Metacat

import java.nio.charset.Charset;

import org.apache.commons.io.input.XmlStreamReader;

    /**

     * test Commons IO detection

     */

    public void testXMLEncodingDectection()

    {

        try

        {

        	System.out.println("default charset:" + Charset.defaultCharset().displayName());

        	String sampleXML = "<?xml version='1.0' encoding='UTF-8'><test>my content 你</test>";

        	// get bytes using different encoding - shouldn't matter what we use for the prolog

        	XmlStreamReader xsr = 

        		new XmlStreamReader(

        				new BufferedInputStream(new ByteArrayInputStream(sampleXML.getBytes("ISO-8859-1"))));

        	System.out.println("detected encoding:" + xsr.getEncoding());

        	// read the string [again] using the detected encoding

        	// NOTE: XmlStreamReader consumes the entire stream and does not suport reset()

        	// Besides, we'd have the wrong bytes anyway

        	String result = IOUtils.toString(new ByteArrayInputStream(sampleXML.getBytes(xsr.getEncoding())));

            System.out.println(result);

        	assertTrue(result.equals(sampleXML));

        }

        catch(Exception e)

        {

        	e.printStackTrace();

            fail("Unexpected error in testXMLEncodingDectection: " + e.getMessage());

        }

    }