You're doing a two-stage transformation here: first a transformation using a
stylesheet, which then feeds its output in the form of a SAXSource to a JAXP
identity transform which converts the SAXSource to a StreamResult.

I don't know why you're doing this or why it adds newlines, but it does mean
that the xsl:output is going to be ignored because the output of the
stylesheet is not serialized.

What happens if you remove the (apparently unnecessary) second step?

Michael Kay
http://www.saxonica.com/
 

> -----Original Message-----
> From: John English [mailto:je@xxxxxxxxxxxxxx] 
> Sent: 03 October 2008 13:23
> To: xsl-list@xxxxxxxxxxxxxxxxxxxxxx
> Subject: Re: [xsl] Transformer adds unwanted newlines
> 
> David Carlisle wrote:
> >>Anyone got any ideas what I can do about this?
> > but ignoring indent=no sounds less likely, are you sure you 
> don't have 
> > another xsl:output somewhere, or that you are not outputing an XMl 
> > tree that is then being serialised by some other JDK methord rather 
> > than being serialised by XSLT?
> 
> No other xsl:output, and all I do in the transform is this:
> 
>    SAXParserFactory spf = SAXParserFactory.newInstance();
>    spf.setNamespaceAware(true);
>    SAXParser parser = spf.newSAXParser();
>    reader = parser.getXMLReader();
>    stf = (SAXTransformerFactory) TransformerFactory.newInstance();
>    String s = serializeTree();
>    System.out.println("DEBUG: " + s);     // <------- DEBUG
>    InputSource input = new InputSource(new StringReader(s));
>    filter = stf.newXMLFilter(new StreamSource(stylesheet));
>    filter.setParent(reader);
>    StreamResult result = new StreamResult(response.getWriter());
>    Transformer transformer = stf.newTransformer();
>    SAXSource transformSource = new SAXSource(filter, input);
>    transformer.transform(transformSource, result);
> 
> At the point where the DEBUG output is, the serialized tree 
> is as I've described it; the extra newlines appear in the 
> output also as described.
> 
> In the meantime I have kludged up a workaround when I serialise the
> tree:
> 
>    if (tagName.equals("pre")) {
>      d = d.replaceAll("<br[^>]*>","\n");
>    }
> 
> This works but it offends my sensibilities!
> 
> Thanks,
> 
> ----------------------------------------------------------------------
>   John English              | mailto:je@xxxxxxxxxxxxxx
>   Senior Lecturer           | http://www.it.bton.ac.uk/staff/je
>   School of Computing & MIS | "Those who don't know their history
>   University of Brighton    |  are condemned to relive it" (Santayana)
> ----------------------------------------------------------------------