser,nsIContentSink* aSink=0);
+
+    /**
+     *  
+     *  @update  gess 3/25/98
+     *  @param   aToken -- token object to be put into content model
+     *  @return  0 if all is well; non-zero is an error
+     */
+    NS_IMETHOD HandleToken(CToken* aToken,nsIParser* aParser);
+
+    /**
+     *  This method causes all tokens to be dispatched to the given tag handler.
+     *
+     *  @update  gess 3/25/98
+  	 *  @param   aHandler -- object to receive subsequent tokens...
+	   *  @return	 error code (usually 0)
+     */
+    NS_IMETHOD CaptureTokenPump(nsITagHandler* aHandler);
+
+    /**
+     *  This method releases the token-pump capture obtained in CaptureTokenPump()
+     *
+     *  @update  gess 3/25/98
+  	 *  @param   aHandler -- object that received tokens...
+	   *  @return	 error code (usually 0)
+     */
+    NS_IMETHOD ReleaseTokenPump(nsITagHandler* aHandler);
+
+    /**
+     * 
+     * @update	gess12/28/98
+     * @param 
+     * @return
+     */
+    nsITokenizer* GetTokenizer(void);
+
+    /**
+     * 
+     * @update	gess5/18/98
+     * @param 
+     * @return
+     */
+    NS_IMETHOD WillResumeParse(void);
+
+    /**
+     * 
+     * @update	gess5/18/98
+     * @param 
+     * @return
+     */
+    NS_IMETHOD WillInterruptParse(void);
+
+    /**
+     * Called by the parser to initiate dtd verification of the
+     * internal context stack.
+     * @update	gess 7/23/98
+     * @param 
+     * @return
+     */
+    virtual PRBool Verify(nsString& aURLRef,nsIParser* aParser);
+
+    /**
+     * Set this to TRUE if you want the DTD to verify its
+     * context stack.
+     * @update	gess 7/23/98
+     * @param 
+     * @return
+     */
+    virtual void SetVerification(PRBool aEnable);
+
+    /**
+     *  This method is called to determine whether or not a tag
+     *  of one type can contain a tag of another type.
+     *  
+     *  @update  gess 3/25/98
+     *  @param   aParent -- int tag of parent container
+     *  @param   aChild -- int tag of child container
+     *  @return  PR_TRUE if parent can contain child
+     */
+    virtual PRBool CanContain(PRInt32 aParent,PRInt32 aChild) const;
+
+    /**
+     *  This method gets called to determine whether a given 
+     *  tag is itself a container
+     *  
+     *  @update  gess 3/25/98
+     *  @param   aTag -- tag to test for containership
+     *  @return  PR_TRUE if given tag can contain other tags
+     */
+    virtual PRBool IsContainer(PRInt32 aTag) const;
+
+    /**
+     * Retrieve a ptr to the global token recycler...
+     * @update	gess8/4/98
+     * @return  ptr to recycler (or null)
+     */
+    virtual nsITokenRecycler* GetTokenRecycler(void);
+    
+protected:
+  
+    nsParser*           mParser;
+    nsIContentSink*     mSink;
+    nsString            mFilename;
+    PRInt32             mLineNumber;
+    nsITokenizer*       mTokenizer;
+};
+
+extern NS_HTMLPARS nsresult NS_New_Expat_DTD(nsIDTD** aInstancePtrResult);
+
+
+#endif 
+
+
+

htmlparser/src/nsExpatTokenizer.cpp

@@ -0,0 +1,141 @@
+/* -*- Mode: C++; tab-width: 2; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
+/* 
+ * The contents of this file are subject to the Netscape Public License
+ * Version 1.0 (the "NPL"); you may not use this file except in
+ * compliance with the NPL.  You may obtain a copy of the NPL at
+ * http://www.mozilla.org/NPL/
+ *
+ * Software distributed under the NPL is distributed on an "AS IS" basis,
+ * WITHOUT WARRANTY OF ANY KIND, either express or implied. See the NPL
+ * for the specific language governing rights and limitations under the
+ * NPL.
+ *
+ * The Initial Developer of this code under the NPL is Netscape
+ * Communications Corporation.  Portions created by Netscape are
+ * Copyright (C) 1998 Netscape Communications Corporation.  All Rights
+ * Reserved.
+ */
+
+
+/**
+ * MODULE NOTES:
+ * @update  gess 4/1/98
+ * 
+ */
+
+#include "nsExpatTokenizer.h"
+#include "nsParser.h"
+#include "nsScanner.h"
+#include "nsDTDUtils.h"
+#include "nsParser.h"
+
+/************************************************************************
+  And now for the main class -- nsExpatTokenizer...
+ ************************************************************************/
+
+static NS_DEFINE_IID(kISupportsIID,     NS_ISUPPORTS_IID);                 
+static NS_DEFINE_IID(kITokenizerIID,    NS_ITOKENIZER_IID);
+static NS_DEFINE_IID(kHTMLTokenizerIID, NS_HTMLTOKENIZER_IID);
+static NS_DEFINE_IID(kClassIID,         NS_EXPATTOKENIZER_IID); 
+static nsAutoString gEmpty;
+
+/**
+ *  This method gets called as part of our COM-like interfaces.
