Convert 4 bytes into a float

[Krisando]

I've being trying to convert 4 bytes into a float with no success.

Float:
0.1

The bytes are:
1: 205
2: 204
3: 204
4: 61

A formula for doing so would be great, I have to do this without declaring any float values, purely the formula which will result in the float.

I do not think posting my attempts will benefit this topic, I really have no idea how to do this, It seems extremely complicated all this mantissa kind of stuff.

[jsmith]

You have to look up the IEEE floating point format.

You can get it here:
http://babbage.cs.qc.edu/IEEE-754/Decimal.html

[blackcoder41]

hmm. seems like bit manipulation to me.. do you mean the four bytes you posted should make a float equivalent to 0.1?

[kbw]

I have no idea why you are dong this, but I hope this code helps you understand what's going on.

#include <iostream>

union UStuff
{
        float   f;
        unsigned char   c[0];
};

int main()
{
        UStuff a, b;
        a.f = 0.1;

        for (size_t i = 0; i < sizeof(UStuff); ++i)
        {
                std::cout << "byte " << i << ": "
                        << (unsigned int)(a.c[i])
                        << std::endl;

                b.c[i] = a.c[i];
        }
        std::cout << b.f << std::endl;

        return 0;
}

Output (intel mac):

byte 0: 205
byte 1: 204
byte 2: 204
byte 3: 61
0.1

[blackcoder41]

@kbw
nice!

[Krisando]

Ooo, this looks good.

Having abit of trouble trying to decode this code with 1 letter vars. Hard to tell what is going on. =\ Maybe comments?

[blackcoder41]

hehe.. i am also amaze by that code.. what part don't you understand? so i can explain..

[Krisando]

I'm just not sure what each line is used for.

union UStuff
{
        float   f;
        unsigned char   c[0];
};

I've not seen union before, this looks abit like a struct. Is there any difference between the two?
I do not see what sort of purpose initilizing 1 cell for variable c would have? why not just c without the 1 cell array? Same thing I think.

size_t i = 0
Why would you use size_t? Would it give the "i" a data type suitable for holding the bytesize for sata structure UStuff?

std::cout << "byte " << i << ": "
<< (unsigned int)(a.c[i])
<< std::endl;

b.c[i] = a.c[i];

These one letter variables are confusing me, I cannot tell what they are for. =[

[R0mai]

http://www.cplusplus.com/doc/tutorial/other_data_types/#union

[kbw]

Where did the problem from in the first place. If you can't understand the code, you'll have to explain where the question came from.