Converting Byte String to Float

[Glen Singletary]

I am having difficulty casting a string to float. I am using PCW Compiler PCM version 3.033 and MPLAB IDE for win/16 version 5.50.0.

Consider this test program and the results I get when stepping through:

#include <16F877.h>
#device ICD=TRUE
#include <stdlib.h>
#fuses HS,NOWDT
#zero_ram

float ftemp;
int8 temp[4];

void main(void)
{
temp[0]=0x7f;
temp[1]=0x0;
temp[2]=0x0;
temp[3]=0x0;

ftemp = 1.0;
// RAM at temp = 7F 00 00 00
// RAM at ftemp = 7F 00 00 00
// Watch windows shows 1.000+00 for ftemp
ftemp = (float)*temp;
// RAM at temp = 7F 00 00 00
// RAM at ftemp = 85 7E 00 00
// Watch window shows 1.270+02 for ftemp
ftemp = *temp;
// same results...
}

Any ideas on why I am getting this result and how I can fix it?

thanks
___________________________
This message was ported from CCS's old forum
Original Post ID: 6494

[R.J.Hamlett]

:=I am having difficulty casting a string to float. I am using PCW Compiler PCM version 3.033 and MPLAB IDE for win/16 version 5.50.0.
:=
:=Consider this test program and the results I get when stepping through:
:=
:=#include <16F877.h>
:=#device ICD=TRUE
:=#include <stdlib.h>
:=#fuses HS,NOWDT
:=#zero_ram
:=
:=float ftemp;
:=int8 temp[4];
:=
:=void main(void)
:= {
:= temp[0]=0x7f;
:= temp[1]=0x0;
:= temp[2]=0x0;
:= temp[3]=0x0;
:=
:= ftemp = 1.0;
:= // RAM at temp = 7F 00 00 00
:= // RAM at ftemp = 7F 00 00 00
:= // Watch windows shows 1.000+00 for ftemp
:= ftemp = (float)*temp;
At this point, 'temp' contains the numeric sequence 7F00..., and you are now using this as a pointer!. You need to cast the _address_ of temp, to be a pointer to a float.

:= // RAM at temp = 7F 00 00 00
:= // RAM at ftemp = 85 7E 00 00
:= // Watch window shows 1.270+02 for ftemp
:= ftemp = *temp;
:= // same results...
:= }
:=
:=Any ideas on why I am getting this result and how I can fix it?
:=
:=thanks
You need to be more careful in dealing with pointers. The '&' character gets the address of temp (which you want), and this address, is then used with the '*' character as a pointer. Hence the 'full' expansion would be :

ftemp=*((float *)&temp);

which takes the address of 'temp', and converts this to a pointer to a float, and then uses this pointer to retrieve the value.
You can also use a union to get the same effect.

union {
int8 temp[4];
float ftemp;
} cstore;

cstore.temp[0]=0x7F;
cstore.temp[1]=cstore.temp[2]=cstore.temp[3]=0x00;

Now cstore.ftemp is the floating point value.

Best Wishes
___________________________
This message was ported from CCS's old forum
Original Post ID: 6497

[Glen Singletary]

:=:=I am having difficulty casting a string to float. I am using PCW Compiler PCM version 3.033 and MPLAB IDE for win/16 version 5.50.0.
:=:=
:=:=Consider this test program and the results I get when stepping through:
:=:=
:=:=#include <16F877.h>
:=:=#device ICD=TRUE
:=:=#include <stdlib.h>
:=:=#fuses HS,NOWDT
:=:=#zero_ram
:=:=
:=:=float ftemp;
:=:=int8 temp[4];
:=:=
:=:=void main(void)
:=:= {
:=:= temp[0]=0x7f;
:=:= temp[1]=0x0;
:=:= temp[2]=0x0;
:=:= temp[3]=0x0;
:=:=
:=:= ftemp = 1.0;
:=:= // RAM at temp = 7F 00 00 00
:=:= // RAM at ftemp = 7F 00 00 00
:=:= // Watch windows shows 1.000+00 for ftemp
:=:= ftemp = (float)*temp;
:=At this point, 'temp' contains the numeric sequence 7F00..., and you are now using this as a pointer!. You need to cast the _address_ of temp, to be a pointer to a float.
:=
:=:= // RAM at temp = 7F 00 00 00
:=:= // RAM at ftemp = 85 7E 00 00
:=:= // Watch window shows 1.270+02 for ftemp
:=:= ftemp = *temp;
:=:= // same results...
:=:= }
:=:=
:=:=Any ideas on why I am getting this result and how I can fix it?
:=:=
:=:=thanks
:=You need to be more careful in dealing with pointers. The '&' character gets the address of temp (which you want), and this address, is then used with the '*' character as a pointer. Hence the 'full' expansion would be :
:=
:=ftemp=*((float *)&temp);

Using this expression gave me an "Expecting and Identifier" error because temp was declared as a string. Modifying this slightly to:

ftemp=*((float *)&temp[0]);
or
ftemp=*((float *)temp);

gave the correct results.

:=
:=which takes the address of 'temp', and converts this to a pointer to a float, and then uses this pointer to retrieve the value.
:=You can also use a union to get the same effect.
:=
:=union {
:= int8 temp[4];
:= float ftemp;
:=} cstore;
:=
:=cstore.temp[0]=0x7F;
:=cstore.temp[1]=cstore.temp[2]=cstore.temp[3]=0x00;
:=
:=Now cstore.ftemp is the floating point value.
:=
:=Best Wishes

Thanks for the help.
___________________________
This message was ported from CCS's old forum
Original Post ID: 6518

[R.J.Hamlett]

:=<snipped>
:=:=ftemp=*((float *)&temp);
:=
:=Using this expression gave me an "Expecting and Identifier" error because temp was declared as a string. Modifying this slightly to:
:=
:=ftemp=*((float *)&temp[0]);
:=or
:=ftemp=*((float *)temp);
:=
:=gave the correct results.
At least you saw 'where I was coming from', in that you need the address of the variable, not the contents. :-)
I tend to 'under proofread' what I type at times, so this sort of error is common....

Best Wishes
___________________________
This message was ported from CCS's old forum
Original Post ID: 6519