Inheritance with CRTP - c++

I have these 3 classes.
class A
{
public:
virtual void Func() = 0;
};
template<class T>
class B : public A
{
public:
void Func()
{
cout << "In B" << endl;
static_cast<T*>(this)->Func();
}
};
class C : public B<C>
{
public:
void Func()
{
cout << "In C" << endl;
}
};
And, I do this:
int main(int argc, char **argv)
{
A *a = new C;
a->Func();
return 0;
}
And it prints : "In C".
If I do this,
int main(int argc, char **argv)
{
B<C> *a = new C;
a->Func();
return 0;
}
It again prints "In C"
What is going on?

You're calling a virtual member function of a class C object who has overloaded this function. It calls the function in class C.
Furthermore, this is not CRTP as the templated class B does not inherit from the class template parameter.

Func is virtual, a is a pointer to an instance of C, so C's version of Func is called.

The code is not complete, add #include and "using namespace std;". More importantly, you get the desired behaviour by removing the virtual function declaration in A.
In general, the main reason to use CRTP is to let the template know the type it receives and avoid doing a virtual call (or better, avoid making the method virtual).
template <typename T>
class ClassUsingSomething {
public:
void method1() {
// I need to call method2, I do this by casting, so it doesn't need to be virtual.
static_cast<T *>(this)->method2();
}
};
class A: public ClassUsingSomething<A> {
public:
void method2() {
//do something
}
};

Related

Method nonvisibility in base class despite definition in derived class; polymorphism and using `virtual` keyword

#include <iostream>
class A {
protected:
int foo;
};
class B : public A {
public:
B(int bar) { foo = bar; }
int method() { return foo; }
};
class C {
private:
A baz;
public:
C(A faz) { baz = faz; }
A get() { return baz; }
};
int main(void) {
C boo(B(1));
std::cout << boo.get().method() << std::endl;
return 0;
}
I have a base class A which B is a derived class of. Class C takes an A yet I have passed a derived class (B) in its place. No warnings or errors passing a B to C, but I'd like to have method visibility of method() in the above situation.
I'm not very familiar with virtual but I did try to add virtual int method() = 0; to A which lead to further errors.
Consider were I to add a second derived class:
class D : public A {
public:
D(int bar) { foo = bar; }
int method() { return foo+1; }
};
I'd like C to be able to take either B or D and my best assumption would be to take an A and let it handle it.
How do I use polymorphism correctly in this fashion?
Expected output with the below:
int main(void) {
C boo(B(1));
C boz(D(2));
std::cout << boo.get().method() << std::endl;
std::cout << boz.get().method() << std::endl;
return 0;
}
Would be:
1
3
First of all, in order to use A polymorphically, you need to add a virtual destructor, otherwise you will run into undefined behavior when trying to destroy the object. Then the method that you want to call through A must be virtual as well. If it shouldn't have an implementation in the base class itself, make it pure virtual:
class A {
protected:
int foo;
public:
virtual ~A() {}
virtual int method() = 0;
};
Then in C you need to use pointers or references to A, since polymorphism only works with those.
If you want C to own the A, as your code example to suggest, then you need to provide a destructor deleting the pointer and you need to disable copying of the class (or decide on some useful semantics for it):
class C {
private:
C(const C&); // Don't allow copying
C& operator=(const C&); // Don't allow copying
A* baz;
public:
C(A* faz) : baz(faz) { }
~C() { delete baz; }
A& get() { return *baz; }
};
int main(void) {
C boo(new B(1));
C boz(new D(2));
std::cout << boo.get().method() << std::endl;
std::cout << boz.get().method() << std::endl;
return 0;
}
Ideally you would upgrade to C++11 and use std::unique_ptr<A> instead of A* as member. But even if you can't do that, consider using boost::scoped_ptr<A>, which will manage the deletion for you (you don't need the destructor) and will make the class non-copyable by default. It also provides better exception-safety to encapsulate allocations in smart pointers like that.
If you need to call method() of type B using base class type A there has to be lookup during the runtime. The lookup is necessary to answer the question: Which method should be called? - the one that corresponds the type in a current line? Or other method in inheritance hierarchy?" If you expect method() from class B to be called when you have pointer or reference to A then you have to create a lookup table. This table is called vtable (from virtual functions table) and it's defined by adding virtual keyword to functions.
#include <iostream>
class A {
public:
virtual ~A(){}
virtual int method() = 0;
protected:
int foo;
};
class B : public A {
public:
B(int bar) { foo = bar; }
int method() {
std::cout << "Calling method() from B" << std::endl;
return foo; }
};
class C {
private:
A* baz;
public:
C(A* faz) { baz = faz; }
A* get() { return baz; }
};
int main(void) {
A* element = new B(1);
C boo(element);
boo.get()->method();
return 0;
}
It prints "Calling method() from B". Please keep in mind that the code is for presentation purposes and it's not good from best practices perspective.