+ *  Its purpose is to create an interface to parser object
+ *  of some type.
+ *  
+ *  @update   gess 4/8/98
+ *  @param    nsIID  id of object to discover
+ *  @param    aInstancePtr ptr to newly discovered interface
+ *  @return   NS_xxx result code
+ */
+nsresult nsExpatTokenizer::QueryInterface(const nsIID& aIID, void** aInstancePtr)  
+{                                                                        
+  if (NULL == aInstancePtr) {                                            
+    return NS_ERROR_NULL_POINTER;                                        
+  }                                                                      
+
+  if(aIID.Equals(kISupportsIID))    {  //do IUnknown...
+    *aInstancePtr = (nsExpatTokenizer*)(this);                                        
+  }
+  else if(aIID.Equals(kITokenizerIID)) {  //do ITOkenizer base class...
+    *aInstancePtr = (nsITokenizer*)(this);                                        
+  }
+  else if(aIID.Equals(kHTMLTokenizerIID)) {  //do nsHTMLTokenizer base class...
+    *aInstancePtr = (nsHTMLTokenizer*)(this);                                        
+  }
+  else if(aIID.Equals(kClassIID)) {  //do this class...
+    *aInstancePtr = (nsExpatTokenizer*)(this);                                        
+  }                 
+  else {
+    *aInstancePtr=0;
+    return NS_NOINTERFACE;
+  }
+  NS_ADDREF_THIS();
+  return NS_OK;                                                        
+}
+
+
+/**
+ *  This method is defined in nsIParser. It is used to 
+ *  cause the COM-like construction of an nsParser.
+ *  
+ *  @update  gess 4/8/98
+ *  @param   nsIParser** ptr to newly instantiated parser
+ *  @return  NS_xxx error result
+ */
+NS_HTMLPARS nsresult NS_New_Expat_Tokenizer(nsIDTD** aInstancePtrResult) {
+  nsExpatTokenizer* it = new nsExpatTokenizer();
+  if (it == 0) {
+    return NS_ERROR_OUT_OF_MEMORY;
+  }
+  return it->QueryInterface(kClassIID, (void **) aInstancePtrResult);
+}
+
+
+NS_IMPL_ADDREF(nsExpatTokenizer)
+NS_IMPL_RELEASE(nsExpatTokenizer)
+
+
+/**
+ *  Default constructor
+ *   
+ *  @update  gess 4/9/98
+ *  @param   
+ *  @return  
+ */
+nsExpatTokenizer::nsExpatTokenizer() : nsHTMLTokenizer() {
+  NS_INIT_REFCNT();
+}
+
+/**
+ *  Default constructor
+ *   
+ *  @update  gess 4/9/98
+ *  @param   
+ *  @return  
+ */
+nsExpatTokenizer::~nsExpatTokenizer(){
+}
+
+
+/*******************************************************************
+  Here begins the real working methods for the tokenizer.
+ *******************************************************************/
+
+/**
+ *  This method repeatedly called by the tokenizer. 
+ *  Each time, we determine the kind of token were about to 
+ *  read, and then we call the appropriate method to handle
+ *  that token type.
+ *  
+ *  @update gess 3/25/98
+ *  @param  aChar: last char read
+ *  @param  aScanner: see nsScanner.h
+ *  @param  anErrorCode: arg that will hold error condition
+ *  @return new token or null 
+ */
+nsresult nsExpatTokenizer::ConsumeToken(nsScanner& aScanner) {
+  return nsHTMLTokenizer::ConsumeToken(aScanner);
+}
+

htmlparser/src/nsExpatTokenizer.h

@@ -0,0 +1,59 @@
+/* -*- Mode: C++; tab-width: 2; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
+/*
+ * The contents of this file are subject to the Netscape Public License
+ * Version 1.0 (the "NPL"); you may not use this file except in
+ * compliance with the NPL.  You may obtain a copy of the NPL at
+ * http://www.mozilla.org/NPL/
+ *
+ * Software distributed under the NPL is distributed on an "AS IS" basis,
+ * WITHOUT WARRANTY OF ANY KIND, either express or implied. See the NPL
+ * for the specific language governing rights and limitations under the
+ * NPL.
+ *
+ * The Initial Developer of this code under the NPL is Netscape
+ * Communications Corporation.  Portions created by Netscape are
+ * Copyright (C) 1998 Netscape Communications Corporation.  All Rights
+ * Reserved.
