I have a few simple questions, because I got confused reading all difference responses.
1) If I have an xml with prolog: <?xml version="1.0" encoding="utf-8" ?> and I’m going to unmarshall it with Java (for example: JaXB). I suppose, that I can’t put CROSS OF LORRAINE (http://www.fileformat.info/info/unicode/char/2628/index.htm) inside, but I can put “\u2628”, correct?
2) I’ve also heard that UTF-8 doesn’t contain it, but anything in Unicode can be saved with encoding UTF-8 (or UTF-16), and here is an example from this page:
UTF-8 (hex) 0xE2 0x98 0xA8 (e298a8)
Is my reasoning correct? Can I use this form and put it in the xml with utf-8 encoding?
It should be absolutely fine – UTF-8 can encode any Unicode character.
XML has some restrictions around control characters (U+0000 to U+001F) but U+2628 should be fine.
(Personally I prefer to go to unicode.org for definitive code charts, but U+2628 definitely appears here.)
You shouldn’t need to worry about the UTF-8 side of things – you should be able to put the character in your data directly, and let JAXB do the encoding.