Segmentation fault in test cases

Question

So, the question is as follows: Given any array(reasonably large) of integers, return the maximum difference between any two elements in the array such that the larger element occurs at a higher index than the smaller element. Return -1, if no such pair is found. Example:

7 2 3 10 2 4 8 1

where the first element is the size of the array(or the number of lines being entered), and the rest are the elements. Sample output is 8(10-2) for the above.

My code is as follows:

int A[20],size;

scanf("%d",&size);
for(int i=0;i<size;i++){
scanf("%d\n",&A[i]);
}
int diff = A[1]-A[0];
int currsum = diff;
int maxsum = currsum;

for(int i=1; i<size-1; i++)
{
    // Calculate current difference for the loop
    diff = A[i+1]-A[i];

    // Calculate current sum for the loop
    if (currsum > 0)
       currsum += diff;
    else
       currsum = diff;

    // Update max sum(if needed)
    if (currsum > maxsum)
       maxsum = currsum;
}

printf("%d",maxsum);

This is a question from Hackerrank, but it runs for only three out of 10 possible testcases. The rest of the cases return a segmentation fault. Any idea would be helpful.

Answer 1

As mentioned in the comments, you've declared A to hold just 20 integers. But the question can send up to 1,000,000 integers. That's the mistake!

Answer 2

Using pointers make this more important. First declare A as pointer of integers, then, read the first element of the array, using this integer you can allocate memory dynamically ( malloc() or calloc() function) for your array A . so the size of A will dynamic and you can resize it in function of the first element.