+ */
+
+
+/**
+ * MODULE NOTES:
+ * @update  gess 4/1/98
+ * 
+ */
+
+#ifndef __nsExpatTokenizer
+#define __nsExpatTokenizer
+
+#include "nsISupports.h"
+#include "nsHTMLTokenizer.h"
+#include "prtypes.h"
+
+#define NS_EXPATTOKENIZER_IID      \
+  {0x483836aa, 0xcabe, 0x11d2, { 0xab, 0xcb, 0x0, 0x10, 0x4b, 0x98, 0x3f, 0xd4 }}
+
+
+/***************************************************************
+  Notes: 
+ ***************************************************************/
+
+#if defined(XP_PC)
+#pragma warning( disable : 4275 )
+#endif
+
+CLASS_EXPORT_HTMLPARS nsExpatTokenizer : public nsHTMLTokenizer {
+public:
+            nsExpatTokenizer();
+  virtual   ~nsExpatTokenizer();
+
+            NS_DECL_ISUPPORTS
+
+  virtual nsresult ConsumeToken(nsScanner& aScanner);
+};
+
+extern NS_HTMLPARS nsresult NS_Expat_Tokenizer(nsIDTD** aInstancePtrResult);
+
+#endif
+
+

parser/htmlparser/src/nsExpatDTD.cpp

@@ -0,0 +1,470 @@
+/* -*- Mode: C++; tab-width: 2; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
+/*
+ * The contents of this file are subject to the Netscape Public License
+ * Version 1.0 (the "NPL"); you may not use this file except in
+ * compliance with the NPL.  You may obtain a copy of the NPL at
+ * http://www.mozilla.org/NPL/
+ *
+ * Software distributed under the NPL is distributed on an "AS IS" basis,
+ * WITHOUT WARRANTY OF ANY KIND, either express or implied. See the NPL
+ * for the specific language governing rights and limitations under the
+ * NPL.
+ *
+ * The Initial Developer of this code under the NPL is Netscape
+ * Communications Corporation.  Portions created by Netscape are
+ * Copyright (C) 1998 Netscape Communications Corporation.  All Rights
+ * Reserved.
+ */
+
+/**
+ * MODULE NOTES:
+ * @update  gess 4/8/98
+ * 
+ *         
+ */
+
+
+#include "nsIDTDDebug.h"
+#include "nsExpatDTD.h"
+#include "nsCRT.h"
+#include "nsParser.h"
+#include "nsScanner.h"
+#include "nsIParser.h"
+#include "nsTokenHandler.h"
+#include "nsDTDUtils.h"
+#include "nsIContentSink.h"
+#include "nsIHTMLContentSink.h"
+#include "nsExpatTokenizer.h"
+
+#include "prenv.h"  //this is here for debug reasons...
+#include "prtypes.h"  //this is here for debug reasons...
+#include "prio.h"
+#include "plstr.h"
+
+#ifdef XP_PC
+#include <direct.h> //this is here for debug reasons...
+#endif
+#include "prmem.h"
+
+
+static NS_DEFINE_IID(kISupportsIID, NS_ISUPPORTS_IID);                 
+static NS_DEFINE_IID(kIDTDIID,      NS_IDTD_IID);
+static NS_DEFINE_IID(kClassIID,     NS_EXPAT_DTD_IID); 
+
+static const char* kViewSourceCommand= "view-source";
+
+
+static CTokenRecycler gTokenRecycler;
+
+
+/**
+ *  This method gets called as part of our COM-like interfaces.
+ *  Its purpose is to create an interface to parser object
+ *  of some type.
+ *  
+ *  @update   gess 4/8/98
+ *  @param    nsIID  id of object to discover
+ *  @param    aInstancePtr ptr to newly discovered interface
+ *  @return   NS_xxx result code
+ */
+nsresult nsExpatDTD::QueryInterface(const nsIID& aIID, void** aInstancePtr)  
+{                                                                        
+  if (NULL == aInstancePtr) {                                            
+    return NS_ERROR_NULL_POINTER;                                        
+  }                                                                      
+
+  if(aIID.Equals(kISupportsIID))    {  //do IUnknown...
+    *aInstancePtr = (nsIDTD*)(this);                                        
+  }
+  else if(aIID.Equals(kIDTDIID)) {  //do IParser base class...
+    *aInstancePtr = (nsIDTD*)(this);                                        
+  }
+  else if(aIID.Equals(kClassIID)) {  //do this class...
+    *aInstancePtr = (nsExpatDTD*)(this);                                        
+  }                 
+  else {
+    *aInstancePtr=0;
+    return NS_NOINTERFACE;
+  }
+  NS_ADDREF_THIS();
+  return NS_OK;                                                        
+}
+
+/**
+ *  This method is defined in nsIParser. It is used to 
+ *  cause the COM-like construction of an nsParser.
+ *  
+ *  @update  gess 4/8/98
+ *  @param   nsIParser** ptr to newly instantiated parser
+ *  @return  NS_xxx error result
+ */
+NS_HTMLPARS nsresult NS_New_Expat_DTD(nsIDTD** aInstancePtrResult) {
+  nsExpatDTD* it = new nsExpatDTD();
+  if (it == 0) {
+    return NS_ERROR_OUT_OF_MEMORY;
+  }
+  return it->QueryInterface(kClassIID, (void **) aInstancePtrResult);
+}
+
+
+NS_IMPL_ADDREF(nsExpatDTD)
+NS_IMPL_RELEASE(nsExpatDTD)
+
+
+/**
+ *  Default constructor
+ *  
+ *  @update  gess 4/9/98
+ *  @param   
+ *  @return  
+ */
+nsExpatDTD::nsExpatDTD() : nsIDTD() {
+  NS_INIT_REFCNT();
+  mParser=0;
+  mSink=0;
+  mFilename;
+  mLineNumber=0;
+  mTokenizer=0;
+}
+
+/**
+ *  Default destructor
+ *  
+ *  @update  gess 4/9/98
+ *  @param   
+ *  @return  
+ */
+nsExpatDTD::~nsExpatDTD(){
+  mParser=0; //just to prove we destructed...
