Source: http://projecteuler.net/index.php?section=problems&id=11

Quick overview: Take a 20x20 grid of numbers and compute the largest product of 4 pairs of numbers in either horizontal, vertical, or diagonal.

My current approach is to divide the 20x20 grid up into single rows and single columns and go from there with a much more manageable grid. The code I'm using to divide the rows into rows is

void fillRows
    ( string::const_iterator& fieldIter, 
      list<int>& rowElements, 
      vector<list<int>>& rows )
{
    int count(0);
    for( ; fieldIter < field.end(); ++fieldIter )
    {
        if(isdigit(field[*fieldIter]))
        {           
            rowElements.push_back(toInt(field[*fieldIter]));
            ++count;
        }
        if(count == 40)
        {
            rows.push_back(rowElements);
            count = 0;
            rowElements.clear();
        }
    }
}

Short explanation: I have the field set as static const std::string field and I am filling a vector with lists of rows. Why a list? Because the queue doesn't have a clear function. Also practice using STL container lists and not ones I write myself.

However, this thing isn't working. Oftentimes I see it omitting a character( function toInt parses the const char as int ) and I end up with 18 rows, two rows short of the 20x20 grid. The length of the rows seem good.

Rows: 18

RowElements[0]: 40 (instead of pairs I saved each number individually. Will fix that later)

What am I doing wrong?