Converting floats to and displaying Binary Numbers in C

Question

So I'm just diving into C during my first few days of classes. My professor proposed a question/ more of a challenge for us so I thought I'd propose it here.

float x,y,z;
x = 0.0;
y = 1.0;
z = -2.5;

He wanted us to display these numbers in binary and hexadecimal without using %x etc. He also hinted to use a union:

union U {
float y;
unsigned z;
};

Thank you for your time in advance. I'm just looking for help/ an explanation on how to develop such conversion.

Well, language lawyer would say the professor is committing a crime... — Eugene Sh., Sep 08 '16 at 14:23
Hmm, so you ask **us** to do **your** challenge? Do we get the exam then? — too honest for this site, Sep 08 '16 at 14:31
Ok, so the `union` is just a hint for getting the float marshaled into an integer. Now take a paper and see how you would convert an integer to base-2 or base-16 and then try to code this. — vgru, Sep 08 '16 at 14:41
[This](http://stackoverflow.com/a/35367414/2410359) will convert any object to binary. It does not use `"%x"`. — chux - Reinstate Monica, Sep 08 '16 at 15:21

score 1 · Accepted Answer · edited Sep 08 '16 at 14:53

1

The correct way, not violating the strict aliasing rule would be to alias the float with a char array, which is allowed by the C standard.

union U {
  float y;
  char c[sizeof (float)];
};

This way you will be able to access the individual float y; bytes using the c array, and convert them into binary/hex using the very simple algorithm that can be easily found around (I will leave it up to you, as it is your assignment).

edited Sep 08 '16 at 14:53

unwind

391,730
64
469
606

answered Sep 08 '16 at 14:52

Eugene Sh.

17,802
8
40
61

`unsigned char` has portability advantages over `char` as, rarely, `char` may have trap values. Further, a `char` that is _signed_ , slightly complicates conversion to hex/binary display. – chux - Reinstate Monica Sep 08 '16 at 16:05
@chux I was considering writing `unsigned` as it has advantages with converting based as well. But decided to leave it out to reduce the noise.. – Eugene Sh. Sep 08 '16 at 16:14

Converting floats to and displaying Binary Numbers in C

1 Answers1