+  if(mTokenizer)
+    delete (nsExpatTokenizer*)mTokenizer;
+}
+
+/**
+ * 
+ * @update	gess1/8/99
+ * @param 
+ * @return
+ */
+const nsIID& nsExpatDTD::GetMostDerivedIID(void) const{
+  return kClassIID;
+}
+
+/**
+ * Call this method if you want the DTD to construct a fresh 
+ * instance of itself. 
+ * @update	gess7/23/98
+ * @param 
+ * @return
+ */
+nsresult nsExpatDTD::CreateNewInstance(nsIDTD** aInstancePtrResult){
+  return NS_New_Expat_DTD(aInstancePtrResult);
+}
+
+/**
+ * This method is called to determine if the given DTD can parse
+ * a document in a given source-type. 
+ * NOTE: Parsing always assumes that the end result will involve
+ *       storing the result in the main content model.
+ * @update	gess6/24/98
+ * @param   
+ * @return  TRUE if this DTD can satisfy the request; FALSE otherwise.
+ */
+eAutoDetectResult nsExpatDTD::CanParse(nsString& aContentType, nsString& aCommand, nsString& aBuffer, PRInt32 aVersion) {
+  eAutoDetectResult result=eUnknownDetect;
+
+  if(!aCommand.Equals(kViewSourceCommand)) {
+    if(aContentType.Equals(kXMLTextContentType) ||
+       aContentType.Equals(kRDFTextContentType) ||
+       aContentType.Equals(kXULTextContentType)) {
+      result=eValidDetect;
+    }
+    else {
+      if(-1<aBuffer.Find("<?xml ")) {
+        if(0==aContentType.Length()) {
+          aContentType = kXMLTextContentType; //only reset it if it's empty
+        }
+        result=eValidDetect;
+      }
+    }
+  }
+  return result;
+}
+
+
+/**
+ * 
+ * @update	gess5/18/98
+ * @param 
+ * @return
+ */
+NS_IMETHODIMP nsExpatDTD::WillBuildModel(nsString& aFilename,PRBool aNotifySink,nsIParser* aParser,nsIContentSink* aSink){
+  nsresult result=NS_OK;
+  mFilename=aFilename;
+
+  if(aParser){
+    mSink=aParser->GetContentSink();
+    if((aNotifySink) && (mSink)) {
+      mLineNumber=0;
+      result = mSink->WillBuildModel();
+    }
+  }
+
+  return result;
+}
+
+/**
+  * The parser uses a code sandwich to wrap the parsing process. Before
+  * the process begins, WillBuildModel() is called. Afterwards the parser
+  * calls DidBuildModel(). 
+  * @update	gess  1/4/99
+  * @param	aFilename is the name of the file being parsed.
+  * @return	error code (almost always 0)
+  */
+NS_IMETHODIMP nsExpatDTD::BuildModel(nsIParser* aParser,nsITokenizer* aTokenizer,nsITokenObserver* anObserver,nsIContentSink* aSink) {
+  nsresult result=NS_OK;
+
+  if(aTokenizer) {
+    nsITokenizer*  oldTokenizer=mTokenizer;
+    mTokenizer=aTokenizer;
+    nsITokenRecycler* theRecycler=aTokenizer->GetTokenRecycler();
+
+    while(NS_OK==result){
+      CToken* theToken=mTokenizer->PopToken();
+      if(theToken) {
+        result=HandleToken(theToken,aParser);
+        if(NS_SUCCEEDED(result)) {
+          theRecycler->RecycleToken(theToken);
+        }
+        else if(NS_ERROR_HTMLPARSER_BLOCK!=result){
+          mTokenizer->PushTokenFront(theToken);
+        }
+        // theRootDTD->Verify(kEmptyString,aParser);
+      }
+      else break;
+    }//while
+    mTokenizer=oldTokenizer;
+  }
+  else result=NS_ERROR_HTMLPARSER_BADTOKENIZER;
+  return result;
+}
+
+
+/**
+ * 
+ * @update	gess5/18/98
+ * @param 
+ * @return
+ */
+NS_IMETHODIMP nsExpatDTD::DidBuildModel(nsresult anErrorCode,PRBool aNotifySink,nsIParser* aParser,nsIContentSink* aSink){
+  nsresult result= NS_OK;
+
+  //ADD CODE HERE TO CLOSE OPEN CONTAINERS...
+
+  if(aParser){
+    mSink=aParser->GetContentSink();
+    if((aNotifySink) && (mSink)) {
+        result = mSink->DidBuildModel(1);
+    }
+  }
+  return result;
+}
+
+/**
+ * 
+ * @update	gess8/4/98
+ * @param 
+ * @return
+ */
+nsITokenRecycler* nsExpatDTD::GetTokenRecycler(void){
+  nsITokenizer* theTokenizer=GetTokenizer();
+  return theTokenizer->GetTokenRecycler();
+}
+
+/**
+ * Retrieve the preferred tokenizer for use by this DTD.
