C++ Programming/Operators

From Wikibooks, the open-content textbooks collection

[edit] Operators

Operators are special symbols that are used to represent simple computations, this is significative importance in programming, since it serves to define the interaction of data in a useful way.

Computers are mathematical devices, but compilers and interpreters require a full syntactic theory of all operations in order to parse formulas involving any combinations correctly. In particular they depend on operator precedence rules, on order of operations, that are tacitly assumed in mathematical writing and the same applies to programming languages. Conventionally, the computing usage of operator also goes beyond the mathematical usage (for functions).

C++ like all programming languages uses a set of operators, they are subdivided into several groups:

• arithmetic operators (like addition and multiplication).
• boolean operators.
• string operators (used to manipulate strings of text).
• pointer operators.
• named operators (operators such as sizeof, new, and delete defined by alphanumeric names rather than a punctuation character).

Most of the operators in C++ do exactly what you would expect them to do, because they are common mathematical symbols. For example, the operator for adding two integers is +. C++ allows the re-definition of some operators (operator overloading) and this be covered later on.

The following are all legal expressions whose meaning is more or less obvious:

• 1+1
• hour-1
• hour*60 + minute
• minute/60

Take this line:

sum = a + b;

it uses the + operator to add the values stored in the locations a and b and the assignment operator (=) to store the result in the location sum. a and b are said to be the operands of +. The combination a + b is called an expression, specifically an arithmetic expression since + is an arithmetic operator. Similarly, = and its operands, sum and a + b together form the assignment expression sum = a + b (Note that the semicolon is not part of the expression). Other arithmetic operations that can be performed on integers (also common in many other languages) include:

• Subtraction, using the - operator
• Multiplication, using the * operator
• Division, using the / operator
• Remainder, using the % operator

Expressions can contain both variables names and integer values. In each case the name of the variable is replaced with its value before the computation is performed.

Addition, subtraction and multiplication all do what you expect, but you might be surprised by division. For example, the following program:

int hour, minute; 
hour = 11; 
minute = 59; 
std::cout << "Number of minutes since midnight: "; 
std::cout << hour*60 + minute << std::endl; 
std::cout << "Fraction of the hour that has passed: "; 
std::cout << minute/60 << std::endl;

would generate the following output:

Number of minutes since midnight: 719
Fraction of the hour that has passed: 0

The first line is what we expected, but the second line is odd. The value of the variable minute is 59, and 59 divided by 60 is 0.98333, not 0. The reason for the discrepancy is that C++ is performing integer division.

When both of the operands are integers (operands are the things operators operate on), the result must also be an integer, and by definition integer division always rounds down, even in cases like this where the next integer is so close.

A possible alternative in this case is to calculate a percentage rather than a fraction:

std::cout << "Percentage of the hour that has passed: "; 
std::cout << minute*100/60 << std::endl;

The result is:

Percentage of the hour that has passed: 98

Again the result is rounded down, but at least now the answer is approximately correct. In order to get an even more accurate answer, we could use a different type of variable, called floating-point, that is capable of storing fractional values.

[edit] Table of Operators

Operators in the same group have the same precedence and the order of evaluation is decided by the associativity (left-to-right or right-to-left). Operators in a preceding group have higher precedence than those in a subsequent group.

[edit] Order of operations

When more than one operator appears in an expression the order of evaluation depends on the rules of precedence. A complete explanation of precedence can get complicated, but just to get you started:

Multiplication and division happen before addition and subtraction. So 2*3-1 yields 5, not 4, and 2/3-1 yields -1, not 1 (remember that in integer division 2/3 is 0). If the operators have the same precedence they are evaluated from left to right. So in the expression minute*100/60, the multiplication happens first, yielding 5900/60, which in turn yields 98. If the operations had gone from right to left, the result would be 59*1 which is 59, which is wrong. Any time you want to override the rules of precedence (or you are not sure what they are) you can use parentheses. Expressions in parentheses are evaluated first, so 2 * (3-1) is 4. You can also use parentheses to make an expression easier to read, as in (minute * 100) / 60, even though it doesn't change the result.

