[XALANJ-2132] Incorrect serialization of supplementary Unicode characters

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Duplicate
• Affects Version/s: 2.6
• Fix Version/s: 2.7
• Environment: Windows XP, Sun JRE 1.5.0_03
• Assignee: Unassigned
• Reporter: Adam P. Lally
• Created: 01/Jun/05 05:36
• Updated: 09/Aug/05 14:46
• Resolved: 08/Jun/05 07:17

Description

Passing a surrogate pair to a TransformerHandler causes incorrect UTF-8 to be generated. The following code illustrates the problem:

SAXTransformerFactory transformerFactory = (SAXTransformerFactory)SAXTransformerFactory.newInstance();
TransformerHandler handler = transformerFactory.newTransformerHandler();
handler.setResult(new StreamResult(System.out));

char[] chars = new char[2];
chars[0] = (char)0xD803;
chars[1] = (char)0xDD75;

handler.startDocument();
handler.startElement("","","foo", new AttributesImpl());
handler.characters(chars, 0, chars.length);
handler.endElement("","","foo");
handler.endDocument();

If you take the output of this program and try to parse it with the Xerces SAX Parser, you get an "Invalid byte 2 of 3-byte UTF-8 sequence" exception.