Chapter 21 - Standard C++ Language Additions

2000 Deitel & Associates, Inc. All rights reserved. 2000 Deitel & Associates, Inc. All rights reserved.

Chapter 21 - Standard C++ Language Additions

Outline21.1 Introduction21.2 bool Data Type21.3 static_cast Operator21.4 const_cast Operator21.5 reinterpret_cast Operator21.6 namespaces21.7 Run-Time Type Information (RTTI)21.8 Operator Keywords21.9 explicit Constructors21.10 mutable Class Members21.11 Pointers to Class Members (.* and ->*)21.12 Multiple Inheritance and virtual Base Classes


21.1Introduction

• We shall cover standard C++ features – data type bool – cast operators – namespaces – run-time type information (RTTI) – operator keywords


21.2bool Data Type

•bool - can be false or true – preferable to 0 (false) and non-zero (true)

• outputting bool variables– numerical default (0 or 1)– stream manipulator boolalpha

• outputs string "true" or "false"

Examples:cout << boolVariable

cout << boolalpha << boolVariable


Outline

1. Initialize bool variable an int.

2. Input a variable.

3. Print the bool value of the int.

3.1 Print the value of bool variable.

1 // Fig. 21.1: fig21_01.cpp

2 // Demonstrating data type bool.

3 #include <iostream>

4

5 using std::cout;

6 using std::endl;

7 using std::cin;

8 using std::boolalpha;

9

10 int main()

11 {

12 bool boolean = false;

13 int x = 0;

14

15 cout << "boolean is " << boolean

16 << "\nEnter an integer: ";

17 cin >> x;

18

19 cout << "integer " << x << " is"

20 << ( x ? " nonzero " : " zero " )

21 << "and interpreted as ";

22

boolean is 0

Enter an integer: 22


Outline

3.2 Print the string value of the bool variable.

Program Output

23 if ( x )24 cout << "true\n";25 else26 cout << "false\n";2728 boolean = true;29 cout << "boolean is " << boolean;30 cout << "\nboolean output with boolalpha manipulator is " 31 << boolalpha << boolean << endl;3233 return 0;34 }

boolean is 0Enter an integer: 22integer 22 is nonzero and interpreted as trueboolean is 1boolean output with boolalpha manipulator is true

Notice how the output varies.

boolean is 1

boolean output with boolalpha manipulator is true

integer 22 is nonzero and interpreted as true


21.3static_cast Operator

• C++ has 4 separate, specific casts

• static_cast - conversion between types– type checking at compile time

– standard conversions: void* to char*, int to float, etc.

– base class pointers to derived class pointers

• Format:static_cast<type to convert to>(object to convert)

int z = 3;

float x = static_cast<int>(z);


Outline

1. Define class BaseClass

1.1 Define member function

1.2 Define class DerivedClass

1.3 Define member function

1.4 Global function prototype(calls function f)

2. Function calls

3. Output results

1 // Fig. 21.2: fig21_02.cpp

2 // Demonstrating the static_cast operator.


4

5 using std::cout;

6 using std::endl;

7

8 class BaseClass {

9 public:

10 void f( void ) const { cout << "BASE\n"; }

11 };

12

13 class DerivedClass : public BaseClass {

14 public:

15 void f( void ) const { cout << "DERIVED\n"; }

16 };

17

18 void test( BaseClass * );

19

20 int main()

21 {

22 // use static_cast for a conversion

23 double d = 8.22;

24 int x = static_cast< int >( d );

25

26 cout << "d is " << d << "\nx is " << x << endl;

27

28 BaseClass * basePtr = new DerivedClass;

29 test( basePtr ); // call test

30 delete basePtr;

convert double to int

d is 8.22x is 8


Outline

3.1 Function definition

Program Output

31

32 return 0;

33 }

34

35 void test( BaseClass * basePtr )

36 {

37 DerivedClass *derivedPtr;

38

39 // cast base class pointer into derived class pointer

40 derivedPtr = static_cast< DerivedClass * >( basePtr );

41 derivedPtr->f(); // invoke DerivedClass function f

42 }

d is 8.22x is 8DERIVED

converts a base pointer to a derived pointer, and calls the derived function f


Outline

1. Class definition

1.1 Initialize objects

2. Print

1 // Fig. 21.2: fig21_02.cpp

2 // Demonstrating the static_cast operator.


