Explain what can go wrong with the following C code

[luckyluke]

Please help me to explain what can go wrong with the following C code, why it looping forever? I don't understand.

#include "stdafx.h"
#include <stdio.h>

int main()
{
	int x; 
	float y; 
	x = 1; 
	while (x > 0) 
	{ 
		y = x+1; 
		x = y; 
	}
	return 0;
}

Regards.

[Elysia]

Pretty much floating point inaccuracies: What Every Computer Scientist Should Know About Floating-Point Arithmetic

[luckyluke]

Pretty much floating point inaccuracies: What Every Computer Scientist Should Know About Floating-Point Arithmetic

Oh, Eysia. I'm a newbie, so I have learnt C for 2 week in this summer-semester. I have read your recommended document so I don't understand clearly. Can you detail my problem, many thanks to you!

[Elysia]

In short: there isn't enough storage in a float to exactly store all kinds of numbers. Remember that a float has to store decimals as well as the integer part.
Eventually, you will stumble upon a number that cannot be exactly represented in a float. It happens that two specific numbers are stored in the same way because it cannot be represented exactly (again, lack of enough storage).
At some point, x and x + 1 are the same number in a float. Then you hit your ceiling and cannot proceed.

[sfryett]

	while (x > 0)

x is always > 0?

[Elysia]

It will eventually overflow, resulting in a negative number.

[Adak]

Please help me to explain what can go wrong with the following C code, why it looping forever? I don't understand.

#include "stdafx.h"
#include <stdio.h>

int main()
{
	int x; 
	float y; 
	x = 1; 
	while (x > 0) 
	{ 
		y = x+1; 
		x = y; 
	}
	return 0;
}

Regards.

The while loop will never break out, because you set x equal to 1, then you increase x by 1, in every iteration of the loop (using y as an intermediary).



So how is x ever going to be anything else but greater than zero?

[Elysia]

Of course, it's also a question if a float can overflow?

[luckyluke]

In short: there isn't enough storage in a float to exactly store all kinds of numbers. Remember that a float has to store decimals as well as the integer part.
Eventually, you will stumble upon a number that cannot be exactly represented in a float. It happens that two specific numbers are stored in the same way because it cannot be represented exactly (again, lack of enough storage).
At some point, x and x + 1 are the same number in a float. Then you hit your ceiling and cannot proceed.

I have just understood, many thanks to you!