Chaining Insertion Operators 

std::cout << "The sum of " << a << " and " << b << " is " << sum << "\n";

The line illustrates what is called chaining of insertion operators to print multiple expressions. How this works is as follows:

1. The leftmost insertion operator takes as its operands, std::cout and the string "The sum of ", it prints the latter using the former, and returns a reference to the former.
2. Now std::cout << a is evaluated. This prints the value contained in the location a, i.e. 123 and again returns std::cout.
3. This process continues. Thus, successively the expressions std::cout << " and ", std::cout << b, std::cout << " is ", std::cout << " sum ", std::cout << "\n" are evaluated and the whole series of chained values is printed.

[edit] Precedence (Composition)

At this point we have looked at some of the elements of a programming language like variables, expressions, and statements in isolation, without talking about how to combine them.

One of the most useful features of programming languages is their ability to take small building blocks and compose them (solving big problems by taking small steps at a time). For example, we know how to multiply integers and we know how to output values; it turns out we can do both at the same time:

std::cout << 17 * 3;

Actually, I shouldn't say "at the same time," since in reality the multiplication has to happen before the output, but the point is that any expression, involving numbers, characters, and variables, can be used inside an output statement. We've already seen one example:

std::cout << hour * 60 + minute << std::endl;

You can also put arbitrary expressions on the right-hand side of an assignment statement:

int percentage; 
percentage = ( minute * 100 ) / 60;

This ability may not seem so impressive now, but we will see other examples where composition makes it possible to express complex computations neatly and concisely.

 minute+1 = hour;

[edit] Chaining

std::cout << "The sum of " << a << " and " << b << " is " << sum << "\n";

The above line illustrates what is called chaining of insertion operators to print multiple expressions. How this works is as follows:

1. The leftmost insertion operator takes as its operands, std::cout and the string "The sum of ", it prints the latter using the former, and returns a reference to the former.
2. Now std::cout << a is evaluated. This prints the value contained in the location a, i.e. 123 and again returns std::cout.
3. This process continues. Thus, successively the expressions std::cout << " and ", std::cout << b, std::cout << " is ", std::cout << " sum ", std::cout << "\n" are evaluated and the whole series of chained values is printed.

[edit] Assignment

The most basic assignment operator is the "=" operator. It assigns one variable to have the value of another. For instance, the statement x = 3 assigns x the value of 3, and y = x assigns whatever was in x to be in y. When the "=" operator is used to assign a class or struct, it acts like using the "=" operator on every single element. For instance:

//Example to demonstrate default "=" operator behavior.
 
struct A
 {
  int i;
  float f;
  A * next_a;
 };
 
//Inside some function
 {
  A a1, a2;              // Create two A objects.
 
  a1.i = 3;              // Assign 3 to i of a1.
  a1.f = 4.5;            // Assign the value of 4.5 to f in a1
  a1.next_a = &a2;       // a1.next_a now points to a2
 
  a2.next_a = NULL;      // a2.next_a is guaranteed to point at nothing now.
  a2.i = a1.i;           // Copy over a1.i, so that a2.i is now 3.
  a1.next_a = a2.next_a; // Now a1.next_a is NULL
 
  a2 = a1;               // Copy a2 to a1, so that now a2.f is 4.5. The other two are unchanged, since they were the same.
 }

Assignments can also be chained since the assignment operator returns the value it assigns. But this time the chaining is from right to left. For example, to assign the value of z to y and assign the same value (which is returned by the = operator) to x you use:

x = y = z;

When the "=" operator is used in a declaration, it has special meaning. It tells the compiler to directly initialize the variable from whatever is on the right-hand side of the operator. This is called defining a variable, in the same way that you define a class or a function. With classes, this can make a difference, especially when assigning to a function call:

class A { /* ... */ };
A foo () { /* ... */ };
 
// In some function
 {
  A a;
  a = foo();
 
  A a2 = foo();
 }

In the first case, a is constructed, then is changed by the "=" operator. In the second statement, a2 is constructed directly from the return value of foo(). In many cases, the compiler can save a lot of time by constructing foo()'s return value directly into a2's memory, which makes the program run faster.