+ * @update	gess12/28/98
+ * @param   none
+ * @return  ptr to tokenizer
+ */
+nsITokenizer* nsExpatDTD::GetTokenizer(void) {
+  if(!mTokenizer)
+    mTokenizer=new nsExpatTokenizer();
+  return mTokenizer;
+}
+
+/**
+ * 
+ * @update	gess5/18/98
+ * @param 
+ * @return
+ */
+NS_IMETHODIMP nsExpatDTD::WillResumeParse(void){
+  nsresult result = NS_OK;
+  if(mSink) {
+    result = mSink->WillResume();
+  }
+  return result;
+}
+
+/**
+ * 
+ * @update	gess5/18/98
+ * @param 
+ * @return
+ */
+NS_IMETHODIMP nsExpatDTD::WillInterruptParse(void){
+  nsresult result = NS_OK;
+  if(mSink) {
+    result = mSink->WillInterrupt();
+  }
+  return result;
+}
+
+/**
+ * Called by the parser to initiate dtd verification of the
+ * internal context stack.
+ * @update	gess 7/23/98
+ * @param 
+ * @return
+ */
+PRBool nsExpatDTD::Verify(nsString& aURLRef,nsIParser* aParser) {
+  PRBool result=PR_TRUE;
+  mParser=(nsParser*)aParser;
+  return result;
+}
+
+/**
+ * Called by the parser to enable/disable dtd verification of the
+ * internal context stack.
+ * @update	gess 7/23/98
+ * @param 
+ * @return
+ */
+void nsExpatDTD::SetVerification(PRBool aEnabled){
+}
+
+/**
+ *  This method is called to determine whether or not a tag
+ *  of one type can contain a tag of another type.
+ *  
+ *  @update  gess 3/25/98
+ *  @param   aParent -- int tag of parent container
+ *  @param   aChild -- int tag of child container
+ *  @return  PR_TRUE if parent can contain child
+ */
+PRBool nsExpatDTD::CanContain(PRInt32 aParent,PRInt32 aChild) const{
+  PRBool result=PR_TRUE;
+  return result;
+}
+
+/**
+ *  This method gets called to determine whether a given 
+ *  tag is itself a container
+ *  
+ *  @update  gess 3/25/98
+ *  @param   aTag -- tag to test for containership
+ *  @return  PR_TRUE if given tag can contain other tags
+ */
+PRBool nsExpatDTD::IsContainer(PRInt32 aTag) const{
+  PRBool result=PR_TRUE;
+  return result;
+}
+
+/**
+ *  
+ *  @update  vidur 11/12/98
+ *  @param   aToken -- token object to be put into content model
+ *  @return  0 if all is well; non-zero is an error
+ */
+NS_IMETHODIMP nsExpatDTD::HandleToken(CToken* aToken,nsIParser* aParser) {
+  nsresult        result=NS_OK;
+  CHTMLToken*     theToken= (CHTMLToken*)(aToken);
+  eHTMLTokenTypes theType= (eHTMLTokenTypes)theToken->GetTokenType();
+
+  mParser=(nsParser*)aParser;
+  mSink=aParser->GetContentSink();
+
+  nsCParserNode theNode(theToken,mLineNumber);
+  switch(theType) {
+
+    case eToken_newline:
+      mLineNumber++; //now fall through
+    case eToken_entity:
+    case eToken_whitespace:
+    case eToken_text:
+    case eToken_cdatasection:
+      result=mSink->AddLeaf(theNode); 
+      break;
+
+    case eToken_comment:
+      result=mSink->AddComment(theNode); 
+      break;
+    
+    case eToken_instruction:
+      result=mSink->AddProcessingInstruction(theNode); 
+      break;
+
+    case eToken_start:
+      {
+        PRInt16 attrCount=aToken->GetAttributeCount();
+
+        if(0<attrCount){ //go collect the attributes...
+          int attr=0;
+          for(attr=0;attr<attrCount;attr++){
+            CToken* theToken=mTokenizer->PeekToken();
+            if(theToken)  {
+              eHTMLTokenTypes theType=eHTMLTokenTypes(theToken->GetTokenType());
+              if(eToken_attribute==theType){
+                mTokenizer->PopToken(); //pop it for real...
+                theNode.AddAttribute(theToken);
+              } 
+            }
+            else return kEOF;
+          }
+        }
+        if(NS_OK==result){
+          result=mSink->OpenContainer(theNode); 
+          if(((CStartToken*)aToken)->IsEmpty()){
+            result=mSink->CloseContainer(theNode); 
+          }
+        }
+      }
+      break;
+
+    case eToken_end:
+      result=mSink->CloseContainer(theNode); 
+      break;
+
+    case eToken_style:
+    case eToken_skippedcontent:
+    default:
+      result=NS_OK;
+  }//switch
+  return result;
+}
+
+
+/**
+ *  This method causes all tokens to be dispatched to the given tag handler.
+ *
+ *  @update  gess 3/25/98
+ *  @param   aHandler -- object to receive subsequent tokens...
+ *  @return	 error code (usually 0)
+ */
+nsresult nsExpatDTD::CaptureTokenPump(nsITagHandler* aHandler) {
+  nsresult result=NS_OK;
+  return result;
+}
+
+/**
+ *  This method releases the token-pump capture obtained in CaptureTokenPump()
+ *
+ *  @update  gess 3/25/98
+ *  @param   aHandler -- object that received tokens...
