From 7d4586a4d1bcf1f574cb3634105f501ba4e8ead0 Mon Sep 17 00:00:00 2001
From: "rickg%netscape.com" <rickg%netscape.com>
Date: Thu, 8 Jul 1999 04:39:38 +0000
Subject: [PATCH] fixed bug 9386 and added Terminate() to parser API

---
 htmlparser/src/nsHTMLTokenizer.cpp        |  2 +-
 htmlparser/src/nsHTMLTokens.cpp           | 37 +++++++++--------------
 htmlparser/src/nsIParser.h                |  1 +
 htmlparser/src/nsParser.cpp               | 13 ++++++++
 htmlparser/src/nsParser.h                 |  3 +-
 parser/htmlparser/src/nsHTMLTokenizer.cpp |  2 +-
 parser/htmlparser/src/nsHTMLTokens.cpp    | 37 +++++++++--------------
 parser/htmlparser/src/nsIParser.h         |  1 +
 parser/htmlparser/src/nsParser.cpp        | 13 ++++++++
 parser/htmlparser/src/nsParser.h          |  3 +-
 10 files changed, 64 insertions(+), 48 deletions(-)

diff --git a/htmlparser/src/nsHTMLTokenizer.cpp b/htmlparser/src/nsHTMLTokenizer.cpp
index f8d9dd8cb5a..cf9ad14def9 100644
--- a/htmlparser/src/nsHTMLTokenizer.cpp
+++ b/htmlparser/src/nsHTMLTokenizer.cpp
@@ -462,9 +462,9 @@ nsresult nsHTMLTokenizer::ConsumeStartTag(PRUnichar aChar,CToken*& aToken,nsScan
         if((eHTMLTag_style==theTag) || (eHTMLTag_script==theTag)) {
         nsAutoString endTag(NS_EnumToTag(theTag));
         endTag.Insert("</",0,2);
-        endTag.Append(">");        
         CToken* textToken=theRecycler->CreateTokenOfType(eToken_text,theTag,endTag);
         result=((CTextToken*)textToken)->ConsumeUntil(0,PRBool(eHTMLTag_style==theTag),aScanner,endTag);  //tell new token to finish consuming text...    
+        //endTag.Append(">");        
         CToken* endToken=theRecycler->CreateTokenOfType(eToken_end,theTag,endTag);
         AddToken(textToken,result,mTokenDeque,theRecycler);
         AddToken(endToken,result,mTokenDeque,theRecycler);
diff --git a/htmlparser/src/nsHTMLTokens.cpp b/htmlparser/src/nsHTMLTokens.cpp
index 0225d4ade4f..671f61ae73e 100644
--- a/htmlparser/src/nsHTMLTokens.cpp
+++ b/htmlparser/src/nsHTMLTokens.cpp
@@ -507,11 +507,12 @@ nsresult CTextToken::Consume(PRUnichar aChar, nsScanner& aScanner) {
  *  @return  error result
  */
 nsresult CTextToken::ConsumeUntil(PRUnichar aChar,PRBool aIgnoreComments,nsScanner& aScanner,nsString& aTerminalString){
-  PRBool       done=PR_FALSE; 
-  nsresult     result=NS_OK; 
-  nsString     temp; 
-  PRUnichar    theChar;
-  nsAutoString theRight;
+  PRBool        done=PR_FALSE; 
+  nsresult      result=NS_OK; 
+  nsString      temp; 
+  PRUnichar     theChar;
+  nsAutoString  theRight;
+  PRInt32       rpos=0;
 
   //We're going to try a new algorithm here. Rather than scan for the matching 
  //end tag like we used to do, we're now going to scan for whitespace and comments. 
@@ -519,7 +520,7 @@ nsresult CTextToken::ConsumeUntil(PRUnichar aChar,PRBool aIgnoreComments,nsScann
  //target endtag, or the start of another comment. 
 
