False positive: Either the condition is redundant or there is pointer arithmetic with NULL pointer

[raynebc]

Staring in Cppcheck 1.89 I'm seeing a false positive that I can reproduce with this minimal example:

void function(int value, char *string)
{
    char *ptr1 = NULL, *ptr2 = NULL;
    unsigned long count = 0;

    if(!string)
        return;
    ptr1 = string;
    ptr2 = strrchr(string, 'a');    //The last instance

    if(ptr2 == NULL)
        return; //No such instances

    while(ptr1 < ptr2)
    {   //For each character up to the last instance of 'a'
        count++;
        ptr1++;
    }
}

I know Cppcheck can't assume the programmer is using pointers properly in such an example (ie. both are pointers to the same string), but the warning doesn't seem accurate because both pointers are already checked for NULL.

[Daniel Marjamäki]

Thanks! I created https://trac.cppcheck.net/ticket/9315