Whether or not you define can also matter in a few cases where a definition can result in different linkage, making the variable more or less available to other source files.

[edit] Arithmetic Operators

sum = a + b;

The line above uses the + operator to add the values stored in the locations a and b and the assignment operator (=) to store the result in the location sum. a and b are said to be the operands of +. The combination a + b is called an expression, specifically an arithmetic expression since + is an arithmetic operator. Similarly, = and its operands, sum and a + b together form the assignment expression sum = a + b (Note that the semicolon is not part of the expression). Other arithmetic operations that can be performed on integers (also common in many other languages) include:

• Subtraction, using the - operator
• Multiplication, using the * operator
• Division, using the / operator
• Remainder, using the % operator

The multiplicative operators *, / and % are always evaluated before the additive operators + and -. Among operators of the same class, evaluation proceeds from left to right. This order can be overridden using grouping by parentheses, ( and ); the expression contained within parentheses is evaluated before any other neighboring operator is evaluated. But note that some compilers may not strictly follow these rules when they try to optimize the code being generated, unless violating the rules would give a different answer.

For example the following statements convert a temperature expressed in degrees Celsius to degrees Fahrenheit and vice versa:

deg_f = deg_c * 9 / 5 + 32;
deg_c = ( deg_f - 32 ) * 5 / 9;

[edit] Compound Assignment

One of the most common patterns in software with regards to operators is to update a value:

a = a + 1;
b = b * 2;
c = c / 4;

Since this pattern is used many times, there is a shorthand for it called compound assignment operators. They are a combination of an existing arithmetic operator and assignment operator:

• +=
• -=
• *=
• /=
•  %=
• <<=
• >>=
• |=
• &=
• ^=

Thus the example given in the beginning of the section could be rewritten as

a += 1;  // Equivalent to (a = a + 1)
b *= 2;  // Equivalent to (b = b * 2)
c /= 4;  // Equivalent to (c = c / 4)

TODO Parent topic may need a re-writing. About optimization and distinction on the steps in: a += 1, a = a + 1, ++a or a++.

[edit] Character Operators

Interestingly, the same mathematical operations that work on integers also work on characters.

char letter; 
letter = 'a' + 1; 
std::cout << letter << std::endl;

For the above example, outputs the letter b (on most systems -- note that C++ doesn't assume use of ASCII, EBCDIC, Unicode etc. but rather allows for all of these and other charsets). Although it is syntactically legal to multiply characters, it is almost never useful to do it.

Earlier I said that you can only assign integer values to integer variables and character values to character variables, but that is not completely true. In some cases, C++ converts automatically between types. For example, the following is legal.

int number; 
number = 'a'; 
std::cout << number << std::endl;

On most mainstream desktop computers the result is 97, which is the number that is used internally by C++ on that system to represent the letter 'a'. However, it is generally a good idea to treat characters as characters, and integers as integers, and only convert from one to the other if there is a good reason. Unlike some other languages, C++ does not make strong assumptions about how the underlying platform represents characters; ASCII, EBCDIC and others are possible, and portable code will not make assumptions (except that '0', '1', ..., '9' are sequential, so that e.g. '9'-'0' == 9).

Automatic type conversion is an example of a common problem in designing a programming language, which is that there is a conflict between formalism, which is the requirement that formal languages should have simple rules with few exceptions, and convenience, which is the requirement that programming languages be easy to use in practice.

More often than not, convenience wins, which is usually good for expert programmers, who are spared from rigorous but unwieldy formalism, but bad for beginning programmers, who are often baffled by the complexity of the rules and the number of exceptions. In this book I have tried to simplify things by emphasizing the rules and omitting many of the exceptions.

[edit] Bitwise Operators

These operators deal with a bitwise operations. Bit operations needs the understanding of binary numeration since it will deal with on one or two bit patterns or binary numerals at the level of their individual bits. On most microprocessors, bitwise operations are sometimes slightly faster than addition and subtraction operations and usually significantly faster than multiplication and division operations.

