postfix

Hi,

Could someone please explain how the following ends up outputting 0?

Code:

---

#include <iostream>
int main()
{
        int a=0;
        std::cout << a++ + a++ << std::endl;
        return (0);
}

---

I interpret this as follows.
Compiler reads from right to left. So first the a++ on the right gets processed. Since its postfix, operator + gets 0. and then a is set to 1. Now the a++ on the left gets processed. a=1 is passed to +, and then a is incremented by 1 to 2. So the final result should be 1.

cheers,

pooya

Just.....don't do that.

Honestly I get suspicious enough of code that uses ++ inside a larger expression. Using it twice is just asking for trouble.....

What's probably happening is that the compiler's logic for argument passing is separate from its logic for "all expressions with side-effects must be evaluated". As far as the function-call goes, it sees the same expression twice, says "Oh, I already know what that is", and just passes it. Sometime later it actually does the second increment.

a + a is output, then a is incremented twice. If you want to test it, try

Code:

---

#include <iostream>

int main( )
{
  int x = 1;
  int y = 2;

  std::cout << x++ + y++; // output 3
  return 0;
}

---

And as Lindley said, that type of thing is pure evil. In no small part due to confusion like yours.

Yeah, I have to agree, I wouldn't do that.

For one, the a++ type expressions are generally used in for loops, for things like reading array values, etc..

Code:

---

for ( int i=0; i < 100; i++ )
{
std::cout << "Print me 100 times" << std::endl;
}

---

(Thats not even getting into the problems that can be created by using two of them on the same variable, in one operation.)

Second, cout is for output, it's for printing results, etc, you should do your operations beforehand, and then print the results using cout.

Code:

---

int a=1;
int b=1;
int c=0;

c= a+b;

cout << c << endl;

---

Quote:

---

Originally Posted by pooya

Compiler reads from right to left.

---

Your assumption is wrong. In fact in C++ the order of evaluation of subexpressions within an expression is unspecified (except for a few operators that enforce a specific order).

Quote:

---

Originally Posted by pooya

So first the a++ on the right gets processed. Since its postfix, operator + gets 0. and then a is set to 1. Now the a++ on the left gets processed. a=1 is passed to +, and then a is incremented by 1 to 2. So the final result should be 1.

---

Because the order is unspecified any postfix subexpression may be evaluated first but it doesn't matter which because the example is symmetric.

You have

a = 0;
b = a++ + a++;

When one of a++ is evaluated it's evaluated to 0. Because a is 0 when the other a++ is evaluated a stays 0. The two a++ evaluations have resulted in a being 0 so when then a+a is evaluated it becomes 0+0 = 0, which is assigned to b.

Say you have this instead,

a = 0;
b = ++a + ++a;

When one of ++a is evaluated it becomes 1. Because a is 1 when the other ++a is evaluated a becomes 2. The two ++a evaluations have resulted in a being 2 so when then a+a is evaluated it becomes 2+2 = 4, which is assigned to b.

The above may seem strange but that's what the MS compiler does. I'm sure other compilers could come up with other results. The reason for this is that the evaluation order is unspecified. The MS compiler's strategy seems to be to first perform all increments on a and then use the resulting a value in the sum, something it's entitled to do according to C++.

you should know that the difference between the a++ and ++a.

if a=0, then a++ equals 0, but ++a equals 1.

So you should know what your problem is.





-----------------------------------------------------
codeuu,source code

_uj's explanation is correct. If you are not convinced, then read Stroustrup's answer to the FAQ What's the value of i++ + i++?