Types of C++ Inheritance: public protected private
2645 ワード
In C++ there are three types of inheritance: public protected private
Any of these three types of inheritance can be modified with the
The three access modifiers
public
When a base class is specified as
protected
When a base class is specified as
private
When a base class is specified as
Examples of how this plays out:
struct Y {public:void B() {}};
struct Z {public:void C() {}};
struct Q : public X, protected Y, private Z {public:void Test(){A();//OKB();//OKC();//OK}};
struct R : public Q {public:void Test2(){A();//OKB();//OKC();//NOT OK
Q t;Y *y = &t//OKZ *z = &t//NOT OK}};
int main(int argc, char **argv) {Q t1;t1.A();//OKt1.B();//NOT OKt1.C();//NOT OK
R t2;t2.A();//OKt2.B();//NOT OKt2.C();//NOT OK
X *x = &t1;//OKY *y = &t1;//NOT OKZ *z = &t1;//NOT OK
x = &t2;//OKy = &t2;//NOT OKz = &t2;//NOT OK
}
What about Virtual?
Oh right. Virtual is only useful when multiple inheritance is involved and the same class appears in the inheritance graph more than once. If the inheritance is declared as
Be careful! If the class is inherited sometimes as virtual and sometimes not, the virtual instances are merged and the non-virtual instances are not, giving you a mix of behavior.
Any of these three types of inheritance can be modified with the
virtual
keyword. In my experience interviewing candidates for c++ positions, I've learned that the average programmer does not know how these are used, or even what they mean. So I thought I would go over them here. The three access modifiers
public
, protected
and private
are analogous to the access modifieres used for class members. public
When a base class is specified as
public
ie: class c : public base {};
the base class can be seen by anyone who has access to the derived class. That is, any members inherited from base
can be seen by code accessing c
. protected
When a base class is specified as
protected
ie: class c : protected base {};
the base class can only be seen by subclasses of C
. private
When a base class is specified as
private
ie: class c : private base {};
the base class can only be seen by the class C
itself. Examples of how this plays out:
struct X {
public:
void A() {}
};
struct Y {public:void B() {}};
struct Z {public:void C() {}};
struct Q : public X, protected Y, private Z {public:void Test(){A();//OKB();//OKC();//OK}};
struct R : public Q {public:void Test2(){A();//OKB();//OKC();//NOT OK
Q t;Y *y = &t//OKZ *z = &t//NOT OK}};
int main(int argc, char **argv) {Q t1;t1.A();//OKt1.B();//NOT OKt1.C();//NOT OK
R t2;t2.A();//OKt2.B();//NOT OKt2.C();//NOT OK
X *x = &t1;//OKY *y = &t1;//NOT OKZ *z = &t1;//NOT OK
x = &t2;//OKy = &t2;//NOT OKz = &t2;//NOT OK
}
What about Virtual?
Oh right. Virtual is only useful when multiple inheritance is involved and the same class appears in the inheritance graph more than once. If the inheritance is declared as
virtual
all instances of the class are merged into one sub object and that sub object is initialized once. If the class that appears multiple times in the inheritance graph is NOT declared virtual
one sub object is created for EACH instance of the class and the class is initialized multiple times. Be careful! If the class is inherited sometimes as virtual and sometimes not, the virtual instances are merged and the non-virtual instances are not, giving you a mix of behavior.