Bitwise operations especially important for much low-level programming from optimizations to writing device drivers, low-level graphics, communications protocol packet assembly and decoding.

Although machines often have efficient built-in instructions for performing arithmetic and logical operations, in fact all these operations can be performed just by combining the bitwise operators and zero-testing in various ways.

The bitwise operators work bit by bit on the operands. The operands must be of integral type (one of the types used for integers).

For this section, recall that a number starting with 0x is hexadecimal (hexa, or hex for short or referred also as base-16). Unlike the normal decimal system using powers of 10 and the digits 0123456789, hex uses powers of 16 and the symbols 0123456789abcdef. In the examples remember that Oxc equals 1100 in binary and 12 in decimal. C++ does not directly support binary notation, which would hamper readability of the code.

NOT

~a  

bitwise complement of a.

~0xc produces the value -1-0xc (in binary, ~1100 produces ...11110011 where "..." may be many more 1 bits)

The negation operator is a unary operator which precedes the operand, This operator must not be confused with the "logical not" operator, "!" (exclamation point), which treats the entire value as a single Boolean—changing a true value to false, and vice versa. The "logical not" is not a bitwise operation.

These others are binary operators which lie between the two operands. The precedence of these operators is lower than that of the relational and equivalence operators; it is often required to parenthesize expressions involving bitwise operators.

AND

a & b 

bitwise boolean and of a and b

0xc & 0xa produces the value 0x8 (in binary, 1100 & 1010 produces 1000)

OR

a | b 

bitwise boolean or of a and b

0xc | 0xa produces the value 0xe (in binary, 1100 | 1010 produces 1110)

XOR

a ^ b 

bitwise xor of a and b

0xc ^ 0xa produces the value 0x6 (in binary, 1100 ^ 1010 produces 0110)

Bit shifts

a << b 

shift a left by b (multiply a by 2^b)

0xc << 1 produces the value 0x18 (in binary, 1100 << 1 produces the value 11000)

a >> b 

shift a right by b (divide a by 2^b)

0xc >> 1 produces the value 0x6 (in binary, 1100 >> 1 produces the value 110)

[edit] Derived Types Operators

There are three data types known as pointers, references, and arrays, that have their own operators for dealing with them. Those are *, &, [], ->, .*, and ->*.

Pointers, references, and arrays are fundamental data types that deal with accessing other variables. Pointers are used to pass around a variables address (where it is in memory), which can be used to have multiple ways to access a single variable. References are aliases to other objects, and are similar in use to pointers, but still very different. Arrays are large blocks of contiguous memory that can be used to store multiple objects of the same type, like a sequence of characters to make a string.

[edit] Subscript Operator "[]"

This operator is used to access an object of an array. It is also used when declaring array types, allocating them, or deallocating them.

[edit] Arrays

Arrays store a constant-sized sequential set of blocks, each block containing a value of the elected type under a single name. Arrays often help organize collections of data efficiently and intuitively.

It is easiest to think of an array as simply a list with each value as an item of the list. Where individual elements are accessed by their position in the array called its index, also known as subscript.

Since an array stores values, what type of values and how many values to store must be defined as part of an array declaration, so it can allocate the needed space. The size of array must be a const integral expression greater than zero. That means that you cannot use user input to declare an array. You need to allocate the memory (with operator new[]), so the size of an array has to be known at compile time. Another disadvantage of the sequential storage method is that there has to be a free sequential block large enough to hold the array. If you have an array of 500,000,000 blocks, each 1 byte long, you need to have roughly 500 megabytes of sequential space to be free; Sometimes this will require a defragmentation of the memory, which takes a long time.

To declare an array you can do:

int numbers[30]; // creates an array of 30 integers

or

char letters[4]; // create an array of 4 characters

and so on...

to initialize as you declare them you can use:

int vector[6]={0,0,1,0,0,0};

this will not only create the array with 6 int elements but also initialize them to the given values.

Access a value stored in an array is easy. For example with the above declaration,

int x;
x = vector[2];

will assign x the valued store at index 2 of variable vector which is 1.

