Can someone please explain the output of this code?

[Aarti]

Hi,

Can some one please explain why the output of this program is 15

#include <iostream>

using namespace std;

class A
{
public:
A():i(1)
{}
int i;
};

class B: public A
{
public:
B():j(2)
{}
int j;
};

int f(A* p, int count)
{
int total = 0;
for(int i=0; i<count; ++i)
{
total+=p++->i;
}
return(total);

}

int main()
{
B b[11];
cout << f(b,10);
return 0;
}

Regards,

[Thomas J. Gritzan]

Aarti wrote:

Hi,

Can some one please explain why the output of this program is 15

[...]

class A

[...]

class B: public A

[...]

int f(A* p, int count)
{
int total = 0;
for(int i=0; i<count; ++i)
{
total+=p++->i;
}
return(total);

}

int main()
{
B b[11];
cout << f(b,10);
return 0;
}

You invoked undefined behaviour by treating an array of Bs as array of As.

See Item 3 "Never treat arrays polymorphically" in Scott Meyers's "More
Effective C++"

--
Thomas
http://www.netmeister.org/news/learn2quote.html

[Daniel T.]

Aarti <aa*********@gmail.comwrote:

Can some one please explain why the output of this program is 15

#include <iostream>

using namespace std;

class A
{
public:
A():i(1)
{}
int i;
};

class B: public A
{
public:
B():j(2)
{}
int j;
};

Note, sizeof(B) sizeof(A). For the sake of this explanation, let's
assume that sizeof(B) is 8 and sizeof(A) is 4.

int f(A* p, int count)
{
int total = 0;
for(int i=0; i<count; ++i)
{
total+=p++->i;

The increment above moves 'p' by sizeof(A) bytes. i.e., 4 bytes, but
since the object pointed to by 'p' is really a B, 'p' now points to the
middle of a B object.

}
return(total);

}

int main()
{
B b[11];

In the above line, you create an array of B. If sizeof(B) is 8, and b[0]
is at address location 0x0, then b[1] is at address location 0x8.

cout << f(b,10);
return 0;
}

Regards,

[joe]

You have exercised a big no-no. When you increment a pointer, what
you are effectively doing is adding the sizeof the pointee type to the
value of the pointer. This allows you to visit the next element of an
array automatically. What you have done here is passed an array of B-
s (which are larger than A-s) into your function and trying to access
it as though it were A-s. So, what happens is that you take your
pointer and add sizeof A to it. This means that your pointer now
points into the middle of your B object somewhere. It just so happens
that since this class is simple and nothing faults, but you are now
treating the middle of the B as another A. This is bad. It turns out
that in your example you effectivly access b[0].i + b[0].j + b[1].i +
b[1].j .... and so forth. This gives 5 * 1 + 5 * 2 = 15. Normally, I
might expect a fault if things were more complicated than simple ints
and it is definitely undefined behavior.

joe