Chapter 2

Procedural
Programming

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Preview

 Basic concepts that are similar in both Java and C++, including:
• standard data types
• control structures
• I/O
• functions

 Dynamic memory management, and some basic pointer

operations.

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Main function

Java C++ Comments about

C++

class C { int main() { The main function

public static void ... is not included in
main(String args[])
{
return 0; any class.
... }
}

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Comments

Comments about C++

/*
... C++ does not provide the Java comment /** ... */,
*/ used to create documentation
//

Remember about programming guidelines!

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Primitive Data Types

Comments about C++

char char is one byte long and does not use Unicode

short The language does not define the size and range of numeric
int primitive data types, which are implementation dependent.
long
float There are no predefined wrapper classes for primitive data
double types.

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: More Primitive Data Types
Java C++ Comments about C++

byte -- similar to unsigned char

boolean bool Any non-zero numeric value can be treated as true,

a zero value can be treated as false

String string
To use strings you must include the <string> library

string language = "C" + "++";

bool mark = language < "Java";
char first = language[0];
C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: typedef and sizeof

A synonym for an existing type:

typedef bool Boolean;

Size of data type (in bytes):

int k = 5;
... sizeof(int)... sizeof k …

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Variables, Constants and
Expressions
Java C++
final const --

const int LINES = 25;

Comma operator, e.g. e1, e2

for(i = 1, j = 2; i < N && j < M; ++i, ++j)

...

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Type Conversions

• between bool and int

• Implicit type conversions or casts are performed when:

 the value is promoted to a more precise type:

double d = 3; // 3 is promoted to 3.0

 the value is demoted, resulting in a loss of precision:

double d = 3.14159;
int i = d; // 3.14159 is demoted

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Type Conversions (cont.)

Better to use explicit cast:

i = static_cast<int>(d);

Example

int i, j;
... // initialization of i and j

double d = static_cast<double>(i) / j;
// avoids int division

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Type Conversions (cont.)

• Conversion to remove constness, which allows a const

identifier to be modified:

const int i = 3;
const_cast<int>(i) = 1; // not constant

• reinterpret_cast used to perform a low-level

reinterpretation of the bit pattern to force the type conversion

• dynamic_cast used for derived classes

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Control Structures

Comments about C++

Conditional and switch same

statements

Loops break or continue cannot be

labeled. A goto statement is
available (if and when really
needed).

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Basic I/O

Three predefined streams:

• cout - the standard output stream

<< (put-to) binary operator is used for output

• cin - the standard input stream

>> (get from) operator is used for input

• cerr - the standard error stream.

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Basic I/O (cont.)

cout << "Enter value "; // prompt

int i;
cin >> i; // input integer

cout << "The value is:";

cout << i; // output integer

cout << "The value is:" << i << endl; // chain

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
// File ex2.1.cpp
// Read integer values until a positive value is entered
// If a non-integerExample of I/O abort the program
value is entered,
// Otherwise, output the integer value
#include <iostream>
int main() {
int i = 0;
bool correct = false;
cout << "Enter a positive integer value" << endl;
while(cin >> i) {
if(i > 0) { test for correct input
correct = true;
break;
}
cout << "non-positive value; re-enter" << endl;
}
if(!correct)
cout << "incorrect integer value" << endl;
else cout << "You entered " << i << endl;
}
C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Arrays

C++ arrays are borrowed from C:

Java C++ Comments about C++

int [] marks = int - lower bound equal to zero

new int[10] marks[10] - size must be known at
compile time

The C++ standard library also provides vectors.

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Reference Data Types

A reference data type (which differs from Java references) is of the

form
type&
It is used for:
• passing parameters and returning values
• providing an alias in a variable definition (must be initialized in
the definition)
int i = 1;
int& pi = i; // pi is an alias for i
pi++; // now i is 2, pi not changed

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Reference Data Types
(cont.)

References must be initialized with expressions that are lvalues.

You can declare the reference as constant; for example:

const int& cpi = 8;

(may be initialized with expressions that are not lvalues).

Mostly used with parameters.

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Functions

Terminology:
an expression that appears in a function call, such as x in f(x), is
called an actual parameter, as opposed to the formal parameter
that appears in the function definition.

Construct Comments about C++

int C++ allows global functions, that is functions that are defined
foo(pars) { outside of any class (also called stand-alone functions)
body
}

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Functions (cont.)