+ *  @return	 error code (usually 0)
+ */
+nsresult nsExpatDTD::ReleaseTokenPump(nsITagHandler* aHandler){
+  nsresult result=NS_OK;
+  return result;
+}

parser/htmlparser/src/nsExpatDTD.h

@@ -0,0 +1,238 @@
+
+/* -*- Mode: C++; tab-width: 2; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
+/*
+ * The contents of this file are subject to the Netscape Public License
+ * Version 1.0 (the "NPL"); you may not use this file except in
+ * compliance with the NPL.  You may obtain a copy of the NPL at
+ * http://www.mozilla.org/NPL/
+ *
+ * Software distributed under the NPL is distributed on an "AS IS" basis,
+ * WITHOUT WARRANTY OF ANY KIND, either express or implied. See the NPL
+ * for the specific language governing rights and limitations under the
+ * NPL.
+ *
+ * The Initial Developer of this code under the NPL is Netscape
+ * Communications Corporation.  Portions created by Netscape are
+ * Copyright (C) 1998 Netscape Communications Corporation.  All Rights
+ * Reserved.
+ */
+
+/**
+ * MODULE NOTES:
+ * @update  gess 4/8/98
+ * 
+ *         
+ */
+
+#ifndef __NS_EXPAT_DTD
+#define __NS_EXPAT_DTD
+
+#include "nsIDTD.h"
+#include "nsISupports.h"
+#include "nsHTMLTokens.h"
+#include "nshtmlpars.h"
+#include "nsVoidArray.h"
+#include "nsDeque.h"
+#include "nsIContentSink.h"
+
+
+#define NS_EXPAT_DTD_IID      \
+  {0x5ad4b431, 0xcabb, 0x11d2, \
+  {0xab, 0xcb, 0x0, 0x10, 0x4b, 0x98, 0x3f, 0xd4}}
+
+
+class nsIDTDDebug;
+class nsIParserNode;
+class CITokenHandler;
+class nsParser;
+
+
+
+class nsExpatDTD : public nsIDTD {
+            
+  public:
+
+    NS_DECL_ISUPPORTS
+
+
+    /**
+     *  
+     *  
+     *  @update  gess 4/9/98
+     *  @param   
+     *  @return  
+     */
+    nsExpatDTD();
+
+    /**
+     *  
+     *  
+     *  @update  gess 4/9/98
+     *  @param   
+     *  @return  
+     */
+    virtual ~nsExpatDTD();
+
+    virtual const nsIID&  GetMostDerivedIID(void) const;
+
+    /**
+     * Call this method if you want the DTD to construct a clone of itself.
+     * @update	gess7/23/98
+     * @param 
+     * @return
+     */
+    virtual nsresult CreateNewInstance(nsIDTD** aInstancePtrResult);
+
+    /**
+     * This method is called to determine if the given DTD can parse
+     * a document in a given source-type. 
+     * NOTE: Parsing always assumes that the end result will involve
+     *       storing the result in the main content model.
+     * @update	gess6/24/98
+     * @param   
+     * @return  TRUE if this DTD can satisfy the request; FALSE otherwise.
+     */
+    virtual eAutoDetectResult CanParse(nsString& aContentType, nsString& aCommand, nsString& aBuffer, PRInt32 aVersion);
+
+    /**
+      * The parser uses a code sandwich to wrap the parsing process. Before
+      * the process begins, WillBuildModel() is called. Afterwards the parser
+      * calls DidBuildModel(). 
+      * @update	gess5/18/98
+      * @param	aFilename is the name of the file being parsed.
+      * @return	error code (almost always 0)
+      */
+    NS_IMETHOD WillBuildModel(nsString& aFilename,PRBool aNotifySink,nsIParser* aParser,nsIContentSink* aSink=0);
+
+    /**
+      * The parser uses a code sandwich to wrap the parsing process. Before
+      * the process begins, WillBuildModel() is called. Afterwards the parser
+      * calls DidBuildModel(). 
+      * @update	gess5/18/98
+      * @param	aFilename is the name of the file being parsed.
+      * @return	error code (almost always 0)
+      */
+    NS_IMETHOD BuildModel(nsIParser* aParser,nsITokenizer* aTokenizer,nsITokenObserver* anObserver=0,nsIContentSink* aSink=0);
+
+   /**
+     * The parser uses a code sandwich to wrap the parsing process. Before
+     * the process begins, WillBuildModel() is called. Afterwards the parser
+     * calls DidBuildModel(). 
+     * @update	gess5/18/98
+     * @param	anErrorCode contans the last error that occured
+     * @return	error code
+     */
+    NS_IMETHOD DidBuildModel(nsresult anErrorCode,PRBool aNotifySink,nsIParser* aParser,nsIContentSink* aSink=0);
+
+    /**
+     *  
+     *  @update  gess 3/25/98
+     *  @param   aToken -- token object to be put into content model
+     *  @return  0 if all is well; non-zero is an error
+     */
+    NS_IMETHOD HandleToken(CToken* aToken,nsIParser* aParser);
+
+    /**
+     *  This method causes all tokens to be dispatched to the given tag handler.