4

5 using std::cout;

6 using std::endl;

7

8 class BaseClass {

9 public:

10 void f( void ) const { cout << "BASE\n"; }

11 };

12

13 class DerivedClass : public BaseClass {

14 public:

15 void f( void ) const { cout << "DERIVED\n"; }

16 };

17

18 void test( BaseClass * );

19

20 int main()

21 {

22 // use static_cast for a conversion

23 double d = 8.22;

24 int x = static_cast< int >( d );

25

26 cout << "d is " << d << "\nx is " << x << endl;

27

28 BaseClass * basePtr = new DerivedClass;

29 test( basePtr ); // call test

30 delete basePtr;

In const function print the this pointer is originally const ConstCastTest *. It is cast into type ConstCastTest *, and can then be modified.


Outline

3. Function definition

Program Output

31

32 return 0;

33 }

34

35 void test( BaseClass * basePtr )

36 {

37 DerivedClass *derivedPtr;

38

39 // cast base class pointer into derived class pointer

40 derivedPtr = static_cast< DerivedClass * >( basePtr );

41 derivedPtr->f(); // invoke DerivedClass function f

42 }

d is 8.22x is 8DERIVED


21.4const_cast Operator

• const_cast - cast away const or volatile – cannot be used directly to cast away const-ness

• use pointers


Outline

1. Define class ConstCastTest

1.1 Define member functions

1 // Fig. 21.3: fig21_03.cpp

2 // Demonstrating the const_cast operator.


4

5 using std::cout;

6 using std::endl;

7

8 class ConstCastTest {

9 public:

10 void setNumber( int );

11 int getNumber() const;

12 void printNumber() const;

13 private:

14 int number;

15 };

16

17 void ConstCastTest::setNumber( int num ) { number = num; }

18

19 int ConstCastTest::getNumber() const { return number; }

20

21 void ConstCastTest::printNumber() const

22 {

23 cout << "\nNumber after modification: ";

24

25 // the expression number-- would generate compile error


Outline

2. Create and initialize object

3. Modify and print object with a const function.

Program Output

26 // undo const-ness to allow modification27 const_cast< ConstCastTest * >( this )->number--;2829 cout << number << endl;30 }3132 int main()33 {34 ConstCastTest x;35 x.setNumber( 8 ); // set private data number to 836 37 cout << "Initial value of number: " << x.getNumber();3839 x.printNumber();40 return 0;41 }

Initial value of number: 8Number after modification: 7

Casts the this pointer into type ConstCastTest *. This casts away the "const-ness" and allows number to be modified.


21.5reinterpret_cast Operator

• reinterpret_cast - for nonstandard casts – one pointer type to another pointer type, void* to int, etc.

– cannot be used for standard casts (int to double, etc.).


Outline

1. Initialize variables and pointers

2. Cast a pointer to pointer of a different type

3. Output data

Program Output

1 // Fig. 21.4: fig21_04.cpp

2 // Demonstrating the reinterpret_cast operator.


4

5 using std::cout;

6 using std::endl;

7

8 int main()

9 {

10 int x = 120, *ptr = &x;

11

12 cout << *reinterpret_cast<char *>( ptr ) << endl;

13

14 return 0;

15 }

x

ptr (type int *) cast to a pointer of type (char *).