Arrays are indexed starting at 0, as opposed to starting at 1. The first element of the array above is vector[0]. The index to the last value in the array is the array size minus one. In the example above the subscripts run from 0 through 5. C++ does not do bounds checking on array accesses. The compiler will not complain about the following:

char y;
int z = 9;
char vector[6] = { 1, 2, 3, 4, 5, 6 };
 
// examples of accessing outside the array. A compile error is not raised
y = vector[15];
y = vector[-4];
y = vector[z];

During program execution, an out of bounds array access does not always cause a run time error. Your program may happily continue after retrieving a value from vector[-1]. To alleviate indexing problems, the sizeof() expression is commonly used when coding loops that process arrays.

int ix;
short anArray[]= { 3, 6, 9, 12, 15 };
 
for (ix=0; ix< (sizeof(anArray)/sizeof(short)); ++ix) {
  DoSomethingWith( anArray[ix] );
}

Notice in the above example, the size of the array was not explicitly specified. The compiler knows to size it at 5 because of the five values in the initializer list. Adding an additional value to the list will cause it to be sized to six, and because of the sizeof expression in the for loop, the code automatically adjusts to this change.

You can also use multi-dimensional arrays. The simplest type is a two dimensional array. This creates a rectangular array - each row has the same number of columns. To get a char array with 3 rows and 5 columns we write...

char two_d[3][5];

To access/modify a value in this array we need two subscripts:

char ch;
ch = two_d[2][4];

or

two_d[0][0] = 'x';

There are also weird notations possible:

int a[100];
int i = 0;
if (a[i]==i[a])
  printf("Hello World!\n");

a[i] and i[a] point to the same location. You will understand this better after knowing about pointers.

To get an array of a different size, you must explicitly deal with memory using realloc, malloc, memcpy, etc.

Advantages of arrays include:

• Random access in O(1) (Big O notation)
• Ease of use/port: Integrated into most modern languages

Disadvantages include:

• Constant size
• Constant data-type
• Large free sequential block to accommodate large arrays
• When used as non-static data members, the element type must allow default construction
• Arrays do not support copy assignment (you cannot write arraya = arrayb)
• Arrays cannot be used as the value type of a standard container
• Syntax of use differs from standard containers
• Arrays and inheritance don't mix (an array of Derived is not an array of Base, but can too easily be treated like one)

[edit] Why no bounds checking on array indexes?

C++ does allow for, but doesn't force, bounds-checking implementations, in practice little or no checking is done. It affects storage requirements (needing "fat pointers") and impacts runtime performance. However, the std::vector template class as we will see is an object representing an array, and it provides the at() method, which does enforce bounds checking. Also in many implementations, the standard containers include particularly complete bounds checking in debug mode. They might not support these checks in release builds, as any performance reduction in container classes relative to built-in arrays might prevent programmers from migrating from arrays to the more modern, safer container classes.

[edit] address-of operator "&"

To get the address of a variable so that you can assign a pointer, you use the "address of" operator, which is denoted by the ampersand & symbol. The "address of" operator does exactly what it says, it returns the "address of" a variable, a symbolic constant, or a element in an array, in the form of a pointer of the corresponding type. To use the "address of" operator, you tack it on in front of the variable that you wish to have the address of returned. It is also used when declaring reference types.

Now, do not confuse the "address of" operator with the declaration of a reference. Because use of operators is restricted to expression, the compiler knows that &sometype is the "address of" operator being used to denote the return of the address of sometype as a pointer.

[edit] References

References are a way of assigning a "handle" to a variable. References can also be thought of as "aliases"; they're not real objects, they're just alternative names for other objects.

Assigning References

This is the less often used variety of references, but still worth noting as an introduction to the use of references in function arguments. Here we create a reference that looks and acts like a standard variable except that it operates on the same data as the variable that it references.

int tZoo = 3;       // tZoo == 3
int &refZoo = tZoo; // tZoo == 3
refZoo = 5;         // tZoo == 5

refZoo is a reference to tZoo. Changing the value of refZoo also changes the value of tZoo.