+     *
+     *  @update  gess 3/25/98
+  	 *  @param   aHandler -- object to receive subsequent tokens...
+	   *  @return	 error code (usually 0)
+     */
+    NS_IMETHOD CaptureTokenPump(nsITagHandler* aHandler);
+
+    /**
+     *  This method releases the token-pump capture obtained in CaptureTokenPump()
+     *
+     *  @update  gess 3/25/98
+  	 *  @param   aHandler -- object that received tokens...
+	   *  @return	 error code (usually 0)
+     */
+    NS_IMETHOD ReleaseTokenPump(nsITagHandler* aHandler);
+
+    /**
+     * 
+     * @update	gess12/28/98
+     * @param 
+     * @return
+     */
+    nsITokenizer* GetTokenizer(void);
+
+    /**
+     * 
+     * @update	gess5/18/98
+     * @param 
+     * @return
+     */
+    NS_IMETHOD WillResumeParse(void);
+
+    /**
+     * 
+     * @update	gess5/18/98
+     * @param 
+     * @return
+     */
+    NS_IMETHOD WillInterruptParse(void);
+
+    /**
+     * Called by the parser to initiate dtd verification of the
+     * internal context stack.
+     * @update	gess 7/23/98
+     * @param 
+     * @return
+     */
+    virtual PRBool Verify(nsString& aURLRef,nsIParser* aParser);
+
+    /**
+     * Set this to TRUE if you want the DTD to verify its
+     * context stack.
+     * @update	gess 7/23/98
+     * @param 
+     * @return
+     */
+    virtual void SetVerification(PRBool aEnable);
+
+    /**
+     *  This method is called to determine whether or not a tag
+     *  of one type can contain a tag of another type.
+     *  
+     *  @update  gess 3/25/98
+     *  @param   aParent -- int tag of parent container
+     *  @param   aChild -- int tag of child container
+     *  @return  PR_TRUE if parent can contain child
+     */
+    virtual PRBool CanContain(PRInt32 aParent,PRInt32 aChild) const;
+
+    /**
+     *  This method gets called to determine whether a given 
+     *  tag is itself a container
+     *  
+     *  @update  gess 3/25/98
+     *  @param   aTag -- tag to test for containership
+     *  @return  PR_TRUE if given tag can contain other tags
+     */
+    virtual PRBool IsContainer(PRInt32 aTag) const;
+
+    /**
+     * Retrieve a ptr to the global token recycler...
+     * @update	gess8/4/98
+     * @return  ptr to recycler (or null)
+     */
+    virtual nsITokenRecycler* GetTokenRecycler(void);
+    
+protected:
+  
+    nsParser*           mParser;
+    nsIContentSink*     mSink;
+    nsString            mFilename;
+    PRInt32             mLineNumber;
+    nsITokenizer*       mTokenizer;
+};
+
+extern NS_HTMLPARS nsresult NS_New_Expat_DTD(nsIDTD** aInstancePtrResult);
+
+
+#endif 
+
+
+

parser/htmlparser/src/nsExpatTokenizer.cpp

@@ -0,0 +1,141 @@
+/* -*- Mode: C++; tab-width: 2; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
+/* 
+ * The contents of this file are subject to the Netscape Public License
+ * Version 1.0 (the "NPL"); you may not use this file except in
+ * compliance with the NPL.  You may obtain a copy of the NPL at
+ * http://www.mozilla.org/NPL/
+ *
+ * Software distributed under the NPL is distributed on an "AS IS" basis,
+ * WITHOUT WARRANTY OF ANY KIND, either express or implied. See the NPL
+ * for the specific language governing rights and limitations under the
+ * NPL.
+ *
+ * The Initial Developer of this code under the NPL is Netscape
+ * Communications Corporation.  Portions created by Netscape are
+ * Copyright (C) 1998 Netscape Communications Corporation.  All Rights
+ * Reserved.
+ */
+
+
+/**
+ * MODULE NOTES:
+ * @update  gess 4/1/98
+ * 
+ */
+
+#include "nsExpatTokenizer.h"
+#include "nsParser.h"
+#include "nsScanner.h"
+#include "nsDTDUtils.h"
+#include "nsParser.h"
+
+/************************************************************************
+  And now for the main class -- nsExpatTokenizer...
+ ************************************************************************/
+
+static NS_DEFINE_IID(kISupportsIID,     NS_ISUPPORTS_IID);                 
+static NS_DEFINE_IID(kITokenizerIID,    NS_ITOKENIZER_IID);
+static NS_DEFINE_IID(kHTMLTokenizerIID, NS_HTMLTOKENIZER_IID);
+static NS_DEFINE_IID(kClassIID,         NS_EXPATTOKENIZER_IID); 
+static nsAutoString gEmpty;
+
+/**
+ *  This method gets called as part of our COM-like interfaces.
+ *  Its purpose is to create an interface to parser object
+ *  of some type.
+ *  
+ *  @update   gess 4/8/98
+ *  @param    nsIID  id of object to discover
+ *  @param    aInstancePtr ptr to newly discovered interface
+ *  @return   NS_xxx result code
+ */
+nsresult nsExpatTokenizer::QueryInterface(const nsIID& aIID, void** aInstancePtr)  
+{                                                                        
+  if (NULL == aInstancePtr) {                                            
+    return NS_ERROR_NULL_POINTER;                                        
+  }                                                                      
+
+  if(aIID.Equals(kISupportsIID))    {  //do IUnknown...
