Object Oriented Programming

[CSE202]

Classes in C++ - Part 2

(Static Members, Copy Constructors, Single Inheritance and Friend Functions)
SAURABH SRIVASTAVA
ASSISTANT PROFESSOR
DEPARTMENT OF COMPUTER SCIENCE & ENGINEERING
IIT (ISM) DHANBAD
What we know so far …
Namespaces are collections of declarations and definitions
◦ As long as things are unique within a namespace, it is fine !!

Classes can have fields and methods which could be visible to or hidden from the outside world
◦ The former are called public members and the latter are called private members
◦ The default behaviour is that members are private, unless defined in the public section
◦ We did not discuss the static members of the class last time, so we will do it now

Constructors are called whenever a new object for a class is created

◦ A class can have more than one constructor – the job of all of them is to initialise the object’s fields
◦ You can provide initial values for some or all of the fields of the class
◦ We will cover one more type of constructor in this part of the lecture

SAURABH SRIVASTAVA | DEPARTMENT OF COMPUTER SCIENCE & ENGINEERING | IIT (ISM) DHANBAD
Static Members of a class
Static members of a class – fields or methods – are essentially associated with the class …
◦ … as against the non-static members, which are rather associated with an object

Only one copy of a static field is maintained for the whole class
◦ Static variables are stored in a separate memory section and initialised before any object is created

Static methods can only work with static fields of a class

◦ They can however, access the non-static members (including private members) through objects

Static members of a class can be accessed using the class’s name with scope resolution operator
◦ For instance, <class name>::<static field> or <class name>::<static method>
◦ By the way, you can access them using an object as well, but logically, it doesn’t make much sense !!
◦ The access rules (public or private) apply to static members in a similar fashion

SAURABH SRIVASTAVA | DEPARTMENT OF COMPUTER SCIENCE & ENGINEERING | IIT (ISM) DHANBAD
Extended Toy Example – static members

This is an extended version of the Toy class we saw last time

SAURABH SRIVASTAVA | DEPARTMENT OF COMPUTER SCIENCE & ENGINEERING | IIT (ISM) DHANBAD
Extended Toy Example – static members

This is an extended version of the Toy class we saw last time

These are the static members of the class

A static field of a class is initialized by default with 0 (or an

equivalent value)

SAURABH SRIVASTAVA | DEPARTMENT OF COMPUTER SCIENCE & ENGINEERING | IIT (ISM) DHANBAD
Extended Toy Example – static members

At the time of definition, the static methods are defined

exactly like a non-static method

It is just that it can only use the static fields of the class
(because static fields too, are the property of the class and
not associated with a particular object)

SAURABH SRIVASTAVA | DEPARTMENT OF COMPUTER SCIENCE & ENGINEERING | IIT (ISM) DHANBAD
Extended Toy Example – static members

Here, we are using the static field id as a way to provide a

new identification value to each object of the class, as and
when they are created

This is a fairly common use for static fields – keeping a

count of the number of objects created

SAURABH SRIVASTAVA | DEPARTMENT OF COMPUTER SCIENCE & ENGINEERING | IIT (ISM) DHANBAD
Copy Constructors
There is a special type of constructor, that is also added by added by the compiler implicitly
◦ However, a Copy Constructor can never be used for the first object of a class …
◦ … or for any subsequent objects, unless the required action is – creating a copy of an existing object

The constructor, called a Copy Constructor can also be defined explicitly

In most cases, it is not required to provide a user-defined copy constructor
◦ There are some peculiar cases, where you may need to define them (check the homework)

A copy constructor is called implicitly whenever the compiler needs to create a copy of an object
◦ For example, when you pass an object to a function by value

You can also use it to create a copy of an object explicitly in your code

SAURABH SRIVASTAVA | DEPARTMENT OF COMPUTER SCIENCE & ENGINEERING | IIT (ISM) DHANBAD
Extended Toy Example – copy constructer

This is what a constructor looks like

C++ has another way to pass parameters as

compared to C – via a reference (check homework)

Basically, we pass an object of the same type to the

constructor by reference

We can then copy the fields of the passed object to

the newly created object

SAURABH SRIVASTAVA | DEPARTMENT OF COMPUTER SCIENCE & ENGINEERING | IIT (ISM) DHANBAD
(Single) Inheritance in C++
Assume that you now want to make the class Toy a little more “specific”
◦ You also wish to store information about the type and number of batteries required to run the toy
◦ But, not all toys require batteries – some can be played with, without any battery as well

One option is to add some fields and methods in the Toy class to cover this aspect
◦ But then, we will need to adopt a convention like “Battery not required” for some Toys

