[C++] Class (part 1)

The fundamental ideas behind classes are data abstraction and encapsulation.

Data abstraction is a programming technical that relies on the separation of interface(declaration) and implementation(definition)

A class designer desings and implements a class for users of that class.

In this class, the user is a progrmmer, not the utilmate user of the application.

Every member must be declared inside its class, and we can difine a member's body either inside or outside of the class body.

Objects that are const, and references or pointers to const objects may call only const member functions.

The compiler process class in two steps

• the member declarations are compiled first
• the member function bodies are processed.

The name of a member defined outside the class must inlucde the class name.

class Sales_data::arg_print() const{
    if (units_sold){
        return revenue / units_sold;
    } else {
        return ;
    }
}

Define a funcion to return "this" object

Sales_data & Sales_data::combine(const Sales_data & rhs){
    units_sold += rhs.units_sold;
    return *this;    // return current object
}

To return an lvalue, our combine function must return a reference.

Member functions that are part of the interface of a class should be declared in the same header as the class itself.

istream & read(istream & is, Sales_data & item){
    is >> itme.bookNo >> item.units_sold;
    return is;
}
ostream & print(ostream & os, const Sales_data & item){
    os << item.isBN();
    return os;
}

Both read an writer take a reference to thier respective IO class types. The IO class cannot be copied, so only passed by reference.

Reading or Writing to a stream changes that stream, so both functions take ordinary reference, not references to const.

Sales_data add(const Sales_data & lhs, const Sales_data & rhs){
    Sales_data sum = lhs;    // copy data members from lhs into sum. statement[1]
    sum.combine(rhs);
    return sum;
} 

We initialize sum as a copy of lhs. By default, copying a class object copies that object's members.

In Java, statement[1] only assign the reference of lhs to sum. Both sum and lhs points to the same object in heap storage.

In C++, statement[1] copy the all member from lhs to sum. lhs and sum points to different object, and are independent.

The compiler generate a default constructor automatically only if a class declare no constructors.

Many class that need dynamic memory can (and generally should) use a vector or a string to manage the necessary storage. Classes that use vectors and string avoid the complexites of allocating and deallocating memory.

A class may contain zero or more access specifier(e.g. private, public). Each access specifier specify the access level of the succeeding members. The specified access level remain in effect until the next access specifier or the end of the class body.

The only difference between struct and class is the default access level.

• In struct, the members defined before the first accesss specifier are public.
• In class, these members are private.

Friend

A class can allow class or function to access its non-public members by marking that as a frined.

class Sales_data{
    friend Sales_data add(const sales_data&, const Sales_data&);
public:
    Sales_data() default;
private:
    string bookNo;
};

// declarations for members which are part of the Sales_data interface
Sales_data add(const Sales_data&, const Sales.data &);

It is good idea to group friend declaration together at the beginning or end of the class. To make a friend visible to the users of the class, we usually declare each friend(outside the class) in the same header as the class.

Encapsulation provide two important advantage:

• User code cannot corrupt the state of an encapsulated object
• The implementation of an encapsulated class can change over time without requiring changes in user-level code.

Although user code need not change when a class definition changes, the source files that use a class must be re-compiled if the class changes.

Class often have small functions that can benefit from being inline.

Mutable

A mutable data member are never const, even when it is a member of a const class

class Screen{
public:
    void some_member() const;
private:
    mutable size_t access_ctr;    // may change even in a const object
};

void Screen::some_member() const{
    access_ctr++;
};

In-class initializer

When we provide an in-class initializer, we must do so following on = asign or inside braces.

class windows_mgr{
private:
    // By default, a windows_mgr has one blank screen
    std::vector<Screen> screens{screen(, , ' ')};
}

Return *this

class Screen{
public:
    Screen& set(char);
    Screen& move(int i, int j);
    // other members. e.g. cursor
};

inline Screen& Screen::move(int i, int j){
    cursor = i * width + j;
    return *this;    // return this object as lvalue
}
inline Screen& Screen::set(char c){
    contents[cursor] = c;
    return *this;
}

If we concatenate a sequence of these actions into a single expression

    myScreen.move(, ).set('#');

Above single expression is equal to

    myScreen.move(, );
    myScreen.set('#');

Had we defined move and set to return Screen, rather thane Screen &. This statement will be vary different.

    Screen tmp = myScreen.move(, );    // the return value would be copied
    tmp.set('#');    // the content inside myScreen will be unchanged.

A class is not defined until its class body is completed. A class cannot have data members of its own type.

However, a class is considered declared(not yet defined) as soon as its calss name has been seen. Therefore, a class can have data members that are pointers or reference to its own type.

class Link_Sales{
    Screen window;
    Link_Sales *next;
    Link_Sales *prev;
};

class Screen{
    // Windows_mgr can access the non-public part of class Screen.
    friend class Windows_mgr;
}

void Windows_mgr::clear(int i){
    Screen & s = screens[i];
    s.content = '';
}

It is important to understand that friendship is not transitive. If class Windows_mgr has its own friends, that friends have no special access to Screen.

Each class controls which class or function are its friends.

Although overload functions share a common name, they are still different functions. Therefore a class must define as a friend each function in a set of overload functions that it wish to mark as friend.

The lookup rules for friends

Struct X{
    friend void f();
    X() { f(); }    // error: no declaration for f()
    void g();
    void h();
}

void X::g() { return f(); } // error: f has not been declared
void f();    //    declares the function defined in X()
void X:h() { return f(); }    // ok

Remember, some compiler do not enforce the lookup rules for friends.

class Window_mgr{
public:
    screenIndex addScreen( const Screen&);
    // other members
};

// return type is seen before the scope of Window_mgr
Window_mgr::ScreenIndex Window_mgr::addScreen(const Screen & s){
    screens.push_back(s);
    return screens.size() - ;
} 

To use screenIndex for the return type, we must specified.

Member function definitions are processed after all of the declaration in class are processed by compiler. So member functions can use any member declared inside the calss

Even though the class member(height) is hidden, it is accessable by qualified the member's name with the class's name or with this pointer expectly

void Screen::dummy_fcn(pos height){
    cursor = width * this->height;
    // alternative way
    cursor = width * Screen::height;
}

The global object is accessable by the scope opertion.

void Screen::dummy_fcn(pos height){
    cursor = width * ::height;    // the global height
}

Reference:

C++ Primer, Fifth Edition, chapter 7 Classes