Why do I get two different results over a string and a vector

Question

I have two strings s1: "abcd" and s2: "ab" and two int vectors v1: 12,13,14,15 and v2: 12,13 . I am calculating the distance between the two strings and the two vectors. I have got the code from the web for a c file to calculate the distance between the two strings. I am modifying the code in a way to calculate the distance between two vectors, but in c++ . The code is working, but the result are not the same (the distance should be the same for both calculations). The problem comes from:

double transpositions = 0.0;    
for (i = 0; i < s1_len; i++) {
    if (!s1_matches[i]) continue;
    while (!s2_matches[k]) k++;
    if (s1[i] == s2[k]) transpositions++;
    k++;
}

The s1_len is s1_len(s1) (the same for s2) and the s1_match is int *s1_matches = (int*) calloc(s1_len, sizeof(int)) . I am running the c code with correct answers, but when I change the code to c++ and change the string variables to vector variables, the number that I get for transposition is different. To change the string to vector, I get the size of the vector ( v1_len ) by v1.size() and the v1_match by int *v1_matches = (int*) calloc(v1_len, sizeof(int)) . The code is exactly the same:

double transpositions = 0.0;    
for (i = 0; i < v1_len; i++) {
    if (!v1_matches[i]) continue;
    while (!v2_matches[k]) k++;
    if (v1[i] == v2[k]) transpositions++;
    k++;
}

What is the problem here?

Answer 1

You do not seem to do any bounds checking on s2_matches or v2_matches . This will have undefined behaviour when s1 and s2 are not identical:

while (!s2_matches[k]) k++;
if (s1[i] == s2[k]) transpositions++;
k++;

ie You never check for k overflowing the bounds of s2 or s2_matches . Same with the vector version.