Closed Bug 212903 Opened 21 years ago Closed 21 years ago

XSLT bug -- incorrectly added <meta> element to results tree

Categories

(Core :: XSLT, defect, P3)

Product:
Core
Component:
XSLT
Type:
defect

Tracking

()

Status:
RESOLVED FIXED
Milestone:
mozilla1.6alpha

People

(Reporter: ian.graham, Assigned: peterv)

Details

Attachments

(2 files, 1 obsolete file)

Zipped XML and XSLT files demonstrating the bug (with a readme)
1.17 KB, application/x-zip-compressed
Details
v1.1
5.44 KB, patch
axel
: review+
peterv
: superreview+
Details | Diff | Splinter Review 
I start with XML data rawdata.xml

<?xml version="1.0" ?>
<?xml-stylesheet href="transform-raw.xsl" type="text/xsl"  ?>
<stuff>
   <data> data inside stuff </data> 
</stuff>

and the transform transform-raw.xsl

<?xml version="1.0" ?>
<xsl:stylesheet version="1.0" 
    xmlns:xsl="http://www.w3.org/1999/XSL/Transform" 
    xmlns="http://www.w3.org/1999/xhtml" >
<xsl:output method="xml" indent="yes" />

<xsl:template match="/">
   <html> <head>
   <title> Namespace Fun and Frolic #1 </title>
     </head> <body>
       <xsl:value-of select="//data" />
    </body> </html>
</xsl:template>
</xsl:stylesheet>


I now load rawdata.xml into Mozilla (1.4), which automatically
loads the stylesheet transform-raw.xsl, and displays the generated
results (just the string 'data inside stuff').  

I now use Martin Honnen's javascript: bookmark trick [1] to display
the XML results tree. THe result is:

<a0:html xmlns:a0="http://www.w3.org/1999/xhtml"><a0:head>
<meta http-equiv="Content-Type" content="text/xml; charset=UTF-8"/>
<a0:title> Namespace Fun and Frolic #1 </a0:title>
</a0:head><a0:body> data inside stuff </a0:body></a0:html>

The question: where did the <meta> element come from?  As far
as I can tell, it shouldn't be there.  This looks like an XSLT
bug to me ....

I will attach the example files as a zipped attachment.
OS: Windows 98 → All
Hardware: PC → All
There are two problems with the current code. First of all we should only add
the <meta> in html-output. Second we need to honor the mDocumentIsHTML-flag.
I'll whip up a patch
Assignee: peterv → bugmail
this was mostly fixed by bug 207377, in there i added a test to make sure that
the meta is only added during html-output. We still need test mDocumentIsHTML though
Attached patch v1 (obsolete) — — Splinter Review 

    
  

        Attachment #131032 -
        Flags: superreview?(jst)

        Attachment #131032 -
        Flags: review?(bugmail)

    
    

    
  
Taking.
Assignee: bugmail → peterv
Priority: -- → P3
Target Milestone: --- → mozilla1.6alpha

    
  
Status: NEW → ASSIGNED

    
    

    
  
Comment on attachment 131032 [details] [diff] [review]
v1

- In txMozillaXMLOutput.h:

+    nsresult createHTMLElement(const nsAString& aName,
+				nsIDOMElement** aResult)
+    {
+	 if (mDocumentIsHTML) {
+	     return mDocument->CreateElement(aName,
+					     getter_AddRefs(aResult));
+	 }
+
+	 return
mDocument->CreateElementNS(NS_LITERAL_STRING(kXHTMLNameSpaceURI),
+					   aName, getter_AddRefs(aResult));
+    }

I'd uninline this code, for two reasons. It's not performance critical (at
least not to the point where wether this method is inlined or not matters), and
that NS_LITERAL_STRING expands to a non-trivial ammount of code on some
platforms. But either way works for me...

sr=jst

        Attachment #131032 -
        Flags: superreview?(jst) → superreview+

    
    

    
  

      
      
      
      

        Attached patch
          
          
        v1.1
        — 
          — Splinter Review
      

    


  

        Attachment #131032 -
        Attachment is obsolete: true

    
    

    
  
Comment on attachment 131038 [details] [diff] [review]
v1.1

Transfering sr=jst.

        Attachment #131038 -
        Flags: superreview+

        Attachment #131038 -
        Flags: review?(axel)

    
    

    
  
Comment on attachment 131038 [details] [diff] [review]
v1.1

r=axel@pike.org, I prefer the non-inlined version, too. Esp. as this is going
to go thru more code, once Jonas does his output sink work (getting rid of the
argument to endElement, not calling document.createElement but using element
factories directly)

        Attachment #131038 -
        Flags: review?(axel) → review+

    
  

        Attachment #131032 -
        Flags: review?(bugmail)

    
    

    
  
Checked in.
Status: ASSIGNED → RESOLVED
Closed: 21 years ago
Resolution: --- → FIXED

          You need to log in
          before you can comment on or make changes to this bug.
        






  

    
  







  

    

      
Attachment

      

      
      
      
      
    

    

      

        

          

            
General

            

              Creator: 



            

            
Created: 

            
Updated: 

            
Size: