I have a problem when transforming text containing the swedish letters
"å", "ä" and "ö". If I do

Transformer t =TransformerFactory.newInstance().newTransformer();
t.setOutputProperty( OutputKeys.METHOD, "xml");
t.setOutputProperty("{http://xml.apache.org/xslt}indent-amount", "2");
t.setOutputProperty( OutputKeys.INDENT, "yes");
t.setOutputProperty( OutputKeys.ENCODING, "ISO-8859-1");   <------- *
t.transform( new DOMSource( document), new StreamResult( output ) );
return output.toString( );

I get an xml-file containing broken characters (=?) for the swedish
letters:

<?xml version="1.0" encoding="ISO-8859-1"?>
...
<channelinfo confirmed="true" validate="false" name="Internet">
     <publishdate>1154940455898</publishdate>
     <unpublishdate>1154940455898</unpublishdate>
     <attribute name="rooms"/>
     <attribute name="year"/>
     <attribute name="title">K?pes</attribute>  <------------- *
     <attribute name="price">20000</attribute>
     <attribute name="area"/>
     <attribute name="body">Vill k?pa en truck</attribute>
<-------------- *
</channelinfo>

but if I change the encoding to UTF-8:

t.setOutputProperty( OutputKeys.ENCODING, "UTF-8");   <------- *

the letters are alright:

<?xml version="1.0" encoding="UTF-8"?>
...
<channelinfo confirmed="true" validate="false" name="Internet">
     <publishdate>1154940455898</publishdate>
     <unpublishdate>1154940455898</unpublishdate>
     <attribute name="rooms"/>
     <attribute name="year"/>
     <attribute name="title">Köpes</attribute>  <------------- *
     <attribute name="price">20000</attribute>
     <attribute name="area"/>
     <attribute name="body">Vill köpa en truck</attribute>
<-------------- *
</channelinfo>

But the xml has to be formated in ISO-8859-1 so it would be nice if I
could make it work with that encoding.

Anyone know where I can alter this behavior or why it behaves like
above?