tags:

views:

475

answers:

8

How would you set a variable to equal infinity (or any guaranteed largest number value) in C?

+14  A: 

There is a file called limits.h (at least on Linux there is), which holds this kind of definition e.g.

/* Maximum value an `unsigned short int' can hold.  (Minimum is 0.)  */
#  define USHRT_MAX 65535

/* Minimum and maximum values a `signed int' can hold.  */
#  define INT_MIN   (-INT_MAX - 1)
#  define INT_MAX   2147483647

/* Maximum value an `unsigned int' can hold.  (Minimum is 0.)  */
#  define UINT_MAX  4294967295U
amo-ej1
Tronic
The same as online reference: http://www.utas.edu.au/infosys/info/documentation/C/CStdLib.html#limits.h
h0b0
Also interesting to note is how `INT_MIN` is defined on this particular system. They can't define it as `-2147483648`, because that would mean "negative 2147483648", and 2147483648 overflows an `int`!
Alok
+1  A: 

Normally this is done by 1.0/0.0, but you may get a compile warning on that. I am not aware of other portable ways of doing it in C89, but C99 has macro FP_INFINITE in math.h.

EDIT: Apparently Sam didn't actually want infinity, but integer limits, which can be found in limits.h like others have stated.

Tronic
Hmm didn't think about checking math.h for something like that. This will definitely work. Thanks!EDIT: limits.h works better in my case, sorry for the confusion lol
Sam
`1.0/0.0` is not portable, and in fact constructing infinity bitwise using a union with `float` and `uint32_t` is probably more portable than division by zero.
R..
+7  A: 
#include <limits.h>
int x = INT_MAX;

EDIT: answered before the questioner clarified, I was just guessing what type they wanted.

profjim
thanks limits.h is actually what i needed
Sam
thanks for the accept, but I think amo-ej1's answer is more complete.
profjim
+1  A: 

I generally use the *_MAX macros found in limits.h INT_MAX for integers etc. These will always be correctly set for the variable type. Even the explicitly sized types such as uint32 will have corresponding entries in this header file.

This has the virtue of being the largest possible value that a variable of that type can hold.

For an unsigned integer as you asked for in your question, you would use UINT_MAX

Andrew
+6  A: 

By far the simplest method to get the largest value for an unsigned integer type is to cast (-1) to that type. The standard (§6.2.5/9) requires that unsigned math be carried out modulo a number one greater than the largest value that can be represented, so for any unsigned type T, the expression ((T)-1) will necessarily be the largest value possible in that type.

Jerry Coffin
+1  A: 

Since there's a C++ tag on this question, I'll suggest numeric_limits:

#include <limits>

unsigned x = std::numeric_limits<unsigned>::max();
Fred Larson
Before anybody down votes me, there WAS a C++ tag until it was removed by David Rodríguez. Looking at the edit history, I see it was added by someone other than the OP.
Fred Larson
A: 

I guess you may want to check this link out http://www.gnu.org/s/libc/manual/html_node/Infinity-and-NaN.html

I did this and it works fine on gcc 4.4.1


#include "math.h"

int main(int argc, char**argv)
{ 
    int x = INFINITY; 
    return 0;
}

thelinuxer
+2  A: 

Based upon your comments, you want an unsigned int (although you say "unsigned integer", so maybe you want an integral value, not necessarily an unsigned int).

In C, for unsigned integral type, the value -1, when converted to that type, is guaranteed to be largest value of that type:

size_t size_max = -1;
unsigned int uint_max = -1;
unsigned long ulong_max = -1;

assign the values SIZE_MAX, UINT_MAX and ULONG_MAX to the variables respectively. In general, you should include limits.h and use the appropriate macro, but it is nice to know the rule above. Also, SIZE_MAX is not in C89, so size_t size_max = -1; will work in C89 as well as C99.

Note that the overflow behavior is guaranteed only for unsigned integral types.

Alok
Yes, this is good to know for cases where you don't want to rely on the macro being defined. But is it best coding practice? Someone else might look at the code and be unsure whether there was a mistake (perhaps coming from careless revisions). Using the macros helps document your intent.
profjim
@profjim: as I said, using macros is better, but in case you can't (C89 size_t for example), you should use `-1`.
Alok