120 is the ASCII character code for 'x'


21.6namespaces

• variables with same name and different scopes can overlap– need to distinguish them

• a namespace defines a scope for local and global identifiers.– body delimited by braces {}– use (::) to access namespace members:

namespace_name::member

– or, a using statement must occur before name is usedusing namespace namespace_name;

-members of the namespace do not need a prefix

– not guaranteed to be unique

– can be nested


21.6namespaces (II)

• Unnamed namespaces– occupy global namespace– directly accessible– do not need namespace name– global variables are in global namespace

• accessible in all scopes


Outline

1. Use std namespace

1.1 Declare global variable

1.2 Define namespace Example

1.3 Define namespace inner

1.4 Define unnamed namespace

2. Print variables

1 // Fig. 21.5: fig21_05.cpp

2 // Demonstrating namespaces.


4 using namespace std; // use std namespace

5

6 int myInt = 98; // global variable

7

8 namespace Example {

9 const double PI = 3.14159;

10 const double E = 2.71828;

11 int myInt = 8;

12 void printValues();

13

14 namespace Inner { // nested namespace

15 enum Years { FISCAL1 = 1990, FISCAL2, FISCAL3 };

16 }

17 }

18

19 namespace { // unnamed namespace

20 double d = 88.22;

21 }

22

23 int main()

24 {

25 // output value d of unnamed namespace

26 cout << "d = " << d;

27

28 // output global variable

29 cout << "\n(global) myInt = " << myInt;

Unnamed namespace members do not need qualifiers

d = 88.22

(global) myInt = 98


Outline

2. Print variables


31 // output values of Example namespace

32 cout << "\nPI = " << Example::PI << "\nE = "

33 << Example::E << "\nmyInt = "

34 << Example::myInt << "\nFISCAL3 = "

35 << Example::Inner::FISCAL3 << endl;

36

37 Example::printValues(); // invoke printValues function

38

39 return 0;

40 }

41

42 void Example::printValues()

43 {

44 cout << "\nIn printValues:\n" << "myInt = "

45 << myInt << "\nPI = " << PI << "\nE = "

46 << E << "\nd = " << d << "\n(global) myInt = "

47 << ::myInt << "\nFISCAL3 = "

48 << Inner::FISCAL3 << endl;

49 }

PI = 3.14159

E = 2.71828

myInt = 8

FISCAL3 = 1992

In printValues:

myInt = 8

PI = 3.14159

E = 2.71828

d = 88.22

(global) myInt = 98

FISCAL3 = 1992

Function printValues is a member of Example and does not need a namespace qualifier.


Outline

Program Output

d = 88.22(global) myInt = 98PI = 3.14159E = 2.71828myInt = 8FISCAL3 = 1992 In printValues:myInt = 8PI = 3.14159E = 2.71828d = 88.22(global) myInt = 98FISCAL3 = 1992


21.7Run-Time Type Information (RTTI)

• determines an object's type at run time

• typeid (in <typeinfo>)typeid(object).name()

- returns the name of the object as a C-style string

• dynamic_cast - for polymorphic programming– often used to downcast base-class pointer to derived-class pointer

– used with virtual functions


Outline

1. Define base and derived classes

1 // Fig. 21.7: fig21_07.cpp2 // Demonstrating dynamic_cast.3 #include <iostream>45 using std::cout;6 using std::endl;78 const double PI = 3.14159;910 class Shape {11 public:12 virtual double area() const { return 0.0; }13 };1415 class Circle : public Shape {16 public:17 Circle( int r = 1 ) { radius = r; }1819 virtual double area() const 20 { 21 return PI * radius * radius; 22 };23 protected:24 int radius;25 };2627 class Cylinder : public Circle {28 public:29 Cylinder( int h = 1 ) { height = h; }3031 virtual double area() const 32 { 33 return 2 * PI * radius * height +


