Casting Pointers

[sidhomsam]

Question:

Is it possible to have a portion of the pointers within an Array of float pointers point to an int?

OR... is it possible to have a float pointer point to an int variable (possibly by typecasting the int to a float when initializing the "pointer / variable" association ?



Background / need:

I am writing a communication protocol (Johnson Control's N2) and it requires that I transmit a couple of int values as floats.

[sicarie]

You are aware that doing so would lose the precision of the 'float', correct?

You might consider looking at the structure of a float and breaking that down into two ints, so you could convert back and forth without losing that precision...

If you're not worried about precision, why not just put them into ints in the first place?

[Ganon11]

A float variable can hold an integer value, so you only need float pointers, I think.

[JosAH]

A float variable can hold an integer value, so you only need float pointers, I think.

Nitpick: a float variable can hold quite small int values; on most systems both
ints and floats take four bytes. A float has to store its exponent in there as well
so it can't store every possible mantissa ("significan t") that can be stored in a
simple int. Typically a float can store an int x where |x| < 2^24.

kind regards,

Jos

[sidhomsam]

Here is what I've tried:
[code=c]
void sub_example(voi d);

int int_variable;
float *float_pointer;

main()
{

int_variable = 4;

float_pointer = (float*)&int_va riable;

sub_example();

printf("%3.3f", *float_pointer) ;

}

void sub_example(voi d)
{
int_variable = 12;
}[/code]

But the output gives me a "0.000"

I was expecting / hoping for a "12.000"

[JosAH]

Here is what I've tried:
[code=c]
void sub_example(voi d);

int int_variable;
float *float_pointer;

main()
{

int_variable = 4;

float_pointer = (float*)&int_va riable;

sub_example();

printf("%3.3f", *float_pointer) ;

}

void sub_example(voi d)
{
int_variable = 12;
}[/code]

But the output gives me a "0.000"

I was expecting / hoping for a "12.000"

This has nothing to do with ints vs floats. Study the C/C++ language a bit more
before you attempt to do tricky numerical conversion things.

btw, the bit pattern for an int value 4 doesn't yield a float value of 4 ...

kind regards,

Jos

[kreagan]

I found this article while trying out your problem.

Void Pointers

Hopefully this is exactly what you needed. Plus it shows you how to use function pointers! Yummy!

[sidhomsam]

This has nothing to do with ints vs floats. Study the C/C++ language a bit more
before you attempt to do tricky numerical conversion things.

btw, the bit pattern for an int value 4 doesn't yield a float value of 4 ...

kind regards,

Jos

I understand that an int is stored as 16bits and a float is stored as 32bits - and a float pointer pointed to an int will probably grab the 16 bits of the int it is pointed at then the next 16 bits - which does not translate to a 32 bit float. I am asking of a way to solve this problem

btw - if you don't have a solution you should probably NOT reply instead of replying with a statement like "Study the C/C++ language a bit more
before you attempt to do tricky numerical conversion things"

[JosAH]

I understand that an int is stored as 16bits and a float is stored as 32bits - and a float pointer pointed to an int will probably grab the 16 bits of the int it is pointed at then the next 16 bits - which does not translate to a 32 bit float. I am asking of a way to solve this problem

btw - if you don't have a solution you should probably NOT reply instead of replying with a statement like "Study the C/C++ language a bit more
before you attempt to do tricky numerical conversion things"

You're most likely wrong about the 16 bits for an int or you're running a very old
system. Better dereference the float value from the float pointer first and cast it
to an int after. If you have an int pointer do the same but cast the int value to a
float after.

Better study the C/C++ language quite a bit more before you attempt to do
tricky numerical conversion stuff like the above.

Jos

[sidhomsam]

You're most likely wrong about the 16 bits for an int or you're running a very old
system. Better dereference the float value from the float pointer first and cast it
to an int after. If you have an int pointer do the same but cast the int value to a
float after.

Better study the C/C++ language quite a bit more before you attempt to do
tricky numerical conversion stuff like the above.

Jos

The program runs on a an embedded system (zilog Z8F6421) - and ints are 16bits, and floats are 32bits

btw embedded systems y require little more skill than programming on an operating system

[RRick]

Since your ints are much smaller than a float, you can store the int as a float and not worry about loosing any precision of the int.

When you want an int, just set the array value to an int and the compiler will take care of the conversion. You'll probably get a warning about loosing precision, but it should compile.
[code=cpp]int val = floatArray[99];
int val2 = (int) floatArray[102]; // Should stop compiler warnings[/code]
Your previous attempt to cast between ints and floats is not going to work. You can't cast between pointer types and expect a valid answer. Some numeric conversion needs to take place.

[sidhomsam]

I found this article while trying out your problem.

Void Pointers

Hopefully this is exactly what you needed. Plus it shows you how to use function pointers! Yummy!

KREAGAN!!!

That was it !! Void Pointers with casting did the trick!!!!

Thanks alot

here is what I ended up with:

void sub_example(voi d);


int int_variable;
void *float_pointer[2];
float sam=3.14159;
float sum_sam;


main()
{

int_variable = 4;

float_pointer[0] = &int_variabl e;
float_pointer[1] = &sam;

sub_example();

sum_sam = *(int*)float_po inter[0] + *(float*)float_ pointer[1];
printf("%f",sum _sam );

}

void sub_example(voi d)
{
int_variable = 12;

}





My final output was "15.14159"

Thanks again

[kreagan]

KREAGAN!!!

That was it !! Void Pointers with casting did the trick!!!!

Thanks alot

Yay! Very nice. Thanks for letting us know.

ps. be weary of incrementing void pointers