Help With pointers in classes.

I need a little help on understanding when pointers are involved in classes. Here is an example code.

Code:

---

// Listing 8.10
// Using pointers with const methods

#include <iostream>
using namespace std;


class Rectangle
{
public:
        Rectangle();
        ~Rectangle();
        void SetLength(int length) {itsLength = length;}
        int GetLength() const {return itsLength;}
        void SetWidth(int width) {itsWidth = width;}
        int GetWidth() const { return itsWidth;}

private:
        int itsLength;
        int itsWidth;

};

Rectangle::Rectangle()
{
        itsWidth = 5;
        itsLength =10;
}

Rectangle::~Rectangle()
{}

int main()
{
        Rectangle* pRect = new Rectangle;
        const Rectangle * pConstRect = new Rectangle;
        Rectangle * const pConstPtr = new Rectangle;

        cout << "pRect width: " << pRect->GetWidth()
                << " feet\n";
        cout << "pConstRect width: " << pConstRect->GetWidth()
                << " feet\n";
        cout << "pConstPtr width: " << pConstPtr ->GetWidth()
                << " feet\n";

        pRect->SetWidth(10);
        // pConstRect->SetWidth(10);
        pConstPtr->SetWidth(10);

        cout << "[Rect width: " << pRect->GetWidth()
                << "feet\n";
        cout << "pConstRect width: " << pConstRect->GetWidth()
                << " feet\n";
        cout << "pConstPtr width: " << pConstPtr->GetWidth()
                << " feet\n";
   

        return 0;

}

---

I am wondering about the lines:

Code:

---

Rectangle* pRect = new Rectangle;
const Rectangle * pConstRect = new Rectangle;
Rectangle * const pConstPtr = new Rectangle;

---

Do they mean that basically instead of declaring an object as :

Code:

---

rectangle pRect;

---

and so on, it just declares the object as a pointer? It is a little confusing to me when the difference operator is involved. Also, I am not too clear as to the use of new keyword.

Speaking from your perception, pointers and the new instruction gives you absolute control. For example, using new, you could allocate memory for a Rectangle object and delete it later at any point in the program.

Code:

---


// In this function, rec will get destroyed after the function returns.
void func()
{
Rectangle rec;
...
}

// In thisfunction, pRec will NOT be destroyed pass function scope.
void func()
{
Rectangle *pRec = new Rectangle;
...
}

// Note pRec still exists.

---

Kuphryn

oh I see. 1So basically what

Code:

---

Rectangle *pRec = new Rectangle;

---

is saying is creating an object that is of the class Rectangle and is also a pointer to a new location in Rectangle?

just how it is worded is what i need clarified, also about the.

Quote:

---

I am wondering about the lines:

Code:

---

Rectangle* pRect = new Rectangle;
const Rectangle * pConstRect = new Rectangle;
Rectangle * const pConstPtr = new Rectangle

---

---

The difference between the above declarations and

Code:

---

rectangle pRect;

---

is that the first 3 declaration allocates memory dynamicly and in the second code fragment it is a static memory allocation. What does static and dynamic allocation mean??? Static declaration means that all objects are allocated at compile-time and dynamic declaration mean that objects is allocated at run-time (during program execution).

Quote:

---

Also, I am not too clear as to the use of new keyword.

---

When you use the new keyword in your code it allocates memory(dynamic) on the free store ( or heap) and returns an address. Because objects on the free store are unnamed we have to use a pointer(s) to acces that part of memory. It is very important that we deallocate that part of memory or else we have created a memory leak.

Example code

Code:

---

#include <iostream>

using namespace std;

int main()
{
//normal static declaration
int inumber = 20;

//This is not good, allocated dynamic memory but don't have access to it, memory leak
//new int(10);

//This is better
int *pint = new int(10);
//Clean up after you, avoid memory leak
delete pint;
return 0;
}

---

oh I see. 1So basically what

Code:

---

Rectangle *pRec = new Rectangle;

---

is saying is creating an object that is of the class Rectangle and is also a pointer to a new location in Rectangle?

just how it is worded is what i need clarified.

Let me clairify it more

Code:

---

Rectangle *pRec;

---

This declares a pointer to point to type Rectangle. Because it´s uninitialized it points to something (DANGEROUS)

Code:

---

Rectangle *pRec = new Rectangle;

---

This does the exact same thing and initialize the pointer to a Rectangle-object (or rather the address) on the free store. If Rectangle is a user defined object it´s corresponing constructors are invoked (triggered by new).

I hope that I don´t confuse you!

not too confusing, so basically when pointing like

Code:

---

Rectangle *pRec;

---

since it is not declared, it is not pointing to a specified address. And what did you mean by

Quote:

---

If Rectangle is a user defined object it´s corresponing constructors are invoked (triggered by new).