Outline

1.1 Function prototype

1.2 Declare objects

2. Function calls

3. Define function.

34 2 * Circle::area();35 }36 private:37 int height;38 };3940 void outputShapeArea( const Shape * ); // prototype4142 int main()43 {44 Circle circle;45 Cylinder cylinder;46 Shape *ptr = 0;4748 outputShapeArea( &circle ); // output circle's area49 outputShapeArea( &cylinder ); // output cylinder's area50 outputShapeArea( ptr ); // attempt to output area51 return 0;52 }5354 void outputShapeArea( const Shape *shapePtr )55 {56 const Circle *circlePtr;57 const Cylinder *cylinderPtr;5859 // cast Shape * to a Cylinder *60 cylinderPtr = dynamic_cast< const Cylinder * >( shapePtr );6162 if ( cylinderPtr != 0 ) // if true, invoke area()63 cout << "Cylinder's area: " << shapePtr->area(); 64 else { // shapePtr does not refer to a cylinder65 66 // cast shapePtr to a Circle *

Notice how shapePtr is cast to various types. If it is not of the right type, the cast returns 0.


Outline

3. Define function.

Program Output

67 circlePtr = dynamic_cast< const Circle * >( shapePtr );

68

69 if ( circlePtr != 0 ) // if true, invoke area()

70 cout << "Circle's area: " << circlePtr->area();

71 else

72 cout << "Neither a Circle nor a Cylinder.";

73 }

74

75 cout << endl;

76 }

Circle's area: 3.14159Cylinder's area: 12.5664Neither a Circle nor a Cylinder.

Notice how shapePtr is cast to various types. If it is not of the right type, the cast returns 0.


21.8Operator Keywords

• can use keywords in place of operators (such as !, &, ^, etc.). – use header <iso646.h> (may vary with compiler)

Operator Keyword Description

Logical operator keywords

&& and logical AND

|| or logical OR

! not logical NOT

Inequality operator keyword

!= not_eq inequality

Bitwise operator keywords

& bitand bitwise AND

| bitor bitwise inclusive OR

^ xor bitwise exclusive OR

~ compl bitwise complement

Bitwise assignment operator keywords

&= and_eq bitwise AND assignment

|= or_eq bitwise inclusive OR assignment

^= xor_eq bitwise exclusive OR assignment


Outline

1. Load header

1.1 Initialize variables

2. Use operator keywords

3. Print results

1 // Fig. 21.9: fig21_09.cpp

2 // Demonstrating operator keywords.


4

5 using std::cout;

6 using std::endl;


8

9 #include <iso646.h>

10

11 int main()

12 {

13 int a = 8, b = 22;

14

15 cout << boolalpha

16 << " a and b: " << ( a and b )

17 << "\n a or b: " << ( a or b )

18 << "\n not a: " << ( not a )

19 << "\na not_eq b: " << ( a not_eq b )

20 << "\na bitand b: " << ( a bitand b )

21 << "\na bit_or b: " << ( a bitor b )

22 << "\n a xor b: " << ( a xor b )

23 << "\n compl a: " << ( compl a )

24 << "\na and_eq b: " << ( a and_eq b )

25 << "\n a or_eq b: " << ( a or_eq b )

26 << "\na xor_eq b: " << ( a xor_eq b ) << endl;

27

28 return 0;

29 }

Operator keywords can be used instead of the symbols


Outline

Program Output

a and b: true a or b: true not a: falsea not_eq b: falsea bitand b: 22a bit_or b: 22 a xor b: 0 compl a: -23a and_eq b: 22 a or_eq b: 30a xor_eq b: 30


21.9explicit Constructors

• constructors with one argument can be used for implicit conversion– type received by constructor turned into an object

– automatic conversion sometimes undesirable

• keyword explicit prevents implicit conversion– use before constructor prototype in class definition


Outline

1. Define Array class

1.1 Define constructor

1.2 Define destructor

2. Create object

2.1 Print Array object

2.2 Print int


1 // Fig. 21.9: fig21_09.cpp

2 // Demonstrating operator keywords.


