Thread: What is the diffence with these functions

have a queston that is confusing me....ok say we want to pass a call to a function by refence or a pointer....like this

Code:

void squarebyRefernce( int &numberRef )
{
nuberRef *= numberRef;
}

//or this

void squraebyReference( int *nPtr )
{
*nPtr = *nPtr * *nPtr;
}

My question is...what is the difference... and why would you chose one over the other?

From what i gather they both do the same thing.

in squarebyRefernce( int &numberRef ) you're just passing a variable by reference; everything that will be done to numberRef in the scope of that function, will also be passed back to main.

in the second function, you are simply passing in a pointer to an int...whatever happens to that pointer withing the scope of that function only happens there.

so, the code below should give you an error about passing an int where a pointer is needed.

Code:

#include <iostream>
using namespace std;


void squarebyRefernceA( int &numberRef )
{
	numberRef *= numberRef;
	cout << "A: " << numberRef << endl;
}



void squraebyReferenceB( int *nPtr )
{
	*nPtr = *nPtr * *nPtr;
	cout << "B: " << nPtr << endl;
}

int main(){

	int x = 4;
	squarebyRefernceA( x );
	squraebyReferenceB( x );



	return 0;
}

Im not sure I follow you....They both change the value of the caller in main right ?

What he is saying is that with the referance one you are actully changing the value of x with the pointer one your are just changing the value of the pointer to x

If you forget to clearly name the function ambiguity can arise
consider
func(variable) where func(int &var); it is unclear that variable is being modified by the caller
func(&varibale) where func(int *var) clarifies that the calling function is changing the variable

Strostrup discusses this on pg 99 of the C++ Programing Language as well. He also states that in many cases it is best to pass by value and return a value instead.

consider this
int x = 2;
y = squarebyRefrence(x);//y=4
z = y+x;// opps y= 8 when you wanted 6

or this poly
y=squarebyRefrence(x)+x;

One other point, which was almost touched upon by bithub is that you are guarinteed that the first of the two will work, where as the second may crash your program:

Code:

void squarebyRefernce( int &numberRef )
{
    nuberRef *= numberRef;
}

//or this

void squraebyReference( int *nPtr )
{
    *nPtr = *nPtr * *nPtr;
}

In the first example, numberRef will always be referring to some valid item. But, in the second, nPtr could very well be NULL, and as such, your program may do BadThings(TM) when you dereference it.

With references, you are guarinteed to be referencing something when that point in the program is reached.

Quzah.

Thanks for the reply's folks!That shed's some lite on it for me.So it basicly boils down to preference ? by using &var intstead of *var(which both basicly do the same thing) I am tying &var to an object and know that it is initailized with that object. Where using *var does'nt and could lead to a dangling pointer.