Friend classes
A friend class is a class that can access the private and protected members of another class.
Here is an example:
#include <iostream>
class Storage
{
private:
int m_nValue {};
double m_dValue {};
public:
Storage(int nValue, double dValue)
: m_nValue { nValue }, m_dValue { dValue }
{ }
// Make the Display class a friend of Storage
friend class Display;
};
class Display
{
private:
bool m_displayIntFirst {};
public:
Display(bool displayIntFirst)
: m_displayIntFirst { displayIntFirst }
{
}
// Because Display is a friend of Storage, Display members can access the private members of Storage
void displayStorage(const Storage& storage)
{
if (m_displayIntFirst)
std::cout << storage.m_nValue << ' ' << storage.m_dValue << '\n';
else // display double first
std::cout << storage.m_dValue << ' ' << storage.m_nValue << '\n';
}
void setDisplayIntFirst(bool b)
{
m_displayIntFirst = b;
}
};
int main()
{
Storage storage { 5, 6.7 };
Display display { false };
display.displayStorage(storage);
display.setDisplayIntFirst(true);
display.displayStorage(storage);
return 0;
}Because the Display class is a friend of Storage, Display members can access the private members of any Storage object they have access to.
This program produces the following result:
6.7 5 5 6.7
A few additional notes on friend classes.
First, even though Display is a friend of Storage, Display has no access to the *this pointer of Storage objects (because *this is actually a function parameter).
Second, friendship is not reciprocal. Just because Display is a friend of Storage does not mean Storage is also a friend of Display. If you want two classes to be friends of each other, both must declare the other as a friend.
Author’s note
Sorry if this one hits a little close to home!
Class friendship is also not transitive. If class A is a friend of B, and B is a friend of C, that does not mean A is a friend of C.
For advanced readers
Nor is friendship inherited. If class A makes B a friend, classes derived from B are not friends of A.
Friend member functions
Instead of making an entire class a friend, you can make a single member function a friend. This is done similarly to making a non-member function a friend, except the name of the member function is used instead.
However, in actuality, this can be a little trickier than expected. Let’s convert the previous example to make Display::displayStorage a friend member function. You might try something like this:
#include <iostream>
class Display; // forward declaration for class Display
class Storage
{
private:
int m_nValue {};
double m_dValue {};
public:
Storage(int nValue, double dValue)
: m_nValue { nValue }, m_dValue { dValue }
{
}
// Make the Display::displayStorage member function a friend of the Storage class
friend void Display::displayStorage(const Storage& storage); // error: Storage hasn't seen the full definition of class Display
};
class Display
{
private:
bool m_displayIntFirst {};
public:
Display(bool displayIntFirst)
: m_displayIntFirst { displayIntFirst }
{
}
void displayStorage(const Storage& storage)
{
if (m_displayIntFirst)
std::cout << storage.m_nValue << ' ' << storage.m_dValue << '\n';
else // display double first
std::cout << storage.m_dValue << ' ' << storage.m_nValue << '\n';
}
};
int main()
{
Storage storage { 5, 6.7 };
Display display { false };
display.displayStorage(storage);
return 0;
}However, it turns out this won’t work. In order to make a member function a friend, the compiler has to have seen the full definition for the class of the friend member function (not just a forward declaration). Since class Storage hasn’t seen the full definition for class Display yet, the compiler will error at the point where we try to make the member function a friend.
Fortunately, this is easily resolved by moving the definition of class Display before the definition of class Storage (either in the same file, or by moving the definition of Display to a header file and #including it before Storage is defined).
#include <iostream>
class Display
{
private:
bool m_displayIntFirst {};
public:
Display(bool displayIntFirst)
: m_displayIntFirst { displayIntFirst }
{
}
void displayStorage(const Storage& storage) // compile error: compiler doesn't know what a Storage is
{
if (m_displayIntFirst)
std::cout << storage.m_nValue << ' ' << storage.m_dValue << '\n';
else // display double first
std::cout << storage.m_dValue << ' ' << storage.m_nValue << '\n';
}
};
class Storage
{
private:
int m_nValue {};
double m_dValue {};
public:
Storage(int nValue, double dValue)
: m_nValue { nValue }, m_dValue { dValue }
{
}
// Make the Display::displayStorage member function a friend of the Storage class
friend void Display::displayStorage(const Storage& storage); // okay now
};
int main()
{
Storage storage { 5, 6.7 };
Display display { false };
display.displayStorage(storage);
return 0;
}However, we now have another problem. Because member function Display::displayStorage() uses Storage as a reference parameter, and we just moved the definition of Storage below the definition of Display, the compiler will complain it doesn’t know what a Storage is. We can’t fix this one by rearranging the definition order, because then we’ll undo our previous fix.