4

5 using std::cout;

6 using std::endl;


8

9 #include <iso646.h>

10

11 int main()

12 {

13 int a = 8, b = 22;

14

15 cout << boolalpha

16 << " a and b: " << ( a and b )

17 << "\n a or b: " << ( a or b )

18 << "\n not a: " << ( not a )

19 << "\na not_eq b: " << ( a not_eq b )

20 << "\na bitand b: " << ( a bitand b )

21 << "\na bit_or b: " << ( a bitor b )

22 << "\n a xor b: " << ( a xor b )

23 << "\n compl a: " << ( compl a )

24 << "\na and_eq b: " << ( a and_eq b )

25 << "\n a or_eq b: " << ( a or_eq b )

26 << "\na xor_eq b: " << ( a xor_eq b ) << endl;

27

28 return 0;

29 }

15 is not an Array object, but it is implicitly converted to an Array object using the conversion constructor. This new object is printed using outputArray.

If the keyword explicit comes before the constructor, this program will issue a compiler error - outputArray cannot take an int.


Outline

Program Output

a and b: true a or b: true not a: falsea not_eq b: falsea bitand b: 22a bit_or b: 22 a xor b: 0 compl a: -23a and_eq b: 22 a or_eq b: 30a xor_eq b: 30


Outline

1. Class definition

1.1 Function prototype

1.2 Member variables

----------------------

1. Load header file

1 // Fig 21.10: array2.h

2 // Simple class Array (for integers)

3 #ifndef ARRAY2_H

4 #define ARRAY2_H

5


7

8 using std::ostream;

9

10 class Array {

11 friend ostream &operator<<( ostream &, const Array & );

12 public:

13 Array( int = 10 ); // default/conversion constructor

14 ~Array(); // destructor

15 private:

16 int size; // size of the array

17 int *ptr; // pointer to first element of array

18 };

19

20 #endif

21 // Fig 21.10: array2.cpp22 // Member function definitions for class Array23 #include <iostream>2425 using std::cout;26 using std::ostream;2728 #include <cassert>29 #include "array2.h"30


Outline

1.1 Function definitions

31 // Default constructor for class Array (default size 10)

32 Array::Array( int arraySize )

33 {

34 size = ( arraySize > 0 ? arraySize : 10 );

35 cout << "Array constructor called for "

36 << size << " elements\n";

37

38 ptr = new int[ size ]; // create space for array

39 assert( ptr != 0 ); // terminate if memory not allocated

40

41 for ( int i = 0; i < size; i++ )

42 ptr[ i ] = 0; // initialize array

43 }

44

45 // Destructor for class Array

46 Array::~Array() { delete [] ptr; }

47

48 // Overloaded output operator for class Array

49 ostream &operator<<( ostream &output, const Array &a )

50 {

51 int i;

52

53 for ( i = 0; i < a.size; i++ )

54 output << a.ptr[ i ] << ' ' ;

55

56 return output; // enables cout << x << y;

57 }


Outline

1. Load header

1.1 Initialize object

2. Print object


58 // Fig 21.10: fig21_10.cpp

59 // Driver for simple class Array


61

62 using std::cout;

63

64 #include "array2.h"

65

66 void outputArray( const Array & );

67

68 int main()

69 {

70 Array integers1( 7 );

71

72 outputArray( integers1 ); // output Array integers1

73

74 outputArray( 15 ); // convert 15 to an Array and output

75

76 return 0;

77 }

78

79 void outputArray( const Array &arrayToOutput )

80 {

81 cout << "The array received contains:\n"

82 << arrayToOutput << "\n\n";

83 }

outputArray needs a parameter of type const Array &, so 15 is converted into an Array by the conversion constructor.


Outline

Program Output

Array constructor called for 7 elementsThe array received contains:0 0 0 0 0 0 0 Array constructor called for 15 elementsThe array received contains:0 0 0 0 0 0 0 0 0 0 0 0 0 0 0


21.10 mutable Class Members

• mutable data member– always modifiable, even in a const function or object

– permanently allows a const data member to be modified

• const_cast: – used every time a const data member must be

modified

– reduces risk of accidentally modifying a const variable


21.11 Pointers to Class Members (.* and ->*)

• pointers to class members are different from normal pointers

• use .* and ->* instead of . and -> when accessing class members (functions and data)


