Thursday, December 9, 2010

An introduction to pointers

Pointers are an extremely powerful programming tool. They can make some things
much easier, help improve your program's efficiency, and even allow you to
handle unlimited amounts of data. For example, using pointers is one way to
have a function modify a variable passed to it. It is also possible to use
pointers to dynamically
allocate memory
, which means that you can write
programs that can handle nearly unlimited amounts of data on the fly--you
don't need to know, when you write the program, how much memory you need.
Wow, that's kind of cool. Actually, it's very cool, as we'll see in some of
the next tutorials. For now, let's just get a basic handle on what pointers
are and how you use them.

What are pointers? Why should you care?

Pointers are aptly named: they "point" to locations in memory. Think of a row
of safety deposit boxes of various sizes at a local bank. Each safety
deposit box will have a number associated with it so that the teller can quickly
look it up. These numbers are like the memory addresses of variables. A
pointer in the world of safety deposit boxes would simply be anything that
stored the number of another safety deposit box. Perhaps you have a rich
uncle who stored valuables in his safety deposit box, but decided to put the
real location in another, smaller, safety deposit box that only stored a card
with the number of the large box with the real jewelry. The safety deposit
box with the card would be storing the location of another box; it would be
equivalent to a pointer. In the computer, pointers are just variables that
store memory addresses, usually the addresses of other variables.

The cool thing is that once you can talk about the address of a variable, you'll
then be able to go to that address and retrieve the data stored in it. If you
happen to have a huge piece of data that you want to pass into a function,
it's a lot easier to pass its location to the function than to copy every
element of the data! Moreover, if you need more memory for your program, you
can request more memory from the system--how do you get "back" that memory?
The system tells you where it is located in memory; that is to say, you get a
memory address back. And you need pointers to store the memory address.

A note about terms: the word pointer can refer either to a memory
address itself, or to a variable that stores a memory address. Usually,
the distinction isn't really that important: if you pass a pointer variable
into a function, you're passing the value stored in the pointer--the memory
address. When I want to talk about a memory address, I'll refer to it as a
memory address; when I want a variable that stores a memory address, I'll call
it a pointer. When a variable stores the address of another variable, I'll
say that it is "pointing to" that variable.

Pointer Syntax

Pointers require a bit of new syntax because when you have a pointer, you need
the ability to request both the memory location it stores and the value stored
at that memory location. Moreover, since pointers are somewhat special, you
need to tell the compiler when you declare your pointer variable that the
variable is a pointer, and tell the compiler what type of memory it points to.

The pointer declaration looks like this:
<variable_type> *<name>; 

For example, you could declare a pointer that stores the address of an integer
with the following syntax:

int *points_to_integer;

Notice the use of the *. This is the key to declaring a pointer; if you add
it directly before the variable name, it will declare the variable to be a
pointer. Minor gotcha: if you declare multiple pointers on the same line,
you must precede each of them with an asterisk:

// one pointer, one regular int
int *pointer1, nonpointer1;

// two pointers
int *pointer1, *pointer2;

As I mentioned, there are two ways to use the pointer to access information:
it is possible to have it give the actual address to another variable. To do
so, simply use the name of the pointer without the *. However, to access the
actual memory location, use the *. The technical name for this doing this is
dereferencing the pointer; in essence, you're taking the reference to some
memory address and following it, to retrieve the actual value. It can be
tricky to keep track of when you should add the asterisk. Remember that the
pointer's natural use is to store a memory address; so when you use the


then it evaluates to the address. You have to add something extra, the
asterisk, in order to retrieve the value stored at the address. You'll
probably do that an awful lot. Nevertheless, the pointer itself is supposed
to store an address, so when you use the bare pointer, you get that address

Pointing to Something: Retrieving an Address

In order to have a pointer actually point to another variable it is necessary
to have the memory address of that variable also. To get the memory address of
a variable (its location in memory), put the & sign in front of the
variable name. This makes it give its address. This is called the address-of
operator, because it returns the memory address. Conveniently, both ampersand
and address-of start with a; that's a useful way to remember that you use
& to get the address of a variable.

For example:
#include <iostream>

using namespace std;

int main()
  int x;            // A normal integer
  int *p;           // A pointer to an integer

  p = &x;           // Read it, "assign the address of x to p"
  cin>> x;          // Put a value in x, we could also use *p here
  cout<< *p <<"\n"; // Note the use of the * to get the value

The cout outputs the value stored in x. Why is that? Well, let's look at the
code. The integer is called x. A pointer to an integer is then defined as p.
Then it stores the memory location of x in pointer by using the address-of
operator (&) to get the address of the variable. Using the ampersand is a
bit like looking at the label on the safety deposit box to see its number
rather than looking inside the box, to get what it stores. The user then
inputs a number that is stored in the variable x; remember, this is the same
location that is pointed to by p.

The next line then passes *p into cout. *p performs the "dereferencing"
operation on p; it looks at the address stored in p, and goes to that address
and returns the value. This is akin to looking inside a safety deposit box
only to find the number of (and, presumably, the key to ) another box, which
you then open.

Notice that in the above example, pointer is initialized to point to a
specific memory address before it is used. If this was not the case, it could
be pointing to anything. This can lead to extremely unpleasant consequences to
the program. For instance, the operating system will probably prevent you
from accessing memory that it knows your program doesn't own: this will cause
your program to crash. To avoid crashing your program, you should always
initialize pointers before you use them.

It is also possible to initialize pointers using free memory. This allows
dynamic allocation of array memory. It is most useful for setting up
structures called linked lists. This difficult topic is too complex for this
text. An understanding of the keywords new and delete will, however, be
tremendously helpful in the future.

The keyword new is used to initialize pointers with memory from free store (a
section of memory available to all programs). The syntax looks like the
int *ptr = new int;
It initializes ptr to point to a memory address of size int (because variables
have different sizes, number of bytes, this is necessary). The memory that is
pointed to becomes unavailable to other programs. This means that the careful
coder should free this memory at the end of its usage.

The delete operator frees up the memory allocated through new. To do so, the syntax is as in the example.

delete ptr;
After deleting a pointer, it is a good idea to reset it to point to 0. When 0
is assigned to a pointer, the pointer becomes a null pointer, in other words,
it points to nothing. By doing this, when you do something foolish with the
pointer (it happens a lot, even with experienced programmers), you find out
immediately instead of later, when you have done considerable damage.

In fact, the concept of the null pointer is frequently used as a way of
indicating a problem--for instance, some functions left over from C return 0
if they cannot correctly allocate memory (notably, the malloc function). You want to be sure to handle
this correctly if you ever use malloc or other C functions that return a "NULL
pointer" on failure.

In C++, if a call to new fails because the system is out of memory, then it
will "throw an exception". For the time being, you need not worry too much
about this case, but you can read more about what happens when new fails.

Taking Stock of Pointers

Pointers may feel like a very confusing topic at first but I think anyone can
come to appreciate and understand them. If you didn't feel like you absorbed
everything about them, just take a few deep breaths and re-read the lesson.
You shouldn't feel like you've fully grasped every nuance of when and why you
need to use pointers, though you should have some idea of some of their basic

No comments: