Puzzled C++ code

[yqwen]

please run the following code, and kindly advise why the result is 20 instead of 10? thanks in advance

#include <iostream>
#include <string>
#include <vector>
 
using namespace std;
 
class A
{
public:
    A() : i(1) {} 
    int i;
};
 
class B : public A
{
public:
    B() : j(3) {} 
    int j; 
};
 
int f(A* p, int count)
{
    int total=0;
 
    for(int k=0; k<count; k++)
    {
        int temp = p->i;
        p++;
        cout << " k =  " << k << ", " << temp << endl; 
        total += temp;
    }
 
    return (total);
}
 
int main() {
    B b[10];
    cout << f(b,10);
 
    return 0;
}
 

[r035198x]

please run the following code, and kindly advise why the result is 20 instead of 10? thanks in advance

#include <iostream>
#include <string>
#include <vector>

using namespace std;

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

class B : public A
{
public:
B() : j(3) {}
int j;
};

int f(A* p, int count)
{
int total=0;

for(int k=0; k<count; k++)
{
int temp = p->i;
p++;
cout << " k = " << k << ", " << temp << endl;
total += temp;
}

return (total);
}

int main() {
B b[10];
cout << f(b,10);

return 0;
}

1.) Please use code tags when posting code
2.)Why don't you tell us why you expect 10 instead of 20 first?

[LaneWolf]

please run the following code, and kindly advise why the result is 20 instead of 10? thanks in advance

#include <iostream>
#include <string>
#include <vector>

using namespace std;

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

class B : public A
{
public:
B() : j(3) {}
int j;
};

int f(A* p, int count)
{
int total=0;

for(int k=0; k<count; k++)
{
int temp = p->i;
p++;
cout << " k = " << k << ", " << temp << endl;
total += temp;
}

return (total);
}

int main() {
B b[10];
cout << f(b,10);

return 0;
}

Try print out sizeoff(A) and sizeof(b) to see the problem ;)

[yqwen]

1.) Please use code tags when posting code
2.)Why don't you tell us why you expect 10 instead of 20 first?

1. Sorry, I am new here, didn't know the norm and everything.
2. The reason I am expecting 10 is because I thought (still think) that for the array of b, data member i should equal 1 for every element, so the summation of such should turn out to be 10. I know something is wrong but I cannot figure out what that is.

thank you very much.

[yqwen]

Try print out sizeoff(A) and sizeof(b) to see the problem ;)

I am sorry but are you trying to help? I hope you can understand that I am pretty puzzled by this problem.

[Banfa]

I am sorry but are you trying to help? I hope you can understand that I am pretty puzzled by this problem.

I can understand that because LaneWolf's clue is rather oblique however it is at the root of the problem.

Examine these 2 lines of your code

    p++;
    cout << " k = " << k << ", " << temp << endl; 
 

The first is the line that is manifesting the error (that is causing the error to be apparent, the actual error is a little more fundamental), can you explain what you thing it is doing? Do you understand how pointer arithmetic works in C++?

The second line is outputting data that should give you a clue as to what is going wrong.

[yqwen]

Thanks.

1. This is a question from an online C++ aptitude test.
2. I understand that array should not be used in a polymorphism setting.
3. I just wanted to know why it still produces a seemingly OK result, which happens to be the correct choice in the test. (I mean no matter how absurd it is, it still follow some logic to produce 20, so I wanted to know why it's 20)

I can understand that because LaneWolf's clue is rather oblique however it is at the root of the problem.

Examine these 2 lines of your code

    p++;
    cout << " k = " << k << ", " << temp << endl; 
 

The first is the line that is manifesting the error (that is causing the error to be apparent, the actual error is a little more fundamental), can you explain what you thing it is doing? Do you understand how pointer arithmetic works in C++?

The second line is outputting data that should give you a clue as to what is going wrong.

[Ganon11]

Each time the statement p++ is executed, p is set to point to some spot in front of the previous spot. The difference between the old and new address is determined by the size of the object type p points to. Since the function tells us p points to an object of type A, it assumes the only data member included is i (an integer), and increments p by the proper amount (1 byte, the size of an integer).

p really points to an object of type B, which has 2 integers; thus taking up twice the room of A. So when p is incremented 'to a new A object', it is really pointing to the B portion of that object - that is, the memory which the compiler assumes to be an A object is actually a B object. Since both classes have only one integer data member, you are able to access j as if it were an i in an A object.

Thus, you are actually adding both i and j from the first five B objects in your array with this function.

