Problems with String.getBytes() method

Question

Consider this code
    for(int k=0 ; k<256 ; k++)
       char[] cc = {(char)k};
       String xxx = new String(cc);
       byte b1 = (byte)xxx.charAt(0);
       byte[] bb = xxx.getBytes();
       byte b2 = bb[0];
       boolean diff = (b1!=b2);
       if (diff)
       System.out.println(k+" - "+b1+" - "+b2);
     }for k from 128 to 159 b1 and b2 are different and in particular b2 (that is the result of getString() method) is every time equals to 63.
Does someone know if it is a bug or what else.

Answer 1

char variables store 16 bit UNICODE characters. If you cast char to byte you're just taking the low order 8-bits of a unicode character.
On the other hand getBytes attempts to convert a character into a byte, or sequence of bytes using the default character encoding of the machine, which won't generally corespond. National character encodings meen that non-ASCII characters, including national stuff like pound sign, may well map from UNICODE characters above 255 into a single byte.
The character encoding which does corespond to the bottom 8 bits of UNICODE is ISO-8859-1. If you try it with that your discrepancies will disappear.