   static nsAutoString theWhitespace2("\b\t ");
-
+  PRInt32 termStrLen=aTerminalString.Length();
   while((!done) && (NS_OK==result)) { 
     result=aScanner.GetChar(aChar); 
     if((NS_OK==result) && (kLessThan==aChar)) { 
@@ -553,17 +554,20 @@ nsresult CTextToken::ConsumeUntil(PRUnichar aChar,PRBool aIgnoreComments,nsScann
       temp+=aChar; 
       result=aScanner.ReadUntil(temp,kLessThan,PR_FALSE); 
     } 
-    temp.Right(theRight,aTerminalString.Length());
-    done=PRBool(0==theRight.Compare(aTerminalString,PR_TRUE)); 
+    temp.Right(theRight,termStrLen+10); //first, get a wad of chars from the temp string
+    rpos=theRight.RFind('<');   //now scan for the '<'
+    if(-1<rpos)
+      rpos=theRight.RFind(aTerminalString,PR_TRUE,rpos);
+    done=PRBool(-1<rpos); 
   }  //while
   int len=temp.Length(); 
-  temp.Truncate(len-aTerminalString.Length()); 
+  temp.Truncate(len-(theRight.Length()-rpos)); 
   mTextValue=temp; 
 
   // Make aTerminalString contain the name of the end tag ** as seen in **
   // the document and not the made up one.
-  theRight.Cut(0,2);
-  theRight.Cut((theRight.Length()-1),1);
+  theRight.Cut(0,rpos+2);
+  theRight.Truncate(theRight.Length()-1);
   aTerminalString = theRight;
   return result; 
 }
@@ -854,17 +858,6 @@ nsresult ConsumeComment(PRUnichar aChar, nsScanner& aScanner,nsString& aString)
 nsresult CCommentToken::Consume(PRUnichar aChar, nsScanner& aScanner) {
   PRBool theStrictForm=PR_FALSE;
   nsresult result=(theStrictForm) ? ConsumeStrictComment(aChar,aScanner,mTextValue) : ConsumeComment(aChar,aScanner,mTextValue);
-
-/*
-  //this change is here to make the editor teams' life easier.
-  //I'm removing the leading and trailing markup...
-
-  if(0==mTextValue.Find("<!"))
-    mTextValue.Cut(0,2);  //trim off 1st 2 chars...
-  if(kGreaterThan==mTextValue.Last())
-    mTextValue.Truncate(mTextValue.Length()-1); //trim off last char
-*/
-
   return result;
 }
 
diff --git a/htmlparser/src/nsIParser.h b/htmlparser/src/nsIParser.h
index acd84c7f286..5c19a438b7d 100644
--- a/htmlparser/src/nsIParser.h
+++ b/htmlparser/src/nsIParser.h
@@ -170,6 +170,7 @@ class nsIParser : public nsISupports {
     virtual nsresult  Parse(nsIURI* aURL,nsIStreamObserver* aListener = nsnull,PRBool aEnableVerify=PR_FALSE, void* aKey=0) = 0;
     virtual nsresult	Parse(nsIInputStream& aStream, PRBool aEnableVerify=PR_FALSE, void* aKey=0) = 0;
     virtual nsresult  Parse(nsString& aSourceBuffer,void* aKey,const nsString& aContentType,PRBool aEnableVerify,PRBool aLastCall) = 0;
+    virtual nsresult  Terminate(void) = 0;
 
     virtual PRBool    IsValidFragment(const nsString& aSourceBuffer,nsITagStack& aStack,PRUint32 anInsertPos,const nsString& aContentType)=0;
     virtual nsresult  ParseFragment(const nsString& aSourceBuffer,void* aKey,nsITagStack& aStack,PRUint32 anInsertPos,const nsString& aContentType)=0;
diff --git a/htmlparser/src/nsParser.cpp b/htmlparser/src/nsParser.cpp
index a1c97857655..4cd2eef67f7 100644
--- a/htmlparser/src/nsParser.cpp
+++ b/htmlparser/src/nsParser.cpp
@@ -575,6 +575,19 @@ void nsParser::SetUnusedInput(nsString& aBuffer) {
   mUnusedInput=aBuffer;
 }
 
+/**
+ *  Call this when you want to *force* the parser to terminate the
+ *  parsing process altogether. This is binary -- so once you terminate
+ *  you can't resume without restarting altogether.
+ *  
+ *  @update  gess 7/4/99
+ *  @return  should return NS_OK once implemented
+ */
+nsresult nsParser::Terminate(void){
+  NS_NOTYETIMPLEMENTED("Call again later");
+  nsresult result=NS_ERROR_NOT_IMPLEMENTED;
+  return result;
+}
 
