What I meant is that L'U+code' is probably not one character, but is actually the same as writing L'123456'. You see, what I ask is, have you read somewhere that the VS compiler has such input syntax for unicode - with U+ prefix? As I said, I am not current with MS technology, but AFAIK this is not the proper way to specify unicode codepoints in wchar_t. Why don't you use \xNNNN?