C program:output seems confusing

[Arjun Srivastava]

#include<stdio.h>
#include<conio.h>
#define PROD(x) (x*x)
void main()
{
clrscr();
int p=3,k;
k=PROD(p+1); //here i think value 3+1=4 would be passed to macro
printf("\n%d",k);
getch();
}

Hi,in my opinion,out should be 16. but OUTPUT: 7 Anyone please tell me how? Thanks.

[Ernest Friedman-Hill]

An important C programing lesson: macros aren't functions! They are just expanded as text, and then the text is compiled as part of the normal code. This macro expands to the text "x*x" for a given value of "x"; here "x" is "p+1", so we get "p+1*p+1", which (just as in Java) evaluates to p + (1 * p) + 1 = 7 (since p is 3.)

[Arjun Srivastava]

Ernest Friedman-Hill wrote: here "x" is "p+1", so we get "p+1*p+1",evaluates to p + (1 * p) + 1 = 7

Yeah,Nice.
Thanks. That was really simple.
idk where my mind is these days.

[Campbell Ritchie]

Change PROD(p+1) to PROD((p+1)).

PROD is an inappropriate name for your macro; you should have used SQUARE or similar.

[Campbell Ritchie]

. . . or better still, #define SQUARE(x) ((x)*(x))

[Ernest Friedman-Hill]

Campbell Ritchie wrote: . . . or better still, #define SQUARE(x) ((x)*(x))

That's the classic dodge, although then the follow-on question is "Why does this give 12 rather than 9?"

int x = 2;
int result = SQUARE(++x);

[Arjun Srivastava]

Ernest Friedman-Hill wrote:

Campbell Ritchie wrote: . . . or better still, #define SQUARE(x) ((x)*(x))

That's the classic dodge, although then the follow-on question is "Why does this give 12 rather than 9?"

int x = 2;
int result = SQUARE(++x);

Ok if you seriously want to know how 12 comes,i will tell you

((++x)*(++x))=((3)*(4))=12

Thanks for the guidance.

[Campbell Ritchie]

Ernest Friedman-Hill wrote:

Campbell Ritchie wrote: . . . or better still, #define SQUARE(x) ((x)*(x))

That's the classic dodge, although then the follow-on question is "Why does this give 12 rather than 9?"

int x = 2;
int result = SQUARE(++x);

What about SQUARE(x++)?

[Arjun Srivastava]

Campbell Ritchie wrote:What about SQUARE(x++)?

First of all,please be serious and stop laughing,we are having serious discussion here.
Above one is still tricky, but never mind here we go

(2)*(3)=6.

[Campbell Ritchie]

Arjun Srivastava wrote:. . . serious discussion here. . . .

You jest


I challenge you to find a C implementation which returns 12 from SQUARE(x++). It shouldn't take you more than an hour.

[Arjun Srivastava]

Campbell Ritchie wrote:You jest
find a C implementation which returns 12 from SQUARE(x++).

Phew. You are G(enius).
I have tried more than an hour ,but didn't able to crack that joke oops sorry that equation.
What is the answer?please tell me.
Answer again confusing to below code

#include<stdio.h>
#include<conio.h>
#define SQUARE(x) (x*x)
void main()
{
clrscr();
int i=3,j,k;
j=SQUARE(i++);
k=SQUARE(++i);
printf("\n%d\n%d\n%d",j,k,i);
getch();
}

Answer again confusing: 9 49 7
I was thinking j=3*4=12, k=6*7=42 , i=7
What is happening?did i miss something?
(x*x)=((x)*(x)) same here.it doesn't matter.

[Campbell Ritchie]

You need to try different implementations, as I said. The behaviour of i++ is undefined in C, and if you try enough different C compilers, you will find some which will return 6 from that code, and some which will return 12.

SQUARE(x++) will expand to (x++ * x++) You are obviously getting 3 returned the first time, and squaring that to 9. By the end of the first line x is 5, and by the end of the second line it is 7. So you are obviously the final value of ++x twice, and the initial value of x++ twice in the first line. That is a different behaviour from what I expected. I tried your code on "gcc" and got 9 49 7 as output.

[Arjun Srivastava]

It's undefined because the standard says so. Put it through a decent compiler like gcc - Wall -std=c99 to get warning: operation on 'i' may be undefined. The fact that it compiles on a particular implementation (the ancient Turbo C by the looks of it) is not relevant. You can do it, just be aware that the results may not always be what you expect.

Campbell Ritchie wrote:I tried your code on "gcc" and got 9 49 7 as output.

Hi sir, did you also get any warning while compiling this code on gcc?

#include<stdio.h>
#include<conio.h>
#define SQUARE(x) (x)*(x)  // line1
void main()
{
clrscr();
int i=3,j,k,l,m;
j=SQUARE(i++);         // i=3,4
k=SQUARE(++i);        //  i=6,7
l=SQUARE(++i);       //   i=8,9
m=SQUARE(i++);      //    i=9,10
printf("\n%d\n%d\n%d\n%d",j,k,l,m);
getch();
}

j=9
k=49
l=81
m=81
I saw the pattern here.
for post increment i++ :it gives square of i(first value)
For pre increment ++i :it gives square of i(second value)
Also if line 1 changes to ((x)+(x))
j=6
k=14
l=18
m=18
It follows the same pattern.
Does this make any sense? can i use this technique for solving pseudo codes output?
Thanks

[Campbell Ritchie]

Arjun Srivastava wrote: . . . Hi sir, did you also get any warning while compiling this code on gcc?

No, but I didn't use -Wall

. . . It follows the same pattern.
Does this make any sense?

Yes

can i use this technique for solving pseudo codes output?

Probably not.

[Arjun Srivastava]

Campbell Ritchie wrote:
I challenge you to find a C implementation which returns 12 from SQUARE(x++).

Btw what was that implementation?
I didn't get through.
Please declare the answer.

[Campbell Ritchie]

I can't remember. Sorry. I used to have three C/C++ compilers running on a Windows box; one was Dev-C++, which is a nice tool which returns 4 from

int i = 4;
i = i++;

. . . so it wasn't that. (If you want a C compiler for Windows, I think Dev-C++ is a good bet).

There are lots of hits on this Google search for "c compiler for windows"; there was one maybe Miracle-C which returned 5 from the above code snippet.
Sorry, I can't remember any more :o

[Arjun Srivastava]

Yeah, now am using Dev-C++ compiler, it is much more better than old turbo C and friendly too.
Thanks.

[Campbell Ritchie]

You're welcome. Dev-C++ is a nice application, isn't it.