That looks like a bug in BytecodeName.parseBytecodeName. The 'new String' shortly after the error point should have been 'new Object'. Time for more unit testing! -- John