Another option is to extend the Toy class and create a new class for “battery operated toys”
◦ Basically, a battery operated toy has all the properties and behaviour of a toy …
◦ … plus, some more properties and behaviour
◦ There is no point creating a copy of all fields and methods of the Toy class …
◦ … instead, through Inheritance we “import” this behaviour, and add more to the new class

SAURABH SRIVASTAVA | DEPARTMENT OF COMPUTER SCIENCE & ENGINEERING | IIT (ISM) DHANBAD
The Battery Operated Toy Example
Here, we created another class to as a specific
type of toys – those that require batteries

SAURABH SRIVASTAVA | DEPARTMENT OF COMPUTER SCIENCE & ENGINEERING | IIT (ISM) DHANBAD
The Battery Operated Toy Example
This is how we inherit the class Toy2 in
BatteryOperatedToy

The syntax is
class <D> : <a> <B>
where D is the derived class, B is the base class
and a could be public, private or
protected

We will discuss the protected access

specifier shortly …

The access specifier here decides the access

rights of the “inherited members”

SAURABH SRIVASTAVA | DEPARTMENT OF COMPUTER SCIENCE & ENGINEERING | IIT (ISM) DHANBAD
The protected access specifier
Other than public and private, there is one more access specifier in C++ for class members
The specifier, protected, is similar to the specifier private, if there is no inheritance
The private members of a class cannot be inherited by any derived class
◦ One way to let a member be inherited is by making it public
◦ But, it also provides access to the member to the outside world
◦ A mid path is to make these members protected

A protected member behaves the same way as a private member, with just one exception
◦ The protected members can be inherited by a derived class
◦ To reiterate, for every other case, protected and private mean exactly the same

SAURABH SRIVASTAVA | DEPARTMENT OF COMPUTER SCIENCE & ENGINEERING | IIT (ISM) DHANBAD
Access specifiers in inheritance
Scenario:
class <D> : public <B>

When the inheritance access specifier is public,

the public members join the public section of
the derived class, and the protected members
join the protected section of the derived class

Remember, private members are not inherited

SAURABH SRIVASTAVA | DEPARTMENT OF COMPUTER SCIENCE & ENGINEERING | IIT (ISM) DHANBAD
Access specifiers in inheritance
Scenario:
class <D> : protected <B>

When the inheritance access specifier is protected,

both public as well as protected members join
the protected section of the derived class

Remember, private members are not inherited

SAURABH SRIVASTAVA | DEPARTMENT OF COMPUTER SCIENCE & ENGINEERING | IIT (ISM) DHANBAD
Access specifiers in inheritance
Scenario:
class <D> : private <B>

When the inheritance access specifier is private,

both public as well as protected members join
the private section of the derived class

Remember, private members are not inherited

SAURABH SRIVASTAVA | DEPARTMENT OF COMPUTER SCIENCE & ENGINEERING | IIT (ISM) DHANBAD
Constructors and Inheritance
When a class inherits another class, there is a hierarchical creation of the objects
This involves, creation of an object of the base class first, and then adding more fields to it …
◦ … thus making it an object of the derived type
◦ Keep in mind that space is allocated in the memory for even the private fields of the base class …
◦ … it is just that they are not accessible directly to the derived class

Thus, when a constructor of the derived class is called, “some” base class constructor is also invoked
◦ The base class constructor must be invoked before the derived class constructor is executed

If there is a default constructor available in the base class, no additional measures are required …
◦ … the default constructor is automatically invoked, when any derived class constructor is called

However, if a default constructor is not present in the base class, some additional work is required
◦ Basically, all derived class constructors must make a call to some base class constructor

SAURABH SRIVASTAVA | DEPARTMENT OF COMPUTER SCIENCE & ENGINEERING | IIT (ISM) DHANBAD
Calling base class constructors
This is how an explicit call to a base class
constructor is achieved in the derived class

SAURABH SRIVASTAVA | DEPARTMENT OF COMPUTER SCIENCE & ENGINEERING | IIT (ISM) DHANBAD
Calling base class constructors
Syntax:
derived constructor header :
base constructor call

Notice the difference – header vs call

We are making a “call” to a base
constructor from the ”header” of the
derived constructor

The only parameters that can be passed in

the base constructor call are either the
formal parameters of the derived
constructor, or constants

SAURABH SRIVASTAVA | DEPARTMENT OF COMPUTER SCIENCE & ENGINEERING | IIT (ISM) DHANBAD
Accessing private members externally
The general rule of thumb is that private members of a class are inaccessible externally
However, in some peculiar cases, you may wish to make an exception …
◦ … i.e. allow external access to the private members of the class

To do so, you have the option of creating friendships in you code !!