 /**
  *  Call this when you want control whether or not the parser will parse
diff --git a/htmlparser/src/nsParser.h b/htmlparser/src/nsParser.h
index 8b93b9b0245..504e9d4acb9 100644
--- a/htmlparser/src/nsParser.h
+++ b/htmlparser/src/nsParser.h
@@ -204,7 +204,8 @@ friend class CTokenHandler;
      *  @param   aState determines whether we parse/tokenize or just cache.
      *  @return  current state
      */
-    virtual PRBool EnableParser(PRBool aState);
+    virtual PRBool    EnableParser(PRBool aState);
+    virtual nsresult  Terminate(void);
 
     /**
      * Call this to query whether the parser is enabled or not.
diff --git a/parser/htmlparser/src/nsHTMLTokenizer.cpp b/parser/htmlparser/src/nsHTMLTokenizer.cpp
index f8d9dd8cb5a..cf9ad14def9 100644
--- a/parser/htmlparser/src/nsHTMLTokenizer.cpp
+++ b/parser/htmlparser/src/nsHTMLTokenizer.cpp
@@ -462,9 +462,9 @@ nsresult nsHTMLTokenizer::ConsumeStartTag(PRUnichar aChar,CToken*& aToken,nsScan
         if((eHTMLTag_style==theTag) || (eHTMLTag_script==theTag)) {
         nsAutoString endTag(NS_EnumToTag(theTag));
         endTag.Insert("</",0,2);
-        endTag.Append(">");        
         CToken* textToken=theRecycler->CreateTokenOfType(eToken_text,theTag,endTag);
         result=((CTextToken*)textToken)->ConsumeUntil(0,PRBool(eHTMLTag_style==theTag),aScanner,endTag);  //tell new token to finish consuming text...    
+        //endTag.Append(">");        
         CToken* endToken=theRecycler->CreateTokenOfType(eToken_end,theTag,endTag);
         AddToken(textToken,result,mTokenDeque,theRecycler);
         AddToken(endToken,result,mTokenDeque,theRecycler);
diff --git a/parser/htmlparser/src/nsHTMLTokens.cpp b/parser/htmlparser/src/nsHTMLTokens.cpp
index 0225d4ade4f..671f61ae73e 100644
--- a/parser/htmlparser/src/nsHTMLTokens.cpp
+++ b/parser/htmlparser/src/nsHTMLTokens.cpp
@@ -507,11 +507,12 @@ nsresult CTextToken::Consume(PRUnichar aChar, nsScanner& aScanner) {
  *  @return  error result
  */
 nsresult CTextToken::ConsumeUntil(PRUnichar aChar,PRBool aIgnoreComments,nsScanner& aScanner,nsString& aTerminalString){
-  PRBool       done=PR_FALSE; 
-  nsresult     result=NS_OK; 
-  nsString     temp; 
-  PRUnichar    theChar;
-  nsAutoString theRight;
+  PRBool        done=PR_FALSE; 
+  nsresult      result=NS_OK; 
+  nsString      temp; 
+  PRUnichar     theChar;
+  nsAutoString  theRight;
+  PRInt32       rpos=0;
 
   //We're going to try a new algorithm here. Rather than scan for the matching 
  //end tag like we used to do, we're now going to scan for whitespace and comments. 
@@ -519,7 +520,7 @@ nsresult CTextToken::ConsumeUntil(PRUnichar aChar,PRBool aIgnoreComments,nsScann
  //target endtag, or the start of another comment. 
 
   static nsAutoString theWhitespace2("\b\t ");
-
+  PRInt32 termStrLen=aTerminalString.Length();
   while((!done) && (NS_OK==result)) { 
     result=aScanner.GetChar(aChar); 
     if((NS_OK==result) && (kLessThan==aChar)) { 
@@ -553,17 +554,20 @@ nsresult CTextToken::ConsumeUntil(PRUnichar aChar,PRBool aIgnoreComments,nsScann
       temp+=aChar; 
       result=aScanner.ReadUntil(temp,kLessThan,PR_FALSE); 
     } 
-    temp.Right(theRight,aTerminalString.Length());
-    done=PRBool(0==theRight.Compare(aTerminalString,PR_TRUE)); 
+    temp.Right(theRight,termStrLen+10); //first, get a wad of chars from the temp string
+    rpos=theRight.RFind('<');   //now scan for the '<'
+    if(-1<rpos)
+      rpos=theRight.RFind(aTerminalString,PR_TRUE,rpos);
+    done=PRBool(-1<rpos); 
   }  //while
   int len=temp.Length(); 
-  temp.Truncate(len-aTerminalString.Length()); 
+  temp.Truncate(len-(theRight.Length()-rpos)); 
   mTextValue=temp; 
 
