The original bytes were probably meant as UTF-8, not UTF-16

...not that it matters much /what/ characters are used in the sample text, of
course.  And no-argument java.lang.String.getBytes() appears to indeed use the
legacy Mac OS Roman encoding on Mac OS X...

Change-Id: I51cd8c16f14e2df8b75d9961150518c8a03d53ca

unoxml/qa/complex/unoxml/DOMTest.java

@@ -2706,7 +2706,7 @@ public class DOMTest
             "<xlink:test/>" +
             "<office:automatic-styles teststyle=\"test\"/>" +
             "<moretest/>" +
-            "some text \uC3B6\uC3A4\uC3BC" +
+            "some text \u00F6\u00E4\u00FC" +
             "</office:document-content>";
 
         XDocumentBuilder xBuilder =