For example say we want to have a function to swap 2 integers

void swap(int &a, int &b){
  int temp = a; 
  a = b; 
  b = temp;
}

int main(){
   int x = 5; 
   int y = 6; 
   int &refx = x; 
   int &refy = y; 
   swap(refx, refy); // now x = 6 and y = 5
   swap(x, y); // and now x = 5 and y = 6 again
}

[edit] Dereferencing Operator "*"

This operator is used to get the variable pointed to by a pointer. It is also used when declaring pointer types.

[edit] Pointers

Pointers are important data types due to special characteristics. They may be used to indicate a variable without actually creating a variable of that type. They can be a difficult concept to understand, some special effort should be spent on understanding the power they give to programmers.

Pointers have a very descriptive name. Pointers variables only store memory addresses, usually the addresses of other variables. Essentially, they point to another variable memory location, a reserved location on the computer memory. You can use a pointer to pass the location of a variable to a function, this enables the function's pointer to use the variable space, so that it can retrieve or modify its data. You can even have pointers to pointers, and pointers to pointers to pointers and so on and so forth.

[edit] Declaring Pointers

Pointers are declared by adding a * before the variable name in the declaration, as in the following example:

int* x;  // pointer to int.
int * y; // pointer to int. (legal, but rarely used)
int *z;  // pointer to int.
int*i;   // pointer to int. (legal, but rarely used)

// C codestyle
int *z;
 
// C++ codestyle
int* z;

Watch out, though, because the * associates to the following declaration only:

int* i, j;  // CAUTION! i is pointer to int, j is int.
int *i, *j; // i and j are both pointer to int.

You can also have multiple pointers chained together, as in the following example:

int **i;  // Pointer to pointer to int.
int ***i; // Pointer to pointer to pointer to int (rarely used).

[edit] Null Pointer

The null pointer is a very special pointer. It means that the pointer points to absolutely nothing. It is an error to attempt to dereference (using the * or -> operators) a null pointer. A null pointer can be referred to using the constant zero, as in the following example:

int i;
int *p;
 
p = 0; //Null pointer.
p = &i; //Not the null pointer.

Note that you can't assign a pointer to an integer, even if it's zero. It has to be the constant. The following code is an error:

int i = 0;
int *p = i; //Error: 0 only evaluates to null if it's a pointer

There is a macro defined in the standard library called NULL, which is always equal to a null pointer value (essentially, 0). It is good practice to always use NULL when referring to the null pointer, as it strongly improves readability.

Since a null pointer is 0, it will always compare to 0. Like an integer, if you use it in a true/false expression, it will return false if it is the null pointer, and true if it's anything else:

#include <iostream>
 
void IsNull (int * p)
{
  if (p)
    std::cout<<"Pointer is not NULL"<<std::endl;
  else
    std::cout<<"Pointer is NULL"<<std::endl;
}
 
int main()
{
  int * p;
  int i;
 
  p = NULL;
  IsNull(p);
  p = &i;
  IsNull(&i);
  IsNull(p);
  IsNull(NULL);
 
  return 0;
}

This program will output that the pointer is NULL, then that it isn't NULL twice, then again that it is.

TODO Make short introduction to pointers as data members (so it can be cross linked from the function and class sections of the texts)

[edit] Pointer Indirection Operator "->"

This operator is used to access a member of a class pointer.

[edit] Pointer-to-Member Dereferencing Operator ".*"

This operator is used to access the variable associated with a specific class instance, given an appropriate pointer.

[edit] Pointer-to-Member Indirection Operator "->*"

This operator is used to access the variable associated with a class instance pointed to by one pointer, given another pointer-to-member that's appropriate.

[edit] Pointers to functions

When used to point to functions, pointers can be exceptionally powerful. A call can be made to a function anywhere in the program, knowing only what kinds of parameters it takes. Pointers to functions are used several times in the standard library, and provide a powerful system for other libraries which need to adapt to any sort of user code. This case is examined more in depth in the Functions Section of this book.

