Divide by zero problem

[Hunt_guy]

her's two piece of Code
1.
{
int a=7,b=0,c;
c=a/b;
}
this code give's a error at run time divide by zero
2.
int c;
c=7/0;
printf("%d",c);
this code gives a compile time warning divide by zero
but at run-time No Warning :(
also c containc 7
can someone explain me wht the hell is happing here
thanx in advance

[the_JinX]

I think the compiler takes the error out (in the second code)..

the compiler "knows" you can't devide by zero and just removes the devision..

the compiler isn't "smart" enough to do the same with your first peice of code..

[steve.milner]

Originally posted here (http://www.AntiOnline.com/showthread.php?threadid=253475#post710687) by the_JinX
the compiler isn't "smart" enough to do the same with your first peice of code..

It will be a long time before that are ever smart enough...

How would a compiler deal with this - Forgive me this isnt in a prog. language.

Set Vairable to be 10 divided by a random integer between -5 and 5


Steve

[Algaen]

I would think the compiler wouldn't do a thing about it. Maybe an very well written compiler could spit out a warning that there is a potential divide by 0 error, but the ones in use now would let it go. In many languages there is no way that the random number could be known at compile since their random number generators are seeded from the system clock. (therefore the number generated depends on the time the program was run at).

If the random number generated at run-time turned out to be 0 then the program would crash with a run-time error (or throw an exception depending on the language).

[Tim_axe]

Question interperted as:

int c;
c=7/0;
printf("%d",c);

Why does it print 7, and not some uninitilized value? Why does it get c=7 when the only time a value is given is c=7/0, which isn't really do-able for the program?


My best guess is that the compiler's *.o or something file that goes with your *.c file removed the /0 part, so it turned into c=7 instead of c=7/0. I don't quite know how to check that, but it probably happens in an intermediate step of compiling where macros, #defines, etc., are replaced. The only way to find out would be to do some digging through those files if you can find them before they are deleted, or to read up on different things about your specific compiler. I haven't tried it so I don't know how DevC++ handles it, but it might be the same way that MSVC++ and/or Borland or anything else does it... Maybe it is part of the C language standard? I don't know for sure, so you'd probably have to look it up...


And I wonder what would a "smart enough" compiler would do when have have a part in your own random function used in "int error=5/neverzero();" that is like "decision = true; if (decision != true) { return 0; } " and then at some point in time there is acturally a way to set decision to be false overriding the piece of code before it... *confusing*

*edited

[White_Eskimo]

If the random number generated at run-time turned out to be 0 then the program would crash with a run-time error (or throw an exception depending on the language).

This is why humans make code and not machines...humans are capable of throwing in an extra line of code just to check to make sure that you arent deviding by zero.


Set Vairable to be 10 divided by a random integer between -5 and 5
If random integer is equal too 0
Display error message


Thats why there are if statements guys :) so we can just to make sure the user isnt entering 0 when deviding.

Maybe an very well written compiler could spit out a warning that there is a potential divide by 0 error, but the ones in use now would let it go.

If that were the case, i would not want to own a well written compiler :) Look at this bit of code (yes i know it is in C++...)

int main (void)
{
int x;
int y=4;
int z=0;
cin >> x;
z=y/x;
return 0;
}


It would be annoying if you got warning messages because the dumb user could enter 0 for x, and therefore you would be trying to devide by zero which isnt possible. instead, compilers allow the programmer to check, so a better version of the code above (still a very lame version) would be:

int main (void)
{
int x;
int y=4;
int z=0;
cin >> x;
if (x==0)
{
cout << "no you dumb**** you cant do that";
return 0;
}
z=y/x;
return 0;
}


I think a nice friendly error message is a lot better than a bunch of lame compiler warnings :)

[Hunt_guy]

Hello guys i have done some digging in this problem and got a close solution the divide by zero error comes only when the 00h interrupt is generated or is trigrred by some external device or by thr programmer by using INT 00H assembly instruction but this happens only ate execution time of the program wht is happing here is that we r givingvariable names ie a/b where a=7,b=0 and due to this at the run time the processoer fetches the data from these address and then does the division and 00H interrupt is generated but in case of immidiate operands ie 7/0 the divide by zero condition is generated at the compile time so no question of a intrrupt and therefore no divide by zeror error only warning don't know how come this a=7/0, here 'a' got the value 7 ??

[gothic_type]

White_Eskimo, you're almost there, but not quite. You're being too nice to the user. A much more user friendly version would clear the screen and then put up exactly the same prompt that was already there to make the user think that the program's crashed :P

Please enter a number: 0
[screen clears]
Please enter a number: 0
[screen clears]
Please enter a number: 0
[user punches screen]
["Please enter a number" starts scrolling down the screen]
[user shoots self]


Now that's a friendly program :P

#include <iostream>
#include <stdlib>
#include <dos> // doubt this is needed

void main(void)
{
int number=0;

std::cout << "Please enter a number: ";
std::cin >> number;

while(number==0)
{
system("cls");
std::cout << "Please enter a number: ";
std::cin >> number;
}
}


That's how you find out if someone is worthy to use your program or not :)

ac

[White_Eskimo]

lol hahaha :) amen