---

When the pointer is first declared it does not point to anything specific. Thus you should always point to NULL.

int *pNum = NULL;

The key to understanding pointers is knowing that pointers refer directly to memory addresses or to other pointers. That is what pointers so powerful.

Kuphryn

Quote:

---

And what did you mean by

Quote:

---

If Rectangle is a user defined object it´s corresponing constructors are invoked (triggered by new).

---

---

When you create a class (like yours Rectangle), all objects of that class are called user defind objects and every time you create such an object the constructor is invoked. And likewise when an object is deleted the destructor is invoked. This all is done implicity (automatic), so you don´t have to invoke them.

This is the construtor

Code:

---

Rectangle::Rectangle()
{
        itsWidth = 5;
        itsLength =10;
}

---

And this is the destructor

Code:

---

Rectangle::~Rectangle()
{}

---

The keyword new creates an object (dynamicly).

so by saying

Code:

---

Rectangle* pRect = new Rectangle;

---

as opposed to saying

Code:

---

Rectangle myRect;

---

you are saying that pRect is a pointer to a Newly assigned address of the Rectangle class?

First, lose the barrier between the datatype you created (rectangle) and any other datatype. They are all just datatypes and pointers work the same for all of them.

int *intPtr = new int;
char *charPtr = new char;
foo *fooPtr = new foo ();

Another idea to get into your mind. A pointer is a datatype, it's purpose is to hold a memory address.

The keyword "new" allocates the amount of memory needed for the datatype you are instantiating.

fooPtr = new foo (); // we are assigning fooPtr the memory address of the new foo object.

To access the object you need to dereference it by using the *

*fooPtr->FooPrintMethod ();

fooPtr->FooPrintMethod (); // bad!...you'll get yelled at by the compiler ;)
*fooPtr->FooPrintMethod (); // good!

Make anymore sense?

Code:

---

class Rectangle
{
  public:
    Rectangle ();
    Rectangle (int, int);
    int width;
    int length;
};

Rectangle::Rectangle()
{
  width = 0;
  length = 0;
}

Rectangle::Rectangle(int w, int len)
{
  width = w;
  lenght = len;
}

---

Using above class:
Rectangle rect; //declares on object called rect of type Rectangle with memory for rect allocated on the project stack using static memory using the default constructor. Width and lenght are bothassigned 0; This could also be written as this:
REctangle rect();

Rectangle rect1(3, 4);//declares an object called rect1 of type Rectangle with memory for rect1 allocated on the project stack using static memory using the constructor with two ints for arguments. Width is assigned w, 3 in this case, and length is assigned len, 4 in this case.

Rectangle * pRect;//declares a pointer to type Rectangle called pRect. It is pointing to something, but who know what. Using pRect as it is is dangerous because we don't know what it is pointing to.

Rectangle * pRect = NULL; //now pRect is pointing to a gauranteed memory spot called NULL. You can't use this memory spot for anything, but at least now pRect is not pointing to something it's not supposed to be pointing to.

Rectangle * pRect = ▭ //this time pRect is declared and initialzed to the address of rect.

Rectangle * pRect = new Rectangle;//now pRect is pointing to memory allocated on the heap(heap is also called free store and dynamic memory) for an object of type Rectangle. The object is declared using the default constructor. It could also be written like this:
Rectangle * pRect = new Rectangle();

Rectangle * pRect = new Rectangle(3, 4);//now pRect is pointing to memory allocated on the heap for an object of type Rectangle that is constructed using the two int constructor.

const Rectangle rect(3, 4);//now the value of rect can not be changed from width of 3 and lenght of 4

From here on out I _hope_ I get it correct. If not I'm sure someone will help out.

The various possible uses of keyword const with pointers:
const Rectangle * pRect;
Rectangle * const pRect1;
const Rectangle * const pRect;

pRect is a pointer to a constant Rectangle. The value of width and length in the Rectangle that is being pointed to can't be changed.

pRect1 is a constant pointer to a Rectangle. The value of width and length can be changed, but pRect1 can't point to anything else but the address it is pointing to.

pRect2 is a constant pointer to a constant Rectangle. The value of width and length in the Rectangle that is being pointed to can't be changed, and pRect2 can't be changed to point to anything else.

You can use all the flavors of keyword const noted above with dynamic memory (ie with the new keyword) just as you can non-constant pointers, etc.

And, last, but not least, to access public members for an object from outside the class you would use the dot operator like this:

Rectangle rect(3, 4);
cout << rect.width << ',' << rect.length;

and for a pointer to rect, irrespective of whether pointer is pointing to static or dynamic memory;

Rectangle * pRect = &rect;
cout << (*pRect).width << ','; //using dereference and dot operatorin combination
cout << pRect->length;//using arrow operator


Phewwwwww......

indigo0086

Yes!