A function may be defined as inline:

inline int sum(int i, int j) {

return i + j;
}

which is a request to the compiler to generate code inline.

Each call to the inlined function is replaced at compile time by its body:

int i = sum(1, k);

// replaced by: int i = 1 + k

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Declarations and
Definitions

A function declaration merely provides a function signature and the

return type:

bool prime(int); // note the semicolon

A function definition includes the declaration and the implementation of

the function. A function can be called only if it has been declared or
defined.

Each function declaration must come with the documentation to be used

by the client. A different kind of information is needed by the
implementor of the function, and so the function definition comes
with its own documentation.
C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Pass by Value, Pass by
Reference, and Constant Pass
by Reference

Passing parameters by value means that the formal parameter is assigned

the value of the actual parameter, and the actual parameter is not
modified:
void swap(int x, int y) {
int temp = x;
x = y;
y = temp;
}
int i = 1;
int j = 2;
swap(i, j);
C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Pass by Value, Pass by
Reference, and Constant Pass
by Reference
Passing by reference allows the function to modify the actual
parameter. The type of the parameter passed by reference is a
reference type, with an & following the parameter type:
void swap(int& x, int& y) {
int temp = x;
x = y;
y = temp;
}

int i = 3;
int j = 4;
swap(i, j); // i and j passed by reference
C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Pass by Value, Pass by
Reference, and Constant Pass
by Reference
Constant pass by reference does not allow the client to inadvertently
modify the actual parameter:

double product(const double& block, int size);

If a function has a non-constant parameter, then it may not be called with

a constant actual parameter.

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 Pass by Value and by
Reference

If you do not want to modify the value of the actual parameter,

then use
- pass by value for parameters of primitive data types,
- constant pass by reference for parameters of structured data
types, such as classes.

If you do want to modify the value of the actual parameter, then use
pass by reference.

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Default Values of
Parameters

A default value of a formal function parameter is useful if this function

is frequently called with the same actual parameter.
The client can call this function either without the actual parameter for
this formal parameter (and so using the default value), or with an
actual parameter to override the default value.

A function move() that typically moves to position 0 by a step of 1

void move(int from, int to = 0, int by = 1);

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Default Values of
Parameters (cont.)

Only trailing parameters may be assigned default values:

void move(int from, int to = 0, int by);

Overriding values of the actual parameters are assigned from left to right:
void move(int from, int to = 0, int by = 1);

move(2); // move(2, 0, 1)
move(2, 3); // move(2, 3, 1)
move(2, 3, 4); // move(2, 3, 4)

It is not possible to override the second most trailing default parameter,

without overriding the first default parameter.
C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Overloading Functions

The name of a function should identify the action performed by this

function.
Overloading a function means that you can have several functions with
the same name in a single scope, provided that these functions have
different signatures:

double max(double, double);

int max(int, int); // overloaded

bool max(int, int); // identical signatures

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Overloading Functions
(cont.)

Sometimes, overloading can be replaced by using default parameter

values:

void move(int from, int to);

void move(int from, int to, int by);

can be replaced by a single function:

void move(int from, int to, int by = 0);

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Functions and Arrays
When arrays are used as function parameters, their size has to be passed
as another parameter:

void maxMin(const double arr[], int size,

double& max, double& min) {
int i;
for(max = min = arr[0], i = 1; i < size; ++i) {
if(max < arr[i])
max = arr[i];
if(min > arr[i])
min = arr[i];
}
}
C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Functions and Arrays
(cont.)

double arr[] = {1.3, 1.2, 1.1};

double maxi, mini;
maxMin(arr, 3, maxi, mini);

Consider a different maxMin1():

const double x[] = {3, 1, 7.8};

void maxMin1(double arr[],int size,
double& max, double& min);

maxMin1(x, 3, max, min); // error, x is const

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Functions Returning
References

A function may return a reference type, and if so, a function call is an

lvalue:

int a[5];
int& index(int x[], int i) {
// gives read/write access
// to the i-th element of x
return x[i];
}

cin >> index(a, 1);

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Functions Returning
Constant References

Functions may also be designed to provide read-only access, by

returning a constant reference:

const int& get(const int x[], int i) {

return x[i];
}

cout << get(a,2);

cin >> get(a,1); // can't change the value

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Pointers and Dynamic
Memory Management

• Static memory is used to store the values of global variables, as well

as function and class static variables.