Outline

1. Class definition

1.1 Function prototypes

1.2 Initialize object

2. Function calls

1 // Fig. 21.13 fig21_13.cpp

2 // Demonstrating operators .* and ->*


4

5 using std::cout;

6 using std::endl;

7

8 class Test {

9 public:

10 void function() { cout << "function\n"; }

11 int value;

12 };

13

14 void arrowStar( Test * );

15 void dotStar( Test * );

16

17 int main()

18 {

19 Test t;

20

21 t.value = 8;

22 arrowStar( &t );

23 dotStar( &t );

24 return 0;

25 }


Outline

3. Function definitions

Program Output

26

27 void arrowStar( Test *tPtr )

28 {

29 void ( Test::*memPtr )() = &Test::function;

30 ( tPtr->*memPtr )(); // invoke function indirectly

31 }

32

33 void dotStar( Test *tPtr )

34 {

35 int Test::*vPtr = &Test::value;

36 cout << ( *tPtr ).*vPtr << endl; // access value

37 }

function8

arrowStar declares and initializes memPtr to point to a function in Test that takes no parameters and returns no value.

&Test::function to get the offset into the class for member function function.

Without Test::, memPtr is a standard pointer.

dotStar declares and initializes vPtr to point to value.

The .* operator is then used to access the member to which vPtr points.


21.12 Multiple Inheritance and virtual Base Classes

• Ambiguities can result with multiple inheritance

• iostream could have duplicate subobjects (data from ios inherited into ostream and istream).– Upcasting an iostream pointer to an ios object creates a problem.

Two ios subobjects could exist: which one is used?

– Ambiguous, results in syntax error (of course, iostream does not actually have this problem)

ios

ostreamistream

iostream


21.12 Multiple Inheritance and virtual Base Classes (II)

• Use virtual base class inheritance– only one subobject of the base is inherited into the multiply derived class.

Second Derived Class

Base Class

First Derived Class

Multiply-Derived Class

virtual inheritance

virtual inheritance


Outline1 // Fig. 21.15: fig21_15.cpp

2 // Attempting to polymorphically call a function

3 // multiply inherited from two base classes.


5

6 using std::cout;

7 using std::endl;

8

9 class Base {

10 public:

11 virtual void print() const = 0; // pure virtual

12 };

13

14 class DerivedOne : public Base {

15 public:

16 // override print function

17 void print() const { cout << "DerivedOne\n"; }

18 };

19

20 class DerivedTwo : public Base {

21 public:

22 // override print function

23 void print() const { cout << "DerivedTwo\n"; }

24 };

25

26 class Multiple : public DerivedOne, public DerivedTwo {

27 public:

28 // qualify which version of function print

29 void print() const { DerivedTwo::print(); }

30 };

31

1. Define base class

1.1 Define non-virtual derived classes

1.2 Define multiply derived class


Outline

1.3 Create class objects

1.4 Create an array of base class pointers.

2. Initialize array elements

Program Output

32 int main()

33 {

34 Multiple both; // instantiate Multiple object

35 DerivedOne one; // instantiate DerivedOne object

36 DerivedTwo two; // instantiate DerivedTwo object

37

38 Base *array[ 3 ];

39 array[ 0 ] = &both; // ERROR--ambiguous

40 array[ 1 ] = &one;

41 array[ 2 ] = &two;

42

43 // polymorphically invoke print

44 for ( int k = 0; k < 3; k++ )

45 array[ k ] -> print();

46

47 return 0;

48 }

Compiling...Fig21_15.cppfig21_15.cpp(39) : error C2594: '=' : ambiguous conversions from 'class Multiple *' to 'class Base *'

The address of both is implicitly converted to a base class pointer. This is ambiguous because class Multiple has duplicate subobjects inherited from Base.

If DerivedOne and DerivedTwo had used virtual inheritance, no errors would result and the objects will be printed.

Chapter 21 - Standard C++ Language Additions

Documents

Transcript of Chapter 21 - Standard C++ Language Additions