Re: int/float binary conversion

[Lars Christian Jensen]

On Tue, 13 Sep 2005, Lars Christian Jensen wrote:

> The program below prints "f1 = 4.500000, i = 0x08048484, f2 = 0.000000", 
> when compiled with c++ (GCC) 4.0.1 20050727 (Red Hat 4.0.1-5).

Sorry, I forgot the options: -O2 -march=i686

> This don't looks like un-defined behaviour to me, but the same program 
> outputs "f1 = 4.500000, i = 0x40900000, f2 = 4.500000", when compiled 
> with other compilers.
>
>
> #include <cstdio>
>
> int main() {
>   float f1 = 4.5;
>   int i =    *((int*)&f1);
>   float f2 = *((float*)&i);
>
>   printf("f1 = %f, i = 0x%08x, f2 = %f\n", f1, i, f2);
>
>   return 0;
> }
>
> --
> Lars Christian Jensen