• Stack-based memory is implicitly managed by function calls and

function returns.

• Heap-based memory is used to dynamically manage memory as a

result of a program’s request, such as Java's new call.

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Pointers and Dynamic
Memory Management

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Pointers and Dynamic
Memory Management (cont.)

• Dynamic memory allocation means that a block of memory

requested by a program is removed from the heap, and can be used by
this program.

• Dynamic memory deallocation means that a program returns a block

of memory to the heap - the de-allocated memory can be used the next
time a memory request occurs.

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Pointers and Dynamic
Memory Management (cont.)

Garbage collector decides when inaccessible objects are deallocated.

Java objects are always allocated on the heap and never on the run-time
stack, and all memory deallocation is done automatically by the
garbage collector.

Memory leakage occurs when memory that is no longer needed by the

program has not been freed and made available for reuse.

What's a reference to an object in Java? It is a memory address;

specifically, the address of a block of memory allocated for this object
on the heap.
C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Pointers and Dynamic
Memory Management (cont.)

In C++:
• objects may be allocated both on the stack and in the heap
• there is no standard garbage collector
• the programmer is fully responsible for explicitly deallocating
memory.

This may lead to various programming errors:

• dangling reference problem - the result of a variable referencing a
memory block whose lifetime has expired (allocated on the stack and
then deallocated)
• memory leakage.

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Basic Pointers Operations

Pointer variables, or just pointers have names, types, and values.

The value of a pointer is the address of a memory block.
The type of a pointer determines the size of that block (the size is
essential to access data stored at that address).

For any data type T, for example int, you can define a variable of type
"pointer to T", such as "pointer to int“:

int* p; // pointer to int

char* s; // pointer to char

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Basic Pointers Operations
(cont.)

Dereferencing – using pointer’s name, prefixed with the asterisk *.

For example, *p is the contents of the memory block that p points to,
and *p behaves as an int variable.

You can take a non-pointer variable and get a pointer by applying the
address operator & to it:
int i = 1;
...i... // an int variable
...&i... // like int pointer, pointing to i
int* p;
p = &i; // p points to i
... *p ... // dereferenced p is equal to i
C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Basic Pointers Operations
(cont.)

C++ provides NULL, but rather than using it, use 0

Two common errors associated with C++ pointers is the use of

• un-initialized pointers,
• pointers that point to 0.

The language's run-time system does not provide the same safety as
Java's exception handling mechanism; instead, your program may
continue to execute following the use of an un-initialized pointer, and
crash unexpectedly at a later time.

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Constant Pointers

Read these pointer declarations right-to-left:

const int* p;
a pointer to an integer that is constant; the value of p may change,
but the value of *p cannot

int* const cp;

a constant pointer to integer; the value of *cp can change,
but the value of cp cannot

const int* const cpc;

a constant pointer to a constant integer.
C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Constant Pointers (cont).

In some exceptional cases you may want to remove const'ness:

const_cast<type>(value)

For example, to make pointer p

const int* p;

point to the value 3 (i.e. to assign 3 to *p), you can use

*(const_cast<int*>(p)) = 3;

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Pointers as Parameters

Array parameters are treated like pointers.

void maxMin(double* arr, int size, double& max,

double& min);
void maxMinConst(const double* arr, int size,
double& max, double& min);

const double a[] = {1.5, 3.1};

double max, min;
maxMin(a, 2, max, min); // a is const
maxMinConst(a, 2, max, min);
maxMin(const_cast<double*>(a), 2, max, min);
C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 Pointer Parameter

1. Functions that do not modify values pointed to by their pointer

parameters should have these parameters specified as const; for
example, const double*. This allows us to call the function
both with non-constant and constant pointer parameters.

2. Functions that modify their pointer parameters should have these

specified as passed by reference.

