Question

im trying to implement bit shifting to bigints. The BigInt is represented by an array of bytes that should be interpreted as a single integer N bits in two’s complement. So i want to make something like:

Example array bigint:
{0xFE, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF} would represent the integer 0xFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFE (-2) N = 128 bits.

typedef unsigned char *BigInt;

when i try to create my bigint im getting errors

#include <stdlib.h>

    void bi_init (int nbits)
    {
        nbytes = (nbits/8);
    }

    BigInt bi_new (int val)
    {
        BigInt new = (BigInt)malloc(nbytes*sizeof(unsigned char));

        new=val --> problem, can someone give me a hint on how can i implement this array?

        return novo;
    }

Answer 1

typedef unsigned char *BigInt;

BigInt *new = (BigInt)malloc(nbytes*sizeof(unsigned char));

You declare new as an unsigned char**, that’s one level of pointers too many (and you shouldn’t cast the result of malloc in C).

Assuming that novo and new are the same variable and only one occurrence has been translated for the post,

novo=val

overwrites the just allocated address with the passed-in val.

To fill the allocated buffer with the bytes of val,

int i = 0;
// make it unsigned, so that right-shifting works correctly
unsigned int u_val = val;
while(u_val) {
    new[i++] = u_val&0xFF;
    u_val >>= 8;
}

Then there remains the problem

nbytes = (nbits/8);

that you’re allocating too little memory if nbits is not a multiple of 8, make that

nbytes = (nbits+7)/8;