Virtual function behavior in derived classes

Can I have a virtual function in the base class and some of my derived classes do have that function and some don't have.
class A{
virtual void Dosomething();
};
class B : public A{
void Dosomething();
};
class C : public A{
//Does not have Dosomething() function.
};
From one of my c++ textbook:
Once a function is declared virtual, it remains virtual all the way down the inheritance, even if the function is not explicitly declared virtual when the derived class overrides it.
When the derived class chooses not to override it, it simply inherits its base class's virtual function.
Therefore to your question the answer is No. Class c will use Class A's virtual function.
Derived classes do not have to implement all the virtual functions, unless it is a pure virtual function. Even in this case, it will cause an error only when you try to instantiate the derived class( without implementing the pure virtual function ).
#include <iostream>
class A{
public :
virtual void foo() = 0;
};
class B: public A{
public :
void foo(){ std::cout << "foo" << std::endl;}
};
class C: public A{
void bar();
};
int main() {
//C temp; The compiler will complain only if this is initialized without
// implementing foo in the derived class C
return 0;
}
I think the closest you might get, is to change the access modifier in the derived class, as depicted below.
But, I would consider it bad practice, as it violates Liskov's substitution principle.
If you have a situation like this, you might need to reconsider your class design.
#include <iostream>
class A {
public:
virtual void doSomething() { std::cout << "A" << std::endl; }
};
class B : public A {
public:
void doSomething() override { std::cout << "B" << std::endl; };
};
class C : public A {
private:
void doSomething() override { std::cout << "C" << std::endl; };
};
int main(int argc, char **args) {
A a;
a.doSomething();
B b;
b.doSomething();
C c;
//c.doSomething(); // Not part of the public interface. Violates Liskov's substitution principle.
A* c2 = &c;
c2->doSomething(); // Still possible, even though it is private! But, C::doSomething() is called!
return 0;
}

Understanding pure-virtual functions

The following code compiles fine:
struct A
{
const int a;
virtual void foo() = 0;
};
struct B : A{ };
void A::foo(){ std::cout << "foo" << std::endl; }
DEMO
The thing is the struct A is an abstract therefore we can't instanciate it. But we can subclass it and
struct A
{
const int a;
virtual void foo() = 0;
};
struct B : A{ };
void A::foo(){ std::cout << "foo" << std::endl; }
int main(int argc, char ** argv)
{
B b;
b.foo(); //error: implement pure-virtual
}
DEMO
still can't use the A's implementation of foo and I suspect it will never called. So, I have no idea about application of such definition... Yes, it's useful to provide a definition for a virtual destructors, but that's not the case.
Where the definition of pure-virtuals can be used?
You can call it explicitly.
struct A
{
const int a;
virtual void foo() = 0;
};
struct B : A
{
void foo();
};
void A::foo()
{
std::cout << "A::foo" << std::endl;
}
void B::foo()
{
A::foo(); // here
std::cout << "B::foo" << std::endl;
}
int main(int argc, char ** argv)
{
B b;
b.foo(); // prints A::foo followed by B::foo
}
But you can (and MUST) implement (if B is to be instantiable) an override for A's void foo(). And THAT implementation CAN (but definitely not required to) call the BASE implementation:
struct B : public A
{
virtual void foo() {A::foo();}
};
I HAVE implemented this scenario, where I had a very simple "base" implementation, but required all leafs to ACTIVELY decide to utilize this base (or not) by chaining back to this common implementation.
You can call it from within B. And since A::foo() is pure virtual, B needs to define foo:
struct A {
virtual void foo() = 0;
};
void A::foo(){ std::cout << "A::foo() ran" << std::endl; }
struct B : A {
void foo() {
A::foo(); // <--
std::cout << "B::foo() ran" << std::endl;
}
};
int main(int argc, char ** argv)
{
B b;
b.foo();
}
This is how you implement Interfaces in C++. You force subclasses to implement that method to be instantiated, and you operate polymorphically in your own code on that abstract base class, leaving the specific implementation to clients of your interface.
As far as the implementing in base class goes, I can only think of two reasons. One, to allow them to not have to special case destructors. And two, B can define its own implementation in terms of A::foo(), IE, a default implementation.