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Functions Returning
Pointers
int* change(int arr[], int size, int i) {
// return a pointer to the i-th element
if(i < 0 || i >= size)
return 0;
return &arr[i];
}
const int* nonChange (const int arr[], int size, int i);
... // as above
int x[] = {1, 2, 3};
int* i = change(x, 3, 2);
*i = 3; // OK, can modify through i
const int* j = nonChange(x, 3, 2);
cout << *j;
*j = 3; // can't assign to a const
i = nonChange(x, 3, 2); // can't: non-const
C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: C-style Strings and main
Function Parameters
C++ provides two kinds of strings:
string
char* - pointer to char, (C-style)

C-style strings cannot be avoided in the main function, and for file
operations
A C-style string can be passed as a parameter of a function:

int foo(char* str);

Within the body of this function, str is treated as an array of characters

(terminated by the zero character). The size of this array is the length
of the string, which can be obtained by using the strlen()
C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: C-style Strings and main
Function Parameters

Java C++ Comments about C++

void main( int main( Integer function with two parameters:

String[] int argc, - the number of command line
args) char*argv[]) parameters
- an array of C-style strings representing
parameters

The name of the program is taken from the command line as argv[0],
and therefore argc is always greater than zero

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: C-style Strings and main
Function Parameters
// File: ex2.5.cpp
// Copy words from the input file to the output
// file; passing filenames on the command line
// Output the total number of words to the standard output
// Return 1 if can't open input file;
// 2 if can't open output file
#include <iostream>
#include <fstream>
#include <string>
int main(int argc, char* argv[]) {
// check command line
if(argc != 3) {
cerr << "Usage: " << argv[0] << " file1 file2" << endl;
return 1;
}
C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 ifstream
ifstream inFile(argv[1]); and
if(!inFile) {
Example ofstream
cerr << "can't open " << argv[1] << endl; constructors
return 1; use C-style
} strings
ofstream outFile(argv[2]);
if(!outFile) {
cerr << "can't open " << argv[2] << endl;
return 2;
}
string word; // copy words
long wordCounter = 0;
while(inFile >> word) { // end-of-file
outFile << word << endl;
wordCounter++;
}
cout << "Read " << wordCounter << " words" << endl;
return 0;
}
C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Static Local Variables

Local variables in a function are allocated on the stack, in the so-

called instance, or frame for this function. It is sometimes useful
to limit the scope but not the lifetime of a variable.

void login() {
static int counter = 0;
counter++;
…
}

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: File Organization and
Standard Facilities

A Java program:
• typically consists of multiple files
• when you compile it, you usually use the name of single file,
which contains the main function to be invoked.
• the Java compiler checks what other files are needed, and whether
these files have been recently compiled. All files that have been
modified since the last compilation are recompiled, and finally,
all compiled files are linked.

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: File Organization and
Standard Facilities (cont.)

A C++ program:
• is a set of compilation units, which must be linked together to
form the executable code
• typically consists of multiple files; each file is a unit of
compilation.
• the programmer is responsible for compiling all the necessary
files
• you typically use separate compilation, and a tool such as make
to update the executable code.

Separate compilation: you can compile parts of your program,

where each part is a file
C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Preprocessing and Header
Files

Related function declarations are typically stored in header files.

You typically include the header file containing required
declarations.

Any line of C++ code that begins with the # (pound symbol) is
referred to as a preprocessor command line:
include command, which is used to include external files in the
current file
conditional compilation command, which tells the compiler that a
part of the source file is included conditionally

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Preprocessing and Header
Files (cont.)

Standard header files

#include <filename>

User-defined header files:

#include "filename.h"

(User-defined header file typically have the ‘h’ extension, for

example "screen.h".)
These two formats differ in how the specified file is located by the
compiler.

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Preprocessing and Header
Files (cont.)

Conditional compilation directive is based on the definition of a

name, called a macro, which specifies whether this file has
already been included. By convention, an include file screen.h
will have the macro name SCREEN_H. The header file might look
like this:

#ifndef SCREEN_H // conditional compilation

#define SCREEN_H
... // contents of the header file
#endif

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 Include Guards

The header file should be conditionally included; for example

#ifndef SCREEN_H // conditional compilation

#define SCREEN_H
... // contents of the header file
#endif

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002
 2: Namespaces

To avoid name collisions, C++ provides namespaces, (similar to

Java's packages):
Each namespace defines a new scope, and a name used in one
namespace does not conflict with the same name used in another
namespace.
The C++ standard library is defined in a namespace called std.

Practically, in all programs, you include the header file:

#include <iostream>
In order to make every name from this namespace directly available,
you specify the namespace with the using keyword:
using namespace std;

C++ Programming with Design Patterns Revealed Tomasz Müldner Copyright: Addison-Wesley Publishing Company, 2002