Some of my explanation is probably incorrect regarding the specifics of how each object is stored in memory, in which case I'm sure Banfa will be glad to help out. However, this is my analysis of the output you are getting.

[Banfa]

Some of my explanation is probably incorrect regarding the specifics of how each object is stored in memory, in which case I'm sure Banfa will be glad to help out. However, this is my analysis of the output you are getting.

No need, you have it pretty much spot on.

[yqwen]

Ganon11:

Thank you very much for the clarafiction and I admire your patience and time for putting up the words. The concept is now crystal clear to me.

-yqwen

Each time the statement p++ is executed, p is set to point to some spot in front of the previous spot. The difference between the old and new address is determined by the size of the object type p points to. Since the function tells us p points to an object of type A, it assumes the only data member included is i (an integer), and increments p by the proper amount (1 byte, the size of an integer).

p really points to an object of type B, which has 2 integers; thus taking up twice the room of A. So when p is incremented 'to a new A object', it is really pointing to the B portion of that object - that is, the memory which the compiler assumes to be an A object is actually a B object. Since both classes have only one integer data member, you are able to access j as if it were an i in an A object.

Thus, you are actually adding both i and j from the first five B objects in your array with this function.

Some of my explanation is probably incorrect regarding the specifics of how each object is stored in memory, in which case I'm sure Banfa will be glad to help out. However, this is my analysis of the output you are getting.

[Atli]

... (1 byte, the size of an integer) ...

Hi. This is probbly a stupid question but...
An integer is 1 byte in C++? Dont bytes have a max value of 255?

[Banfa]

An integer is 1 byte in C++?

No at least 2 bytes actually I assume that's a typo on Ganons part.

Dont bytes have a max value of 255?

No it depends on how many bits are in the byte (which is not fixed by the standard).

[Atli]

Ahh ok I see, thats a stupid thing not to have a fixed standard for.

[Banfa]

Ahh ok I see, thats a stupid thing not to have a fixed standard for.

No it's not, if your platform is 7 bit or 9 bit having your programming language fixed to 8 bit bytes would be a stupid thing to do (oh and 7 and 9 bit platforms do exist).

You are falling into the trap (and TBH so do most people) of assuming that all platforms share the same characteristics as the platforms(s) you use.

[Ganon11]

No at least 2 bytes actually I assume that's a typo on Ganons part.

Yeah, dont know what I was thinking when I typed that. Thanks for the catch on that one.

[Atli]

You are falling into the trap (and TBH so do most people) of assuming that all platforms share the same characteristics as the platforms(s) you use.

Yea thats true, we (humas in general) have a habit of assuming that what we dont know doesn't exist.

No it's not, if your platform is 7 bit or 9 bit having your programming language fixed to 8 bit bytes would be a stupid thing to do (oh and 7 and 9 bit platforms do exist).

What I dont understand is why have 7 and 9 bit computer systems. I've always thought computer systems were built arround 2 in the power of x (or n)... how does 7 or 9 fit into that pattern?

[JosAH]

What I dont understand is why have 7 and 9 bit computer systems. I've always thought computer systems were built arround 2 in the power of x (or n)... how does 7 or 9 fit into that pattern?

They don't and why should they? PDP/8 -> 12 bit words, PDP/9 -> 9 bit bytes,
IBM/370 -> 36 bit words but variable, Cray -> ditto, Saturn -> 4 bit bytes. The
extra bits can come in handy you know; e.g. the first LISP implementations
couldn't live without them ;-)

kind regards,

Jos

[Atli]

Good point. Never acctually had to use any of this so I've forgotten most of it.
Also might have helped if I had been paying attention when I was learning this stuff :P

[Banfa]

What I dont understand is why have 7 and 9 bit computer systems. I've always thought computer systems were built around 2 in the power of x (or n)... how does 7 or 9 fit into that pattern?

7, 9, 12 and anything else you care to mention fits the pattern because they are the x value.

The power of 2 thing comes into maximum integer values and bit values, for instance on an 8 bit machine the bits have the values

1 2 4 8 16 32 64 128 maximum unsigned char = 256

all powers of 2, the equivalent values for 7 and 9 bits are

7: 1 2 4 8 16 32 64 maximum unsigned char = 128
9: 1 2 4 8 16 32 64 128 256 maximum unsigned char = 512

x can be anything you want, but everything else will be a power of 2 due to the binary nature of computers. (actually even this is not quite true).

[Atli]

Ahh I see, its all coming back to me now :P

Thanks.