   // Make aTerminalString contain the name of the end tag ** as seen in **
   // the document and not the made up one.
-  theRight.Cut(0,2);
-  theRight.Cut((theRight.Length()-1),1);
+  theRight.Cut(0,rpos+2);
+  theRight.Truncate(theRight.Length()-1);
   aTerminalString = theRight;
   return result; 
 }
@@ -854,17 +858,6 @@ nsresult ConsumeComment(PRUnichar aChar, nsScanner& aScanner,nsString& aString)
 nsresult CCommentToken::Consume(PRUnichar aChar, nsScanner& aScanner) {
   PRBool theStrictForm=PR_FALSE;
   nsresult result=(theStrictForm) ? ConsumeStrictComment(aChar,aScanner,mTextValue) : ConsumeComment(aChar,aScanner,mTextValue);
-
-/*
-  //this change is here to make the editor teams' life easier.
-  //I'm removing the leading and trailing markup...
-
-  if(0==mTextValue.Find("<!"))
-    mTextValue.Cut(0,2);  //trim off 1st 2 chars...
-  if(kGreaterThan==mTextValue.Last())
-    mTextValue.Truncate(mTextValue.Length()-1); //trim off last char
-*/
-
   return result;
 }
 
diff --git a/parser/htmlparser/src/nsIParser.h b/parser/htmlparser/src/nsIParser.h
index acd84c7f286..5c19a438b7d 100644
--- a/parser/htmlparser/src/nsIParser.h
+++ b/parser/htmlparser/src/nsIParser.h
@@ -170,6 +170,7 @@ class nsIParser : public nsISupports {
     virtual nsresult  Parse(nsIURI* aURL,nsIStreamObserver* aListener = nsnull,PRBool aEnableVerify=PR_FALSE, void* aKey=0) = 0;
     virtual nsresult	Parse(nsIInputStream& aStream, PRBool aEnableVerify=PR_FALSE, void* aKey=0) = 0;
     virtual nsresult  Parse(nsString& aSourceBuffer,void* aKey,const nsString& aContentType,PRBool aEnableVerify,PRBool aLastCall) = 0;
+    virtual nsresult  Terminate(void) = 0;
 
     virtual PRBool    IsValidFragment(const nsString& aSourceBuffer,nsITagStack& aStack,PRUint32 anInsertPos,const nsString& aContentType)=0;
     virtual nsresult  ParseFragment(const nsString& aSourceBuffer,void* aKey,nsITagStack& aStack,PRUint32 anInsertPos,const nsString& aContentType)=0;
diff --git a/parser/htmlparser/src/nsParser.cpp b/parser/htmlparser/src/nsParser.cpp
index a1c97857655..4cd2eef67f7 100644
--- a/parser/htmlparser/src/nsParser.cpp
+++ b/parser/htmlparser/src/nsParser.cpp
@@ -575,6 +575,19 @@ void nsParser::SetUnusedInput(nsString& aBuffer) {
   mUnusedInput=aBuffer;
 }
 
+/**
+ *  Call this when you want to *force* the parser to terminate the
+ *  parsing process altogether. This is binary -- so once you terminate
+ *  you can't resume without restarting altogether.
+ *  
+ *  @update  gess 7/4/99
+ *  @return  should return NS_OK once implemented
+ */
+nsresult nsParser::Terminate(void){
+  NS_NOTYETIMPLEMENTED("Call again later");
+  nsresult result=NS_ERROR_NOT_IMPLEMENTED;
+  return result;
+}
 
 /**
  *  Call this when you want control whether or not the parser will parse
diff --git a/parser/htmlparser/src/nsParser.h b/parser/htmlparser/src/nsParser.h
index 8b93b9b0245..504e9d4acb9 100644
--- a/parser/htmlparser/src/nsParser.h
+++ b/parser/htmlparser/src/nsParser.h
@@ -204,7 +204,8 @@ friend class CTokenHandler;
      *  @param   aState determines whether we parse/tokenize or just cache.
      *  @return  current state
      */
-    virtual PRBool EnableParser(PRBool aState);
+    virtual PRBool    EnableParser(PRBool aState);
+    virtual nsresult  Terminate(void);
 
     /**
      * Call this to query whether the parser is enabled or not.