Implement two functions with the same name but different, non-covariant return types due to multiple abstract base classes

If I have two abstract classes defining a pure virtual function with the same name, but different, non-covariant return types, how can I derive from these and define an implementation for both their functions?
#include <iostream>
class ITestA {
public:
virtual ~ITestA() {};
virtual float test() =0;
};
class ITestB {
public:
virtual ~ITestB() {};
virtual bool test() =0;
};
class C : public ITestA, public ITestB {
public:
/* Somehow implement ITestA::test and ITestB::test */
};
int main() {
ITestA *a = new C();
std::cout << a->test() << std::endl; // should print a float, like "3.14"
ITestB *b = dynamic_cast<ITestB *>(a);
if (b) {
std::cout << b->test() << std::endl; // should print "1" or "0"
}
delete(a);
return 0;
}
As long as I don't call C::test() directly there's nothing ambiguous, so I think that it should work somehow and I guess I just didn't find the right notation yet. Or is this impossible, if so: Why?
Okay, it is possible, and the way isn't too ugly. I have to add an additional level of inheritance:
ITestA ITestB <-- These are the interfaces C has to fulfill, both with test()
| |
ITestA_X ITestB_X <-- These classes implement the interface by calling a
| | function with a different name, like ITestA_test
\__________/ which is in turn pure virtual again.
|
C <-- C implements the new interfaces
Now C has no function test(), but when casting a C* to an ITestA*, the implementation of test() in ITestA_test will be used. When casting it to an ITestB*, even by a dynamic_cast from the ITestA*, the implementation of ITestB_test will be used.
The following program prints:
3.14
0
#include <iostream>
class ITestA {
public:
virtual ~ITestA() {};
virtual float test() =0;
};
class ITestB {
public:
virtual ~ITestB() {};
virtual bool test() =0;
};
class ITestA_X : public ITestA {
protected:
virtual float ITestA_test() =0;
virtual float test() {
return ITestA_test();
}
};
class ITestB_X : public ITestB {
protected:
virtual bool ITestB_test() =0;
virtual bool test() {
return ITestB_test();
}
};
class C : public ITestA_X, public ITestB_X {
private:
virtual float ITestA_test() {
return 3.14;
}
virtual bool ITestB_test() {
return false;
}
};
int main() {
ITestA *a = new C();
std::cout << a->test() << std::endl;
ITestB *b = dynamic_cast<ITestB *>(a);
if (b) {
std::cout << b->test() << std::endl;
}
delete(a);
return 0;
}
Does this have any drawbacks you could think of?
When you declare ITestA *a = new C(), you have created a C object. If you invoke test on it with a->test(), it has to use the C virtual table to find the code to execute. But C is trying to have two different implementations of the same signature test(), which isn't allowed. The fact that you declared it as an ITestA * doesn't affect the method resolution. You declared the methods virtual, so they are found via the actual type of the object, regardless of the type of the pointer you used to access it.
You cannot have two methods with the same name and argument types. You'll need to find another way to structure this code.
I don't think this is possible. Functions overload by name (and signature).