Fortunately, this is also fixable in a couple of simple steps. First, we can add class Storage as a forward declaration so the compiler will be okay with a reference to Storage before it has seen the full definition of the class.
Second, we can move the definition of Display::displayStorage() out of the class, after the full definition of Storage class.
Here’s what this looks like:
#include <iostream>
class Storage; // forward declaration for class Storage
class Display
{
private:
bool m_displayIntFirst {};
public:
Display(bool displayIntFirst)
: m_displayIntFirst { displayIntFirst }
{
}
void displayStorage(const Storage& storage); // forward declaration for Storage needed for reference here
};
class Storage // full definition of Storage class
{
private:
int m_nValue {};
double m_dValue {};
public:
Storage(int nValue, double dValue)
: m_nValue { nValue }, m_dValue { dValue }
{
}
// Make the Display::displayStorage member function a friend of the Storage class
// Requires seeing the full definition of class Display (as displayStorage is a member)
friend void Display::displayStorage(const Storage& storage);
};
// Now we can define Display::displayStorage
// Requires seeing the full definition of class Storage (as we access Storage members)
void Display::displayStorage(const Storage& storage)
{
if (m_displayIntFirst)
std::cout << storage.m_nValue << ' ' << storage.m_dValue << '\n';
else // display double first
std::cout << storage.m_dValue << ' ' << storage.m_nValue << '\n';
}
int main()
{
Storage storage { 5, 6.7 };
Display display { false };
display.displayStorage(storage);
return 0;
}Now everything will compile properly: the forward declaration of class Storage is enough to satisfy the declaration of Display::displayStorage() inside the Display class. The full definition of Display satisfies declaring Display::displayStorage() as a friend of Storage. And the full definition of class Storage is enough to satisfy the definition of member function Display::displayStorage().
If that’s a bit confusing, see the comments in the program above. The key points are that a class forward declaration satisfies references to the class. However, accessing members of a class requires that the compiler have seen the full class definition.
If this seems like a pain -- it is. Fortunately, this dance is only necessary because we’re trying to do everything in a single file. A better solution is to put each class definition in a separate header file, with the member function definitions in corresponding .cpp files. That way, all of the class definitions would be available in the .cpp files, and no rearranging of classes or functions is necessary!
Quiz time
Question #1
In geometry, a point is a position in space. We can define a point in 3d-space as the set of coordinates x, y, and z. For example, Point { 2.0, 1.0, 0.0 } would be the point at coordinate space x=2.0, y=1.0, and z=0.0.
In physics, a vector is a quantity that has a magnitude (length) and a direction (but no position). We can define a vector in 3d-space as an x, y, and z value representing the direction of the vector along the x, y, and z axis (the length can be derived from these). For example, Vector { 2.0, 0.0, 0.0 } would be a vector representing a direction along the positive x-axis (only), with length 2.0.
A Vector can be applied to a Point to move the Point to a new position. This is done by adding the vector’s direction to the point’s position to yield a new position. For example, Point { 2.0, 1.0, 0.0 } + Vector { 2.0, 0.0, 0.0 } would yield Point { 4.0, 1.0, 0.0 }.
Such points and vectors are often used in computer graphics (with points representing the vertices of a shape, and vectors represent movement of the shape).
Given the following program:
#include <iostream>
class Vector3d
{
private:
double m_x{};
double m_y{};
double m_z{};
public:
Vector3d(double x, double y, double z)
: m_x{x}, m_y{y}, m_z{z}
{
}
void print() const
{
std::cout << "Vector(" << m_x << ", " << m_y << ", " << m_z << ")\n";
}
};
class Point3d
{
private:
double m_x{};
double m_y{};
double m_z{};
public:
Point3d(double x, double y, double z)
: m_x{x}, m_y{y}, m_z{z}
{ }
void print() const
{
std::cout << "Point(" << m_x << ", " << m_y << ", " << m_z << ")\n";
}
void moveByVector(const Vector3d& v)
{
// implement this function as a friend of class Vector3d
}
};
int main()
{
Point3d p { 1.0, 2.0, 3.0 };
Vector3d v { 2.0, 2.0, -3.0 };
p.print();
p.moveByVector(v);
p.print();
return 0;
}> Step #1
Make Point3d a friend class of Vector3d, and implement function Point3d::moveByVector().
> Step #2
Instead of making class Point3d a friend of class Vector3d, make member function Point3d::moveByVector a friend of class Vector3d.
> Step #3
Reimplement the solution to the prior step using 5 separate files: Point3d.h, Point3d.cpp, Vector3d.h, Vector3d.cpp, and main.cpp.
Thanks to reader Shiva for the suggestion and solution.