Java encoding...

Hello,

I need to write a byte stream onto the network, what i plan to do is take the byte arrays i have, convert them to a string and recreate the byte array using the string i receive on the other end.

Let me illustrate my problem...

1. bss is what i wish to transfer
byte bss[] = new byte[1];
bss[0] = (byte) (129 & 0xff);
System.out.println("Byte value: " + bss[0]);
System.out.println("Int value: " + ( bss[0] & 0xff ));

2. convert bss to a string
StringBuffer ston = new StringBuffer();
ston.append(new String(bss,"UTF-8"));

3. Recreate bss from string
System.out.println(ston.toString());
byte bss2[] = ston.toString().getBytes();
System.out.println("Byte value: " + bss2[0]);
System.out.println("Int value: " + ( bss2[0] & 0xff ));

For some reason my output is:-
Byte value: -127
Int value: 129
?
Byte value: 63
Int value: 63

I need the same byte value and int value after recreating the bytes using getBytes(), i think there is some issue with the encoding issues coming up, wondering if anyone knows a fix for what i'm trying to do..

Thanks
Neville

i'm not really sure,but which jdk do u use????if u use jdk6 , u have to go with somthing new,generics and collections.

Hi.
If I understand it well, than you are sending 129 which is 10000001 in binary form. From what I read on Wikipedia on UTF-8 I think that if the number starts 1 it means that the letter is saved in 2 numbers. The documentation on String(byte[] bytes, String charsetName) says:

So I think that you are getting this unspecified behavior. See the charset and try to send for example number 10. If it works, than you have better use different charset. UTF-8 has too many rules for your case.