32-bit unsigned math with C#?

[Brians256]

Why is it not possible to use bit-wise operators on 32-
bit unsigned integers? Isn't that a bit backwards? How
else can you specify a full 32-bits?

I'm trying to use some C code by translating it to C#.
However, it appears that I must either suffer a
performance hit by moving to 64-bit integers or stuff the
algorithms into a DLL via another more sane language.

Consider this C code:

int foo(void)
{
int i = 0x81234567;
int j = i >> 1;
return (j & ~i);
}

You can't do that easily in C#. A System.Int32 type
cannot contain 0x80000000 or larger value. However, a
System.UInt32 type will not allow bitwise operators.

Is Microsoft insane, or have they restricted the UInt32
type for some good reason? Hopefully I am missing
something simple, so I can port my hashing function to C#.

[Pieter Philippaerts]

"Brians256" <Br*******@hotmail.com> wrote in message

Why is it not possible to use bit-wise operators on 32-
bit unsigned integers?
It is possible.
Consider this C code:
Here are two equivalent C# versions:

// using unsigned integers
public uint foo2() {
uint i = 0x81234567;
uint j = i >> 1;
return (j & ~i);
}
// using signed integers
public int foo() {
int i = -2128394905;
int j = i >> 1;
return (j & ~i);
}

Both compile and work fine on my computer.
Hopefully I am missing
something simple, so I can port my hashing function to C#.

What errors do you exactly get with your code? Posting some code would be
helpful.

Regards,
Pieter Philippaerts
http://www.mentalis.org/

[Brians256]

<insert egg on face>

Ooops, hold on. Wrong project I got that error log
from. foo2() compiles correctly, now I have to figure
out why my prior example code didn't work as expected.

-----Original Message-----

// using unsigned integers
public uint foo2() {
uint i = 0x81234567;
uint j = i >> 1;
return (j & ~i);
}
// using signed integers
public int foo() {
int i = -2128394905;
int j = i >> 1;
return (j & ~i);
}

Both compile and work fine on my computer.
What errors do you exactly get with your code? Postingsome code would be

helpful.

I pasted foo2() into my C# code and it failed just as my
prior code example did. The reason I have not posted

thecode is that the code is proprietary, but the problem is
so easily reproduced that I believed my problem to be
well known and understood. Also, the warnings that I getare truly odd. The error reporting is malformed and mostof the errors have nothing to do with the actual error.

This is what I actually get when I pasted foo2() into an
existing and compiling C# file. Note that none of the
line number actually associate with the foo2() function,
which ranges from line 105 to line 110 in this file
kmd5.cs.

c:\programs\sample\kmd5.cs(114,13): error CS0650: Syntax
error, bad array declarator. To declare a managed array
the rank specifier precedes the variable's identifier
c:\programs\sample\kmd5.cs(134,41): error CS1031: Type
expected
c:\programs\sample\kmd5.cs(134,50): error CS1002: ;
expected
c:\programs\sample\kmd5.cs(134,50): error CS1525: Invalidexpression term ')'

Oh, and I already know that .NET has a built-in MD5 hash
function. This was just part of an existing library thatI'm porting into .Net (I hope).

I'm using VS.2003, if that's any help.
.

[Brians256]

It was all a figment of my imagenation, I suppose.
Another error masqueraded quite successfully to convince
me that you cannot use the uint type with the shift
operator. Believe it or not, but the task list was not
as accurate as the build output window, and I got red-
herrings from bad line numbers.

Thanks for the patience.