Virtual inheritance and static inheritance - mixing in C++

If you have something like this:
#include <iostream>
template<typename T> class A
{
public:
void func()
{
T::func();
}
};
class B : public A<B>
{
public:
virtual void func()
{
std::cout << "into func";
}
};
class C : public B
{
};
int main()
{
C c;
c.func();
return 0;
}
Is func() dynamically dispatched?
How could you implement class A such that if B has a virtual override, that it is dynamically dispatched, but statically dispatched if B doesn't?
Edit: My code didn't compile? Sorry guys. I'm kinda ill right now. My new code also doesn't compile, but that's part of the question. Also, this question is for me, not the faq.
#include <iostream>
template<typename T> class A
{
public:
void func()
{
T::func();
}
};
class B : public A<B>
{
public:
virtual void func()
{
std::cout << "in B::func()\n";
}
};
class C : public B
{
public:
virtual void func() {
std::cout << "in C::func()\n";
}
};
class D : public A<D> {
void func() {
std::cout << "in D::func()\n";
}
};
class E : public D {
void func() {
std::cout << "in E::func()\n";
}
};
int main()
{
C c;
c.func();
A<B>& ref = c;
ref.func(); // Invokes dynamic lookup, as B declared itself virtual
A<D>* ptr = new E;
ptr->func(); // Calls D::func statically as D did not declare itself virtual
std::cin.get();
return 0;
}
visual studio 2010\projects\temp\temp\main.cpp(8): error C2352: 'B::func' : illegal call of non-static member function
visual studio 2010\projects\temp\temp\main.cpp(15) : see declaration of 'B::func'
visual studio 2010\projects\temp\temp\main.cpp(7) : while compiling class template member function 'void A<T>::func(void)'
with
[
T=B
]
visual studio 2010\projects\temp\temp\main.cpp(13) : see reference to class template instantiation 'A<T>' being compiled
with
[
T=B
]
I'm not sure I understand what you're asking, but it appears you are missing the essential CRTP cast:
template<class T>
struct A {
void func() {
T& self = *static_cast<T*>(this); // CRTP cast
self.func();
}
};
struct V : A<V> { // B for the case of virtual func
virtual void func() {
std::cout << "V::func\n";
}
};
struct NV : A<NV> { // B for the case of non-virtual func
void func() {
std::cout << "NV::func\n";
}
};
If T does not declare its own func, this will be infinite recursion as self.func will find A<T>::func. This is true even if a derived class of T (e.g. DV below) declares its own func but T does not.
Test with different final overrider to show dispatch works as advertised:
struct DV : V {
virtual void func() {
std::cout << "DV::func\n";
}
};
struct DNV : NV {
void func() {
std::cout << "DNV::func\n";
}
};
template<class B>
void call(A<B>& a) {
a.func(); // always calls A<T>::func
}
int main() {
DV dv;
call(dv); // uses virtual dispatch, finds DV::func
DNV dnv;
call(dnv); // no virtual dispatch, finds NV::func
return 0;
}
How could you implement class A such that if B has a virtual override, that it is dynamically dispatched, but statically dispatched if B doesn't?
Somewhat contradictory, isn't it? A user of class A may know nothing about B or C. If you have a reference to an A, the only way to know if func() needs dynamic dispatch is to consult the vtable. Since A::func() is not virtual there is no entry for it and thus nowhere to put the information. Once you make it virtual you're consulting the vtable and it's dynamic dispatch.
The only way to get direct function calls (or inlines) would be with non-virtual functions and no indirection through base class pointers.
Edit: I think the idiom for this in Scala would be class C: public B, public A<C> (repeating the trait with the child class) but this does not work in C++ because it makes the members of A<T> ambiguous in C.
In your particular example, there's no need for dynamic dispatch because the type of c is known at compile time. The call to B::func will be hard coded.
If you were calling func through a B*, then you would be calling a virtual function. But in your highly contrived example, that would get you to B::func once again.
It doesn't make much sense to talk about dynamic dispatch from an A* since A is a template class - you can't make a generic A, only one that is bound to a particular subclass.
How could you implement class A such that if B has a virtual override, that it is dynamically dispatched, but statically dispatched if B doesn't?
As others have noticed, it's really hard to make sense of that question, but it made me remember something I have learned a long time ago, so here's a very long shot at answering your question:
template<typename Base> class A : private Base
{
public:
void func()
{
std::count << "A::func";
}
};
Given this, it depends on A's base whether func() is virtual. If Base declares it virtual then it will be virtual in A, too. Otherwise it won't. See this:
class V
{
public:
virtual void func() {}
};
class NV
{
};
class B : public A<V> // makes func() virtual
{
public:
void func()
{
std::count << "B::func";
}
};
class C : public A<NV> // makes func() non-virtual
{
public:
void func()
{
std::count << "C::func";
}
};
Would this happen to answer your question?
Whether the function is dynamically dispatched or not depends on two things:
a) whether the object expression is a reference or pointer type
b) whether the function (to which overload resolution resolves to) is virtual or not.
Coming to your code now:
C c;
c.func(); // object expression is not of pointer/reference type.
// So static binding
A <B> & ref = c;
ref.func(); // object expression is of reference type, but func is
// not virtual. So static binding
A<D>* ptr = new D;
ptr->func(); // object expression is of pointer type, but func is not
// virtual. So static binding
So in short, 'func' is not dynamically dispatched.
Note that :: suppresses virtual function call mechanism.
$10.3/12- "Explicit qualification with
the scope operator (5.1) suppresses
the virtual "call mechanism.
The code in OP2 gives error because the syntax X::Y can be used to invoke 'Y' in the scope of 'X' only if 'Y' is a static member in the scope of 'X'.
Seems you just had to add a little trace and usage to answer your own question...
#include <iostream>
template<typename T> struct A {
void func() {
T::func();
}
};
struct B1 : A<B1> {
virtual void func() {
std::cout << "virtual void B1::func();\n";
}
};
struct B2 : A<B2> {
void func() {
std::cout << "void B2::func();\n";
}
};
struct C1 : B1 { };
struct C2 : B2 { };
struct C1a : B1 {
virtual void func() {
std::cout << "virtual void C1a::func();\n";
}
};
struct C2a : B2 {
virtual void func() {
std::cout << "virtual void C2a::func();\n";
}
};
int main()
{
C1 c1;
c1.func();
C2 c2;
c2.func();
B1* p_B1 = new C1a;
p_B1->func();
B2* p_B2 = new C2a;
p_B2->func();
}
Output:
virtual void B1::func();
void B2::func();
virtual void C1a::func();
void B2::func();
Conclusion: A does take on the virtual-ness of B's func.