Storing large numbers

[Mahkoe]

I realize there are many ways to store large numbers in c++ (and by large I mean too large for long doubles), but none of them were really what I was looking for. I don't just need to store large numbers in which case I would use GMP or something, I need to store them into char arrays. Right now I have a snippet of code that makes logical sense but in fact doesn't work,

char* cp;
cp = new char [32]
*cp = 123456789345123452345345634563456234;
//Doesn't work when I put an L at the end of the number either.

Also, I'm one of those weirdos that needs to know how the code they are using works, so please don't say I can still manage this by screwing around with bigint libraries, I would actually like to know how to do it like this.

Thanks a lot in advance

[Narue]

Right now I have a snippet of code that makes logical sense but in fact doesn't work

Actually, that snippet makes no sense at all. If you're looking to write a bigint library, then your best bet would be to find a simple one and study the code. A good start is to implement the most basic arithmetic in the simplest manner possible. For example:

#include <algorithm>
#include <iostream>
#include <ostream>
#include <string>

namespace jsw {
    class integer {
        std::string _base;
    public:
        integer(const char *init = "0");
        integer(const std::string& init = "0");
        integer add(const integer& rhs);
        friend std::ostream& operator<<(std::ostream& out, const integer& obj);
    };
    
    integer::integer(const char *init)
        : _base(init)
    {}

    integer::integer(const std::string& init)
        : _base(init)
    {}

    integer integer::add(const integer& rhs)
    {
        using std::string;
        
        string::size_type x = _base.size() - 1;
        string::size_type y = rhs._base.size() - 1;
        int carry = 0;
        string result;
        
        // Add two digits while there are more
        while (x < string::npos && y < string::npos) {
            int sum = carry + (_base[x--] - '0') + (rhs._base[y--] - '0');
            carry = sum / 10;
            result.push_back(sum % 10 + '0');
        }
        
        // Fill in the rest
        while (x < string::npos) {
            int sum = carry + (_base[x--] - '0');
            carry = sum / 10;
            result.push_back(sum % 10 + '0');
        }
        
        while (y < string::npos) {
            int sum = carry + (rhs._base[y--] - '0');
            carry = sum / 10;
            result.push_back(sum % 10 + '0');
        }
        
        // The result was built backward, fix it
        std::reverse(result.begin(), result.end());
        
        return integer(result);
    }

    std::ostream& operator<<(std::ostream& out, const integer& obj)
    {
        return out << obj._base;
    }
}

int main()
{
    jsw::integer a = "123", b = "456";
    
    std::cout << a << " + " << b << " = " << a.add(b) << '\n';
}

Writing a big number library isn't inherently difficult; making it efficient is where the real challenge lies, and that part can get pretty complex.

[Mahkoe]

Okay, this'll work too, thanks