Building a float from binary files c - c

Question

I'm trying to build a float number from a binary file with the following format:
first 4 bytes represents the number of digits on the left side of "."
We'll call that X
Next 4 bytes represents the number of digits on the right side of "."
We'll call that Y
Next byte represents the number's sign : "1" => '-', "0" => "+"
Next X+Y bytes represents the number's digits.
Example:
Data.bin:
4 2 0 2 8 3 9 5 3
The number should be : 2839.53
I've the structure :
typedef struct {
unsigned int digitsBefore;
unsigned int digitsAfter;
char sign;
char *digits;
} Number;
Function used to parse and fill the structure :
Number *fromBinaryToNumber(FILE *file){
int counter = 0, size = 0, i = 0;
size_t bytes_read;
unsigned int firstIndex, secondIndex;
char sign;
char *digits;
Number *nr = calloc(1, sizeof(Number));
// reading digits before
bytes_read = fread(&firstIndex, sizeof(unsigned int), 1, file);
nr->digitsBefore = firstIndex;
printf("Bytes read : %zu\n", bytes_read);
printf("First index %d\n", firstIndex);
size += firstIndex;
// reading digits after
bytes_read = fread(&secondIndex, sizeof(unsigned int), 1, file);
nr->digitsAfter = secondIndex;
printf("Bytes read : %zu\n", bytes_read);
printf("Second index %d\n", secondIndex);
size += secondIndex;
// reading sign
bytes_read = fread(&sign, sizeof(char), 1, file);
nr->sign = sign;
printf("Bytes read : %zu\n", bytes_read);
printf("Sign : %.2X\n", sign);
//reading digits
digits = calloc(size, sizeof(unsigned char));
nr->digits = calloc(size, sizeof(unsigned char));
printf("Digits : \n");
bytes_read = fread(digits, size, 1, file);
for(i = 0 ; i < size; i++){
printf("%.2X ", digits[i]);
}
memcpy(nr->digits, digits, size);
free(digits);
return nr;
}
For the following data in Data.bin (Hexdumped):
04 00 00 00 02 00 00 00 00 02 08 03 09 05 03
I get the output:
Bytes read : 1
First index : 4
Bytes read : 1
Second Index : 2
Bytes read : 1
Sign : 00
Digits :
02 08 03 09 05 03
Function to transform Number strunct in float:
double fromNumbertoDouble(Number *number){
int size = 0;
char *buff;
size = number->digitsAfter + number->digitsBefore + 2;
buff = calloc(size, sizeof(char));
if (!buff) {
fprintf(stderr, "Error calloc in fromNumberToDouble");
exit(1);
}
if (number->sign == '01') {
strcat(buff, "-");
}
strncat(buff, number->digits, number->digitsBefore);
strcat(buff, ".");
strncat(buff, number->digits+number->digitsBefore, number->digitsAfter);
return atof(buff);
}
This always return 0.
When I checked to see what is in the number returned from fromBinaryToNumber, number->digits = "\002\b\003\t\005\003".
I'm aware that atof can't convert something like that into a float.
My question is how can i convert the number->digits into something that would get converted into a float by atof.
Any other idea on how to solve this would also be appreciated.
Thank you for your time.

Answer 1

For each of the digits, add '0', i.e. the codepoint value for the digit 0. This turns the array into a a sequence of digit characters. Of course, it must also be 0-terminated for atof(), which requires a proper string.
You do realize that your format is fantastically awkward, right? A single float would use 4 bytes for the entire number, always. If you're going to convert to float anyway you're not going to be able to keep the arbitrary precision that your format supports.

Answer 2

First your buffer size is wrong - you've not taken into account the decimal point, sign or the zero-terminator. Also you don't initialise buff - so strcat may concatenate anywhere- you should either calloc, or memset the buff to 0 before strcat'ing anything to it.