Fesersoft

Menu

Numeric Character References (NCRs):

This is a problem that came up in microsoft.public.dotnet.xml. The message is here.

I have a stylesheet (given below) that's sole purpose is to transform a UTF-8 document to us-ascii using Numeric Character References (NCRs). This allows our software that's not Unicode compatible to at least work with it - to some degree.

However, *ALL I GET* is UTF-8 - it's like .NET XslTransform is somehow ignoring the new encoding. I've tried changing the XmlTextWriter encoding from UTF8 to ASCII but neither helps.

Please Help!

Here's my code and stylesheet:

XPathDocument xmlDoc;
XslTransform xslT;
XmlTextWriter myWriter;

myWriter = new XmlTextWriter(
Server.MapPath( "outXML.xml" ),
System.Text.Encoding.UTF8 );

xmlDoc = new XPathDocument(
Server.MapPath( "XMLCompareFiles\\s2v21chp8f11529.xml" ) );

xslT = new XslTransform();

xslT.Load( Server.MapPath
( "Transform2NCR.xslt" ) );
xslT.Transform( xmlDoc, null,
myWriter );

myWriter.Close();


<xsl:stylesheet
xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
version="1.0">
<xsl:output omit-xml-declaration="no" method="xml" media-type="text/xml" indent="no" encoding="iso8859-1" />
<xsl:template match="node()|@*">
<xsl:copy>
<xsl:apply-templates select="node
()" />
</xsl:copy>
</xsl:template>
</xsl:stylesheet>

The original source was UTF-8 and the goal was to output the document as NCR's because the system the document needed to be sent to did not support Unicode.

We tried about 5 different ways to try to create a .Net solution that would not destroy the output. Even if the encoding method on the XSLT was set to iso8559-1, the output was:

<?xml version="1.0" encoding="iso-8859-1"?><objects><customers><name>?????</name>
<otherInfo>??????????????</otherInfo></customers></objects>

The correct output should be:

<?xml version="1.0" encoding="iso8859-1"?>
<objects><customers><name>第學期第章</name>
<otherInfo>有開機系統命令
路由器會到哪裡</otherInfo>
</customers></objects>

The following .Net Methods were all tried with the output seen in the first example.

public string MemoryStream()
{
XslTransform Xsl = new XslTransform();
string Result;
//string FullPath = xmlUrl;
MemoryStream ms = new MemoryStream();
XmlTextWriter tr = new XmlTextWriter(ms,System.Text.Encoding.GetEncoding(1252));
//Load the stylesheet.
Xsl.Load("Transform2NCR.xslt");
//Create a new XPathDocument and load the XML data to be transformed.
XPathDocument mydata = new XPathDocument("SampleXML.xml");
Xsl.Transform(mydata, null, tr);
tr.Flush();
ms.Position = 1;
StreamReader sr = new StreamReader(ms);
Result = sr.ReadToEnd();
tr.Close();
return (Result);
}

public string StringWriter()
{
XslTransform myXslTransform;
XPathDocument myXPathDocument = new XPathDocument("SampleXML.xml");
myXslTransform = new XslTransform();
myXslTransform.Load("Transform2NCR.xslt");
System.Globalization.CultureInfo ci = new System.Globalization.CultureInfo(0x0409);
System.IO.StringWriter stWrite = new System.IO.StringWriter(ci);
myXslTransform.Transform(myXPathDocument, null, stWrite);
return (stWrite.ToString());
}

public void usingFileStream()
{
XslTransform Xsl = new XslTransform();
//string Result;
//string FullPath = xmlUrl;
Stream fs = new FileStream("outputFileStream.xml",FileMode.Create,FileAccess.Write);
//XmlTextWriter tr = new XmlTextWriter(ms,System.Text.Encoding.ASCII);
//Load the stylesheet.
Xsl.Load("Transform2NCR.xslt");
//Create a new XPathDocument and load the XML data to be transformed.
XPathDocument mydata = new XPathDocument("SampleXML.xml");
Xsl.Transform(mydata, null, fs);
fs.Flush();
//fs.Position = 1;
//StreamReader sr = new StreamReader(fs);
//Result = sr.ReadToEnd();
fs.Close();
//return (Result);
}

public void WriteToFile()
{
XslTransform Xsl = new XslTransform();
//Load the stylesheet.
Xsl.Load("Transform2NCR.xslt");
Xsl.Transform("SampleXML.xml","outFileToFile.xml");
}

As you can tell, we were getting pretty desperate.

Then we Tried to interop with MSXML4.0, performing the transformation using XSLTemplate40 and setting the output to a second MSXML Document. When you moused over the .XML property of the DomDocument, the output was correct, but when you read the result into a StreamWriter, the result was messed up again.
Ok so not really having the time to care if it's an interop problem, I tried a different method. I Transformed the object into an ADODB.Stream object, setting the encoding to ISO-8559-1 then Returned the result using the ReadText Method. IT WORKS!!! The only hack we could figure out was to interop with MSXML and ADODB and Return the Result from a function call.

Here is the sample code that returned the correct result.

Public Function MsXMLADOStreamVB() As String
Dim oXML As New MSXML2.DOMDocument40()
Dim oXSL As New MSXML2.DOMDocument40()
Dim oStream As New ADODB.Stream()
Dim bLoaded As Boolean
Dim Result As String
oXML.async = False
oXSL.async = False
oXML.validateOnParse = False
oXSL.validateOnParse = False
bLoaded = oXML.load("SampleXML.xml")
If Not bLoaded Then
     MsgBox("XML Not Loaded")
End If
bLoaded = oXSL.load("Transform2NCR.xslt")
If Not bLoaded Then
     MsgBox("XSL Not Loaded")
End If

Try
     oStream.Charset = "iso8859-1"
     oStream.Open()
     oXML.transformNodeToObject(oXSL, oStream)
Catch e As Exception
     MsgBox(e.ToString)
End Try
oStream.Position = 0
Result = oStream.ReadText()
Return (Result)
End Function

If anyone knows the real solution, or knows if this is a bug, please email me at nntp@fesersoft.com.

Thank you,

Joe