[edit] Dereferencing

Now that you have a pointer, you need some way to access the memory that it points to. This is the * operator. When it's put in front of a pointer, it gives the variable pointed to. This is an lvalue, so you can assign values to it, or even initialize a reference from it.

#include <iostream>
 
int main()
{
  int i;
  int * p = &i;
  i = 3;
 
  std::cout<<*p<<std::endl; // prints "3"
 
  return 0;
}

Since the result of an & operator is a pointer, *&i is valid, though it has absolutely no effect.

Now, when you combine the * operator with classes, you may notice a problem. It has lower precedence than .! See the example:

struct A { int num; };
 
A a;
int i;
A * p;
 
p = &a;
a.num = 2;
 
i = *p.num; // Error! "p" isn't a class, so you can't use "."
i = (*p).num;

The error happens because the compiler looks at p.num first ("." has higher precedence than "*") and because p does not have a member named num the compiler gives you an error. Using grouping symbols to change the precedence gets around this problem.

It would be very time-consuming to have to write (*p).num a lot, especially when you have a lot of classes. Imagine writing (*(*(*(*MyPointer).Member).SubMember).Value).WhatIWant! As a result, a special operator, ->, exists. Instead of (*p).num, you can write p->num, which is completely identical for all purposes. Now you can write MyPointer->Member->SubMember->Value->WhatIWant. It's a lot easier on the brain!

[edit] sizeof()

The sizeof operator works at compile time to report on the number of bytes of storage occupied by a type (equivalently, by a variable of that type).

Syntactically, sizeof appears like a function call when taking the size of a type, but may be used without parentheses when taking the size of an object. Style guidelines vary on whether using the latitude to omit parentheses in the latter case is desirable.

sizeof has also found new life in recent years in template meta programming in C++, where the fact that it can turn types into numbers, albeit in a primitive manner, is often useful, given that the template metaprogramming environment of C++ typically does most of its calculations with types.

//Examples of sizeof use
std::size_t int_size( sizeof( int ) );// Might give 1, 2, 4, 8 or other values.
 
// or
 
int answer( 42 );
std::size_t answer_size( sizeof( answer ) );// Same value as sizeof( int )
std::size_t answer_size( sizeof answer);    // Equivalent syntax

Note that sizeof measures the size of an object in the simple sense of a contiguous area of storage; for types which include pointers to other storage, the indirect storage is not included in the number of bytes returned by sizeof. A common mistake made by programming newcomers working with C++ is to try to use sizeof to determine the length of a string; the std::strlen or std::string::length functions are more appropriate for that task.

[edit] Dynamic Memory Allocation

Dynamic memory allocation is the allocation of memory storage for use in a w:computer program during the runtime of that program. It is a way of distributing ownership of limited memory resources among many pieces of data and code. Importantly, the amount of memory allocated is determined by the program at the time of allocation and need not be known in advance. A dynamic allocation exists until it is explicitly released, either by the programmer or by a garbage collector implementation; this is notably different from automatic and static memory allocation, which require advance knowledge of the required amount of memory and have a fixed duration. It is said that an object so allocated has dynamic lifetime.

The task of fulfilling an allocation request, which involves finding a block of unused memory of sufficient size, is complicated by the need to avoid both internal and external fragmentation while keeping both allocation and deallocation efficient. Also, the allocator's metadata can inflate the size of (individually) small allocations; chunking attempts to reduce this effect.

Usually, memory is allocated from a large pool of unused memory area called the heap (also called the free store). Since the precise location of the allocation is not known in advance, the memory is accessed indirectly, usually via a reference. The precise algorithm used to organize the memory area and allocate and deallocate chunks is hidden behind an abstract interface and may use any of the methods described below.

You have probably wondered how programmers allocate memory efficiently without knowing, prior to running the program, how much memory will be necessary. Here is when the fun starts with dynamic memory allocation.

[edit] new and delete

For dynamic memory allocation we use the new and delete keywords, the old malloc from C functions can now be avoided but are still accessible for compatibility and low level control reasons.

