Thread: Unions...

Hi.

I have been reading about unions.
I have but together what i think is the solutions to the following:

I have a float value and i would like to easily break it into 4 bytes and view each byte. Pretty straight forward!!

So outside of main() i have this.

Code:

union 
  {
    float AsFloat;
    char AsByte;
  } FloatUnion;

Inside main() i have this.

Code:

char float_union_bytes[4];
    
    FloatUnion.AsFloat = 24.65;
    float_union_bytes[0] = FloatUnion.AsByte[0];
    float_union_bytes[1] = FloatUnion.AsByte[1];
    float_union_bytes[2] = FloatUnion.AsByte[2];
    float_union_bytes[3] = FloatUnion.AsByte[3];
  
    VT100_Print_Hex(float_union_bytes[0], 42, 40, F_BLUE, B_BLACK, 1, 0);
    VT100_Print_Hex(float_union_bytes[1], 43, 40, F_BLUE, B_BLACK, 1, 0);
    VT100_Print_Hex(float_union_bytes[2], 44, 40, F_BLUE, B_BLACK, 1, 0);
    VT100_Print_Hex(float_union_bytes[3], 45, 40, F_BLUE, B_BLACK, 1, 0);

The VT100_Print_hex() will print a value as hex on terminal, this is tested and working.

So when i compile the above, i get this error.

Code:

main.c:60:45: error: subscripted value is neither array nor pointer nor vector
     float_union_bytes[0] = FloatUnion.AsByte[0];

I dont know what its asking me for and to me, it looks correct.
Can anyone help me with what i am missing, please.

Originally Posted by Salem

Try it with

Code:

union
  {
    float AsFloat;
    char AsByte[4];
  } FloatUnion;

This is identical to what I have already...

Originally Posted by john.c

If that's true then why did you show something different?

Ahhhh. My mistake. I see whats different now. I am missing [4]. Its been a long day

I will try this tomorrow when I'm on the PC and let you know if it worked.

Originally Posted by Salem

Try it with

Code:

union
  {
    float AsFloat;
    char AsByte[4];
  } FloatUnion;

This worked. Thank you very much

Originally Posted by awsdert

Personally I prefer to use sizeof(float) rather than using 4 directly, then instead of directly referencing the bytes I would use a loop to access them, just gives that piece of mind that all bytes are covered in the unlikely event a float is not 4 bytes, would be a strange setup but it's not impossible, I've read of a setup where all integers besides char were 8 bytes long so it's not out of the question

Indeed. Some systems might have only one floating-point type so that float is the same as double. Some systems might even have 16- or 32-bit char, which makes sizeof (float) equal 2 or 1. Using sizeof is the safest way to go.