Question

So, I am having trouble with some code.

I want this function to take in a byte array(testing with single byte for now), convert the byte into binary and then append it to a “1.” to use in a calculation.

ex:

ouput: 01110000 —-> 1.01110000

#include <stdlib.h>
#include <stdio.h>
#include <stdint.h>
#include <string.h>

double calcByteValue(uint8_t data[], int size);

int main() {
   uint8_t test[10];
   test[0] = 0x0e;

   double d = calcByteValue(test, 8);
   return 0;
}

double calcByteValue(uint8_t data[], int size) {
    int i;
    uint8_t bits[21];
    char binary[100];
    char str[100] = "1.";

    for (i = 0;i < size;i++) {                    
            bits[i] = (data[0] >> i) & 1;
            if (bits[i] == 0) {
                binary[i] = '0';
                printf("0(%d)\n", i);
            } else {
                binary[i] = '1';
                printf("1(%d)\n", i);
            }
    }

    strcat(str, binary);
    float d = atof(str);
    printf("%f\n", d);
    return 0;
    //return pow(-1, bits[0]) * pow(2, (128-127)) * atof(str));
}

Here is my output, for some reason it is going through the whole loop just fine, but only printing 6 of the original bits, knocking off the last couple ones. What am I doing wrong???

0(0)
1(1)
1(2)
1(3)
0(4)
0(5)
0(6)
0(7)
1.011100

Answer 1

First:

strcat(str, binary);

You never null terminated the binary array. You have to to use it as a string.

 char binary[100];

binary is defined at block scope, and block scope automatic objects that are not explicitly initialized have an indeterminate value.

Here is how to null terminate the array:

binary[size] = '\0';

Second:

you are inserting the bits values in the reverse order. Put (data[0] >> 7) & 1 for binary[0], (data[0] >> 6) & 1 for binary[1] and so on.

Also printf with %f conversion specification prints 6 digits after the decimal point. If you want more digits (e.g., 16), you can specify the precision like this: printf("%.16f\n", d);

You are also using the type float for object d, if you find you don’t have enough precision with float, you can use type double.