usigned int java

[Rajesh.Rapaka]

Hi,

I need to used an unsigned integer in my program. I am actually
developing a protocol which at a certain points asks me to transfer 4
bytes unsigned. long makes 8 bytes. How can I transfer the unsigned int
value.

any help or info is appreciated.
regards
Rajesh Rapaka

 

[Roedy Green]

I need to used an unsigned integer in my program. I am actually
developing a protocol which at a certain points asks me to transfer 4
bytes unsigned. long makes 8 bytes. How can I transfer the unsigned int
value.

when transferring bytes an int will do for unsigned. It is only when
you do a multiplication or compare does signed/unsigned matter.

See http://mindprod.com/jgloss/unsigned.html

 

[Googmeister]

when transferring bytes an int will do for unsigned. It is only when
you do a multiplication or compare does signed/unsigned matter.

I think you mean division/remainder instead of multiplication.
I'd also add casting to the list.

 

[Roedy Green]

I think you mean division/remainder instead of multiplication.
I'd also add casting to the list.

compare, multiplication, division, remainder all are different from a
bit point of view for signed/unsigned. addition, subtraction, move are
not.

see http://mindprod.com/jgloss/unsigned.html for details.

 

[Googmeister]

compare, multiplication, division, remainder all are different from a
bit point of view for signed/unsigned. addition, subtraction, move are
not.

Are you sure about multiplication? Can you provide
a counterexample? Thanks.

 

[Roedy Green]

Are you sure about multiplication? Can you provide
a counterexample? Thanks.

In Forth there are signed * and unsigned U* multiplication operators.
I implemented them in BBL Forth. There I was multiplying to 32 bit
quantities to get a 64 bit result.

If you have a hex calculator with more than 32 bits, try some
experiments with two +, two - and mixed, treated both as unsigned and
signed. pick some large and small numbers.

Or try the same in Java with 16 bits where you have char and short to
play with to give you a 32 bit result.

 

[Mark Thornton]

In Forth there are signed * and unsigned U* multiplication operators.
I implemented them in BBL Forth. There I was multiplying to 32 bit
quantities to get a 64 bit result.

If you have a hex calculator with more than 32 bits, try some
experiments with two +, two - and mixed, treated both as unsigned and
signed. pick some large and small numbers.

Or try the same in Java with 16 bits where you have char and short to
play with to give you a 32 bit result.

The low order 32 bits of a 32 bit multiply is the same for signed and
unsigned. All that differs is testing for overflow.

Mark Thornton

 

[Googmeister]

The low order 32 bits of a 32 bit multiply is the same for signed and
unsigned. All that differs is testing for overflow.

Thanks, that's the same conclusion I reached. Since Java truncates
overflow, you get exactly the same bits (just like for add/subtract).

 

[Roedy Green]

The low order 32 bits of a 32 bit multiply is the same for signed and
unsigned. All that differs is testing for overflow.

Here is an example demonstrating this is so. A 32x32->64 bit differs
between signed and unsigned however.

// unsigned 32-bit multiply
long a = ( 0x7fffffffL * 0xffffffffL ) & 0xffffffffL;

// signed 32-bit multiply
long b = ((long) ( 0x7fffffff * 0xffffffff )) & 0xffffffffL;

// prints 2147483649
System.out.println( a );
// prints 2147483649
System.out.println( b );