unsigned integers in C

CoderMan · 03-24-2009, 06:06 PM

From what I understand, if I attempt to increment an unsigned int which is already at its max value, it will flip back to zero.

Question: is this guaranteed to happen by the C language? That is, can I count on this happening on all platforms?

(Of course, I recognize that the max value of the int will change from platform to platform...)

khodeir · 03-24-2009, 06:10 PM

Quote:

Originally Posted by CoderMan

From what I understand, if I attempt to increment an unsigned int which is already at its max value, it will flip back to zero.

Question: is this guaranteed to happen by the C language? That is, can I count on this happening on all platforms?

(Of course, I recognize that the max value of the int will change from platform to platform...)

when you exceeds the limit the register overflows this will happen on any operating system try factorial 444444
you must take care when you make a program
try using float number because it has 8 or 10 bytes as i remembered and this is a greet resolution
if you want more : there are libraries for encryption that has 1000 bit as i think

that will make all things large

pixellany · 03-24-2009, 06:15 PM

I think it is splitting hairs as to whether it is C or the hardware. I believe this is called overflow---Imagine an 8-bit register containing 11111111 (hex FF, dec 255). Add 1, and it tries to go to 100000000, but that "1" is the ninth bit, which--as they say--falls on the floor. With an 8 bit register, you don't know if your count is at 0 or 256 (or 512,768, etc.)

The header files define exactly what different data types are on various HW platforms.

johnsfine · 03-24-2009, 06:31 PM

Quote:

Originally Posted by CoderMan

is this guaranteed to happen by the C language?

I don't know. A surprisingly large number of similar basic behaviors in boundary conditions are NOT guaranteed by the language standard (there also may be a question of which version of the language standard).

Quote:

That is, can I count on this happening on all platforms?

But you can count on this happening on all platforms even if it is not specified by the language standard.

There is no platform you might port your C code to in which that isn't the standard hardware behavior and no one implementing a compiler would go to the extra effort required to make that not work, even if they also wouldn't go to any extra effort to make it work.

Hko · 03-24-2009, 06:44 PM

Quote:

Originally Posted by khodeir

if you want more : there are libraries for encryption that has 1000 bit as i think

that will make all things large

If it is about big integers, try for the GMP library for integers as large as fit in memory.

Biddle · 03-24-2009, 07:50 PM

Quote:

is this guaranteed to happen by the C language?

Yes the standard guarantees that the addition of two unsigned integers will silently wrap (not overflow).

Quote:

6.5.2.9
The range of nonnegative values of a signed integer type is a subrange of the
corresponding unsigned integer type, and the representation of the same value in each
type is the same.31) A computation involving unsigned operands can never overflow,
because a result that cannot be represented by the resulting unsigned integer type is
reduced modulo the number that is one greater than the largest value that can be
represented by the resulting type.

This guarantee is used in code for example in a hashing function.

Quote:

there also may be a question of which version of the language standard)

The above quote is from c99 9899 1999