• State Not Answered
  • Locked Locked
  • Replies 19 replies
  • Subscribers 22 subscribers
  • Views 4751 views
  • Users 0 members are here
Options
Related
How was your experience today?
This discussion has been locked.
You can no longer post new replies to this discussion. If you have a question you can start a new discussion

Initialize long with floats binary representation

Jürgen Breuer

Hello,

how can I achieve


long l = WHAT_EVER_MACRO (-12.5);

main()
{
  printf("%X",l);
}

resulting in output "0xC1480000" the binary representation of the float -12.5 .
Without that MACRO the value is 0xFFFFFFF4 = -12 for integers .

Because of given circumstances in an existing project I have to store float values during initialization in long variables.

Is there any possibility?

Thanks in advance

Jürgen

  • 0 in reply to Jürgen Breuer

    As I did write earlier, the floats will not take their binary representation until (not before) they are stored in memory.

    If the value is already available in a variable, it's possible to use a type cast of a pointer to that variable. The other option is to assign the value to a union and later read out the value using one of the other data elements sharing the union.

    The next alternative is to move all such floating point constants into a separate header file with named constants that you preprocess into their binary representation.

  • 0 in reply to ImPer Westermark

    Now I got you!

    I tryed

    float f32ConstInitval = -12.5;
uint32_t ui32Testvalue = (*(uint32_t *)&f32ConstInitval);


    and got: 'scalar': initializer is not a constant :-(

    With

    static const float f32ConstInitval = -12.5;
uint32_t ui32Testvalue = f32ConstInitval);


    I got ui32Testvalue = 0xFFFFFFF4 as I asked my debugger :-(

    but with

    static const float f32ConstInitval = -12.5;
uint32_t ui32Testvalue = (*(uint32_t *)&f32ConstInitval);

    I yielded my desired ui32Testvalue = 0xC1480000 :-D
    with a litte overhead of 4 Bytes in far flash.

    Thanks a lot

    Jürgen

  • 0 in reply to Jürgen Breuer

    or to make a pointer (void or unsigned long) and points it to the floater.

  • 0 in reply to Jürgen Breuer

    We use this construction since several years for different integer data types (8, 16, 32 bits signed and unsigned) but now it is the first time to use a float variable.

    And that's your problem right there. That construction is wrong, and has been for all those years. It relies on the incorrect assumption that you can stuff whatever data you have into a 32-bit integer and recover it unharmed later. The right way of doing that would been to use a union all along, or simply a buffer of bytes.

    Therefor my question for a computing macro or a compiler directive not to convert the float to an integer.

    No such construction can possibly yield a compile-time constant, which is what you would need for initializing a static variable.