سؤال

Note: This is almost a duplicate of this entry: Abstract classes and Pointers

I need to create a vector of virtual classes. Here the idea:

#include <vector>
using namespace std;

class VirtualFoo {
protected:
    VirtualFoo();
    virtual ~VirtualFoo();
public:
    virtual void doStuff()=0;
};

class ConcreteFoo: public VirtualFoo {
public:
    ConcreteFoo(double a);
    virtual ~ConcreteFoo();
    void doStuff();
private:
    double a;
};

class Foo {
public:
    Foo(std::vector<VirtualFoo> foos);
    virtual ~Foo();
    void doAllStuff();
protected:
    std::vector<VirtualFoo> foos;
};

And I would like to use it this way:

int main(int argc, char *argv[]){

    std::vector<ConcreteFoo> cfoos;
    cfoos.push_back(ConcreteFoo(1.0));
    cfoos.push_back(ConcreteFoo(2.0));

    Foo foo = Foo(cfoos);
    foo.doAllStuff();

}

Of course this does not work because cfoos is a vector of VirtualFoo and not ConcreteFoo.

Now, if I do not use a vector of VirtualFoo but a vector of VirtualFoo* and push back pointers to instances of ConcreteFoo, that seems to work fine.

Just, I am not sure it is the cleanest way to go. It's more like I did not think of any other way of doing this. Is doing this ok ?

هل كانت مفيدة؟

المحلول

The vector with pointers of std::vector<VirtualFoo*> cfoos; is fine. You cannot instantiate abstract classes, hence the compiler fails to instantiate or specialize vector template that use abstract class as value. Pointing elements with base class pointers is fine which is Liskov substitution principle(http://en.wikipedia.org/wiki/Liskov_substitution_principle)

I agree with the comments from others that shared pointers is better solution so that you don't have to worry about memory management:

std::vector<shared_ptr<VirtualFoo>> cfoos;  
cfoos.push_back( shared_ptr<VirtualFoo> (new ConcreteFoo(1.0)) );   
cfoos.push_back( shared_ptr<VirtualFoo> (new ConcreteFoo(2.0)) );
مرخصة بموجب: CC-BY-SA مع الإسناد
لا تنتمي إلى StackOverflow
scroll top