TODO add info on malloc

As covered before, we assign values to pointers using the "address of" operator because it returns the address in memory of the variable or constant in the form of a pointer. Now, the "address of" operator is NOT the only operator that you can use to assign a pointer. You have yet another operator that returns a pointer, which is the new operator. The new operator allows the programmer to allocate memory for a specific data type, struct, class, etc, and gives the programmer the address of that allocated sect of memory in the form of a pointer. The new operator is used as an rvalue, similar to the "address of" operator. Take a look at the code below to see how the new operator works.

By assigning the pointers to an allocated sector of memory, rather than having to use a variable declaration, you basically override the "middleman" (the variable declaration). Now, you can allocate memory dynamically without having to know the number of variables you should declare.

int n = 10; 
SOMETYPE *parray, *pS; 
int *pint; 
 
parray = new SOMETYPE[n]; 
pS = new SOMETYPE; 
pint = new int;

If you looked at the above piece of code, you can use the new operator to allocate memory for arrays too, which comes quite in handy when we need to manipulate the sizes of large arrays and or classes efficiently. The memory that your pointer points to because of the new operator can also be "deallocated," not destroyed but rather, freed up from your pointer. The delete operator is used in front of a pointer and frees up the address in memory to which the pointer is pointing.

delete [] parray;// note the use of [] when destroying an array allocated with new
delete pint;

The memory pointed to by parray and pint have been freed up, which is a very good thing because when you're manipulating multiple large arrays, you try to avoid losing the memory someplace by leaking it. Any allocation of memory needs to be properly deallocated or a leak will occur and your program won't run efficiently. Essentially, every time you use the new operator on something, you should use the delete operator to free that memory before exiting. The delete operator, however, not only can be used to delete a pointer allocated with the new operator, but can also be used to "delete" a null pointer, which prevents attempts to delete non-allocated memory (this actions compiles and does nothing).

You must keep in mind that new T and new T() are not equivalent. This will be more understandable after you are introduced to more complex types like classes, but keep in mind that when using new T() it will initialize the T memory location ("zero out") before calling the constructor (if you have non-initialized members variables, they will be initialized by default).

The new and delete operators do not have to be used in conjunction with each other within the same function or block of code. It is proper and often advised to write functions that allocate memory and other functions that deallocate memory. Indeed, the currently favored style is to release resources in object's destructors, using the so-called resource acquisition is initialization (RAII) idiom.

TODO Move or split some of the information or add references, classes, destructor and constructors have yet to be introduced and bellow we are using a vector for the example

As we will see when we get to the Classes, a class destructor is the ideal location for its deallocator, it is often advisable to leave memory allocators out of classes' constructors. Specifically, using new to create an array of objects, each of which also uses new to allocate memory during its construction, often results in runtime errors. If a class or structure contains members which must be pointed at dynamically-created objects, it is best to sequentially initialize arrays of the parent object, rather than leaving the task to their constructors.

// Example of a dynamic array
 
const int b = 5;
int *a = new int[b];
 
//to delete
delete[] a;

The ideal way is to not use arrays at all, but rather the STL's vector type (a container similar to an array). To achieve the above functionality, you should do:

const int b = 5;
std::vector<int> a;
a.resize(b);
 
//to delete
a.clear();

Vectors allow for easy insertions even when "full." If, for example, you filled up a, you could easily make room for a 6th element like so:

int new_number = 99;
a.push_back( new_number );//expands the vector to fit the 6th element

You can similarly dynamically allocate a rectangular multidimensional array (be careful about the type syntax for the pointers):

const int d = 5;
int (*two_d_array)[4] = new int[d][4];
 
//to delete
delete[] two_d_array;

You can also emulate a ragged multidimensional array (sub-arrays not the same size) by allocating an array of pointers, and then allocating an array for each of the pointers. This involves a loop.

const int d1 = 5, d2 = 4;
int **two_d_array = new int*[d1];
for( int i = 0; i < d1; ++i)
  two_d_array[i] = new int[d2];
 
//to delete
for( int i = 0; i < d1; ++i)
  delete[] two_d_array[i];
 
delete[] two_d_array;