In C++, a non-member function can be declared as a friend of a class
A friend function has access to all the members of the class, even the private ones
◦ It cannot, however, access the members like a member function – an object or class name is still required

A member function of one class, can be a friend of another class

If you wish to make all the member functions of a class as friends of a class …
◦ … you can also make this entire class as a friend for the said class (called a friend class)

SAURABH SRIVASTAVA | DEPARTMENT OF COMPUTER SCIENCE & ENGINEERING | IIT (ISM) DHANBAD
Friend Functions
This is how you specify a function as a friend of
a class – it looks similar to a function
declaration, preceded by the keyword friend

SAURABH SRIVASTAVA | DEPARTMENT OF COMPUTER SCIENCE & ENGINEERING | IIT (ISM) DHANBAD
Friend Functions
While declaring the function, it is a good practice to
provide a declaration of the class for which this
function will be a friend (called a forward
declaration)

SAURABH SRIVASTAVA | DEPARTMENT OF COMPUTER SCIENCE & ENGINEERING | IIT (ISM) DHANBAD
Friend Functions

Friend functions can access the private members of a class using the <object>.<member> notation

SAURABH SRIVASTAVA | DEPARTMENT OF COMPUTER SCIENCE & ENGINEERING | IIT (ISM) DHANBAD
Behaviour Extension Points – Virtual Methods
Sometimes, a method may be defined for a base class …
◦ … but it’s definition may be “more precise” (or “different”) with respect to a derived class

In such cases, what we may wish is …

◦ … if we have a base class object, the expected behaviour should be the one defined in the base class …
◦ … but we have a derived class object, the behaviour expected should be the one defined in the derived class

To achieve this in C++, the method can be defined as virtual in the base class
A virtual method indicates that multiple definitions of this method “may be” found at the runtime
The definition that must be invoked is decided by the type of the object
◦ If the object is of base class type, the definition from the base class is executed, and …
◦ … if the object is of derived class type, the definition executed is the one provided in the derived class

Keep in mind that only non-static methods can be declared as virtual

SAURABH SRIVASTAVA | DEPARTMENT OF COMPUTER SCIENCE & ENGINEERING | IIT (ISM) DHANBAD
Virtual Functions

This is how a virtual function can be declared – by prepending

the keyword virtual

SAURABH SRIVASTAVA | DEPARTMENT OF COMPUTER SCIENCE & ENGINEERING | IIT (ISM) DHANBAD
Virtual Functions

These are the base class definitions for these methods …

SAURABH SRIVASTAVA | DEPARTMENT OF COMPUTER SCIENCE & ENGINEERING | IIT (ISM) DHANBAD
Virtual Functions

… and these are the definitions for the methods in the derived class

SAURABH SRIVASTAVA | DEPARTMENT OF COMPUTER SCIENCE & ENGINEERING | IIT (ISM) DHANBAD
Virtual Functions

The base class versions can also be accessed, using the scope resolution operator

SAURABH SRIVASTAVA | DEPARTMENT OF COMPUTER SCIENCE & ENGINEERING | IIT (ISM) DHANBAD
Virtual Functions

The methods can be invoked either with the base class object or derived class object

SAURABH SRIVASTAVA | DEPARTMENT OF COMPUTER SCIENCE & ENGINEERING | IIT (ISM) DHANBAD
Virtual Functions
You can observe how the output of
the code changes when the same
method is invoked with different
types of objects

This is also an example of overriding

– something we discussed before

We will revisit virtual functions after

we have covered Polymorphism

Note: To build these executables, run the command make clean all in the example directory

SAURABH SRIVASTAVA | DEPARTMENT OF COMPUTER SCIENCE & ENGINEERING | IIT (ISM) DHANBAD
Homework !!
Copy Constructors are “usually” not required to be defined explicitly
◦ It was certainly not required in our example, but in some cases it may be necessary
◦ Try to find out about these cases; may be give this article a read:
https://www.geeksforgeeks.org/copy-constructor-in-cpp/

If you remember our discussion about “passing values by reference” in the previous semester …
◦ … we briefly discussed how the same term means slightly different things in C and C++
◦ Basically, C++ has an explicit way to create references – without the use of pointers
◦ You can, of course, also use the C way to pass parameters by reference, i.e. via pointers
◦ But using references are more popular (and usually convenient)
◦ Read more about these; you may start with these links:
https://www.geeksforgeeks.org/references-in-c/
https://stackoverflow.com/questions/8627956/ways-of-passing-arguments-value-vs-reference-vs-pointer

On the last slide, try to find out why the destructor is being called twice?

SAURABH SRIVASTAVA | DEPARTMENT OF COMPUTER SCIENCE & ENGINEERING | IIT (ISM) DHANBAD