XslTransform.Transform, invalid character output to Response.OutputStream

[Tim Menninger]

When I use the XslTransform.Transform method to write to the
HttpRequest.OutputStream The first byte of the output is always an invalid
character, looks like an ascii zero or some other non-printable ascii char.
The problem is that it messes up my display.

Here is the code:
....
Response.Write("<span>some text</span>");
XmlTextReader reader = new XmlTextReader(xslFile);
XslTransform xsl = new XslTransform();
xsl.Load(reader, resolver, evidence);
xsl.Transform(xmlDoc, args, Request.OutputStream, resolver);

HTML Output:
<span>some text</span>"invalid char"<div>Some transformed html</div>

The problem is that the output should look like this:
some text
Some transformed html

But when the html parser finds the <div> tag it notices that there is
already text on the current line (the invalid char) and starts on the next
line. This adds an extra line break that looks like this:
some text

Some transformed html

Has anyone seen this before? If so is there a fix or workaround? I would
hate to have to transform to a string and then use Response.Write() to write
out the transformation

Thanks,

Tim

[Oleg Tkachenko [MVP]]

Tim Menninger wrote:

When I use the XslTransform.Transform method to write to the
HttpRequest.OutputStream The first byte of the output is always an invalid
character, looks like an ascii zero or some other non-printable ascii char.
The problem is that it messes up my display.

Most likely that's unicode byte order mark (BOM). To avoid BOM
generation declare ISO8859-X output encoding in the stylesheet using
xsl:output instruction. (e.g. ISO8859-1 for Latin1 charset).
If you do need UTF-8 output, but without BOM, there is some trick:

StreamWriter sw = new StreamWriter(Request.OutputStream,
new UTF8Encoding(false));

And then

xsl.Transform(xmlDoc, args, sw, resolver);

--
Oleg Tkachenko [XML MVP, XmlInsider]
http://blog.tkachenko.com