63104 - </script> shown as </</script> in view source

Status: VERIFIED FIXED

(Core :: DOM: HTML Parser, defect)

Description

[HJ]

All build's after 2000121104 are changing the closing tag <script> into someting
like this: </</script>

Build 2000121104 is ok on WinNT4 Sp6b later builds are making a mess of it!

Might this has enything to do with displaying text at rendering of documents?

Comment 1

[HJ]

Attachment: example for test purpose

Comment 2

[HJ]

I have inserted a <\/script> in this example. Don't care, that one is ok!

Also Netscape6/6.0 is having no problems with this! This might be the case that
text pops up on your screen! Is the parser getting drunk of this typo?

Comment 3

[Boris Zbarsky [:bzbarsky]]

I see this on linux CVS build 2000-12-15.  In the view source window, </script>
tags are shown as </</script>

All other closing tags are displayed correctly. OS->All

Comment 4

[HJ]

adding keyword testcase

Comment 5

[vidur (gone)]

Proposed fix (OK, it got checked in by mistake):

Index: nsHTMLTokenizer.cpp
===================================================================
RCS file: /cvsroot/mozilla/htmlparser/src/nsHTMLTokenizer.cpp,v
retrieving revision 3.73
retrieving revision 3.74
diff -u -r3.73 -r3.74
--- nsHTMLTokenizer.cpp 2000/12/18 20:44:09     3.73
+++ nsHTMLTokenizer.cpp 2000/12/18 20:58:53     3.74
@@ -717,12 +717,15 @@

         //if((eHTMLTag_style==theTag) || (eHTMLTag_script==theTag)) {
         if(gHTMLElements[theTag].CanContainType(kCDATA)) {
-          nsAutoString endTag; endTag.AssignWithConversion(nsHTMLTags::GetStrin
gValue(theTag));
-          endTag.InsertWithConversion("</",0,2);
+          nsAutoString endText, endTagName;
+          endTagName.AssignWithConversion(nsHTMLTags::GetStringValue(theTag));
+          endText.Assign(endTagName);
+          endText.InsertWithConversion("</",0,2);
+
           CToken* textToken=theAllocator->CreateTokenOfType(eToken_text,eHTMLTa
g_text);
-          result=((CTextToken*)textToken)->ConsumeUntil(0,PRBool(theTag!=eHTMLT
ag_script),aScanner,endTag,mParseMode,aFlushTokens);  //tell new token to finish
 consuming text...
-          //endTag.Append(">");
-          CToken* endToken=theAllocator->CreateTokenOfType(eToken_end,theTag,en
dTag);
+          result=((CTextToken*)textToken)->ConsumeUntil(0,PRBool(theTag!=eHTMLT
ag_script),aScanner,endText,mParseMode,aFlushTokens);  //tell new token to finis
h consuming text...
+
+          CToken* endToken=theAllocator->CreateTokenOfType(eToken_end,theTag,en
dTagName);
           AddToken(textToken,result,&mTokenDeque,theAllocator);
           AddToken(endToken,result,&mTokenDeque,theAllocator);
         }

Comment 6

[vidur (gone)]

Fix checked in on 12/18/2000.

Comment 7

[harishd]

Will have to teach ConsumeUntil() to not depend on '</' - I'll do that later.

But for now Vidur's patch looks good to me. r=harishd

Comment 8

[Boris Zbarsky [:bzbarsky]]

*** Bug 63226 has been marked as a duplicate of this bug. ***

Comment 9

[Jan Carpenter]

verified: 
2001-02-05-09-Mtrunk : Win32
2001-02-05-11-Mtrunk : Linux
2001-02-05-12-Mtrunk : Mac