I was trying to develop an executable which has standard functionality at product level and additional custom functionality at solution level.
I wrote class A at product level and class B derived from class A at solution level. There is an Init() function at both levels and I want to call the Init() function of class B if executable is built from solution level. Also note main function resides in product level.
My code is listed below;
PRODUCT LEVEL
A.h at product level
#include <iostream>
#define App A::GetInstance()
class A
{
public:
A(int priority = 1);
virtual ~A(){}
virtual void Init() { printf("Called Class A function \n"); }
static A& GetInstance() { return *_instance;}
private:
static A* _instance;
static int _priority;
};
A.cpp at product level
#include "A.h"
A* A::_instance = new A();
int A::_priority = 0;
A::A(int priority)
{
if (priority > _priority)
{
_instance = this;
_priority = priority;
}
}
main.cpp at product level
#include "A.h"
int main(int argc, char** argv)
{
App.Init();
return 0;
}
SOLUTION LEVEL
B.h at solution level
#include "A.h"
class B : public A
{
public:
B(int priority = 2);
virtual ~B(){}
virtual void Init() { printf("Called Class B function \n"); }
};
B.cpp at solution level
#include "B.h"
B* b = new B();
B::B(int priority):A(priority)
{
}
My questions are?
- Instead of global objects is there a better approach to this?
Also if I change the make order of the object files then different output is produced with the same code? is there a way to overcome this?
a.) g++ -o test A.cpp B.cpp main.cpp
./test ------- >outputs ----> Called Class A function
b.) g++ -o test B.cpp A.cpp main.cpp
./test ------- >outputs ----> Called Class B function
Is my implementation is the ideal method for n number of levels? (e.g. solutions that build from another solution)