+    *aInstancePtr = (nsExpatTokenizer*)(this);                                        
+  }
+  else if(aIID.Equals(kITokenizerIID)) {  //do ITOkenizer base class...
+    *aInstancePtr = (nsITokenizer*)(this);                                        
+  }
+  else if(aIID.Equals(kHTMLTokenizerIID)) {  //do nsHTMLTokenizer base class...
+    *aInstancePtr = (nsHTMLTokenizer*)(this);                                        
+  }
+  else if(aIID.Equals(kClassIID)) {  //do this class...
+    *aInstancePtr = (nsExpatTokenizer*)(this);                                        
+  }                 
+  else {
+    *aInstancePtr=0;
+    return NS_NOINTERFACE;
+  }
+  NS_ADDREF_THIS();
+  return NS_OK;                                                        
+}
+
+
+/**
+ *  This method is defined in nsIParser. It is used to 
+ *  cause the COM-like construction of an nsParser.
+ *  
+ *  @update  gess 4/8/98
+ *  @param   nsIParser** ptr to newly instantiated parser
+ *  @return  NS_xxx error result
+ */
+NS_HTMLPARS nsresult NS_New_Expat_Tokenizer(nsIDTD** aInstancePtrResult) {
+  nsExpatTokenizer* it = new nsExpatTokenizer();
+  if (it == 0) {
+    return NS_ERROR_OUT_OF_MEMORY;
+  }
+  return it->QueryInterface(kClassIID, (void **) aInstancePtrResult);
+}
+
+
+NS_IMPL_ADDREF(nsExpatTokenizer)
+NS_IMPL_RELEASE(nsExpatTokenizer)
+
+
+/**
+ *  Default constructor
+ *   
+ *  @update  gess 4/9/98
+ *  @param   
+ *  @return  
+ */
+nsExpatTokenizer::nsExpatTokenizer() : nsHTMLTokenizer() {
+  NS_INIT_REFCNT();
+}
+
+/**
+ *  Default constructor
+ *   
+ *  @update  gess 4/9/98
+ *  @param   
+ *  @return  
+ */
+nsExpatTokenizer::~nsExpatTokenizer(){
+}
+
+
+/*******************************************************************
+  Here begins the real working methods for the tokenizer.
+ *******************************************************************/
+
+/**
+ *  This method repeatedly called by the tokenizer. 
+ *  Each time, we determine the kind of token were about to 
+ *  read, and then we call the appropriate method to handle
+ *  that token type.
+ *  
+ *  @update gess 3/25/98
+ *  @param  aChar: last char read
+ *  @param  aScanner: see nsScanner.h
+ *  @param  anErrorCode: arg that will hold error condition
+ *  @return new token or null 
+ */
+nsresult nsExpatTokenizer::ConsumeToken(nsScanner& aScanner) {
+  return nsHTMLTokenizer::ConsumeToken(aScanner);
+}
+

parser/htmlparser/src/nsExpatTokenizer.h

@@ -0,0 +1,59 @@
+/* -*- Mode: C++; tab-width: 2; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
+/*
+ * The contents of this file are subject to the Netscape Public License
+ * Version 1.0 (the "NPL"); you may not use this file except in
+ * compliance with the NPL.  You may obtain a copy of the NPL at
+ * http://www.mozilla.org/NPL/
+ *
+ * Software distributed under the NPL is distributed on an "AS IS" basis,
+ * WITHOUT WARRANTY OF ANY KIND, either express or implied. See the NPL
+ * for the specific language governing rights and limitations under the
+ * NPL.
+ *
+ * The Initial Developer of this code under the NPL is Netscape
+ * Communications Corporation.  Portions created by Netscape are
+ * Copyright (C) 1998 Netscape Communications Corporation.  All Rights
+ * Reserved.
+ */
+
+
+/**
+ * MODULE NOTES:
+ * @update  gess 4/1/98
+ * 
+ */
+
+#ifndef __nsExpatTokenizer
+#define __nsExpatTokenizer
+
+#include "nsISupports.h"
+#include "nsHTMLTokenizer.h"
+#include "prtypes.h"
+
+#define NS_EXPATTOKENIZER_IID      \
+  {0x483836aa, 0xcabe, 0x11d2, { 0xab, 0xcb, 0x0, 0x10, 0x4b, 0x98, 0x3f, 0xd4 }}
+
+
+/***************************************************************
+  Notes: 
+ ***************************************************************/
+
+#if defined(XP_PC)
+#pragma warning( disable : 4275 )
+#endif
+
+CLASS_EXPORT_HTMLPARS nsExpatTokenizer : public nsHTMLTokenizer {
+public:
+            nsExpatTokenizer();
+  virtual   ~nsExpatTokenizer();
+
+            NS_DECL_ISUPPORTS
+
+  virtual nsresult ConsumeToken(nsScanner& aScanner);
+};
+
+extern NS_HTMLPARS nsresult NS_Expat_Tokenizer(nsIDTD** aInstancePtrResult);
+
+#endif
+
+