float comparation problem

[aBiNg]

Here is the code:

#include <stdio.h>
#define float_cmp(x, y) ((int)((x) * 100 + 0.5) / 100. >= (int)((y) *
100 + 0.5) / 100.) ? 1 : 0

int main(void)
{
float a = 0.014, b = 0.015;

printf("The result is: %d\n", float_cmp(a, b));
return 0;
}

I compiled this source with gcc-3.4.5, and run with it, the output is:
The result is: 1

(b) * 100 + 0.5 is equal to 2.000000, then i think (int)((b) * 100 +
0.5) should be 2, but this is not true. In my system, it is 1.
I don't know the type-cast mechanism of gcc, can anybody show me the
reason?

Any help will be very appreciated.

 

[aBiNg]

Your explainations make this code reasonable. Thanks a lot.

I am pondering how to add a patch to make this code work.

 

[luserXtrog]

Your explainations make this code reasonable. Thanks a lot.

I am pondering how to add a patch to make this code work.

#include <math.h>
double round(double x);