Question

I’ve been having some difficulties with duplicated code in one of my inheritance hierarchy.
How could I avoid duplicating the code inside function smile()?

Given the fact that the variable _a doesn’t exists in the base class I cannot move the function there. Also creating a template function like this template<typename T> void smile(T& a) { a++; } is not really a solution for me. My actual code is a bit more complicated and such a solution would be very messy if not impossible to apply with my current design.

class com
{
public:
   com(int x, float y) : _x(2), _y(1.15f)
   {   }
protected:
   // Common functions go here .We need this base class.
protected:
   int _x;
   float _y;
};

class com_int : public com
{
public:
   void fill()
   { _a = std::max(_x, (int)_y); }
protected:
   int _a;
};

class com_real : public com
{
public:
   void fill()
   { _a = std::min((float)_x, _y); }
protected:
   float _a;
};

class happy_int : public com_int
{
public:
   void smile() { _a ++; } // BAD: Will be duplicated
};

class happy_float : public com_real
{
public:
   void smile() { _a ++; } // BAD: Duplicated code
}

class sad_int : public com_int
{
public:
   frown() { _a --; }
}

Also, does anybody know a good book that teaches how to actually design code in C++ using OOP and template principles?

Answer 1

You can inherit from another helper template:

template <typename T, typename Derived> struct filler
{
    T _a;
    void fill()
    {
        com & b = static_cast<Derived&>(*this);
        _a = std::min(b._x, b._y);
    }
};

Usage:

struct com_int : com, filler<int, com_int> { };