Can I make this function better?

[btripp]

Hey,

My homework was to make a function that added two integers. The function was to throw an exception if the sum of the two integers were larger than INT_MAX. I got it working perfectly. Ive tried many test cases with much success. Now I just want to maybe make it more concise I guess. I just feel like I have excess if statements. Here is my function

int add(const int& a, const int& b)throw(char*)
{
	int total;
	if(a == INT_MAX && b == INT_MAX)
		throw("Sum larger than MAX_INT");
	else if(a == INT_MAX && b > 0)
		throw("Sum larger than MAX_INT");
	else if(b == INT_MAX && a > 0)
		throw("Sum larger than MAX_INT");
	else if(a < INT_MAX && b > (INT_MAX - a))
	{
		if(a < 0)
			total = a + b;
		else
			throw("Sum larger than MAX_INT");
	}
	else if(b < INT_MAX && a > (INT_MAX - b))
	{
		if(b < 0)
			total = a + b;
		else
			throw("Sum larger than MAX_INT");
	}
	else 
		total = a + b;

	return total;
}

If you need me to explain anything let me know.

Thanks,
Brandon

[firedraco]

Since you are using just ints, you could sort of cheat and do this:

long int total = a +b; //a long int can store much bigger numbers than a normal int
if(total > INT_MAX) throw("Sum larger than INT_MAX");
return (int)total;

[MaikCAE]

Hi, I think this part should be enough:

int add(const int& a, const int& b)throw(char*)
{
	int total;
	if(b <= (INT_MAX - a))
	{
		total = a + b;
	}
	else 
		throw("Sum larger than MAX_INT");

	return total;
}

because you want that

a+b <= MAX

Substracting a on boith sides:

b <= MAX - a

If this case is not matching, the sum between a and b is larger than MAX and you can throw the exception

[herbert1910]

i think line 4 (a == INT_MAX && b == INT_MAX) is already covered by the other lines. you can probably get rid of it.


on my compiler, there is an overflow with negative numbers which doesn't affect the outcome.

when

a = (INT_MAX - 1)
b = -3

we get to line 17:
(INT_MAX - b) = INT_MAX - ( -3)) = INT_MAX + 3.

a > (INT_MAX - b) should be false when we put in values.
(INT_MAX - 1) > (INT_MAX + 3) should be false.

INT_MAX + 3 becomes -overflownumber!

a > -overflownumber is true but the value is wrong.

to avoid this, make sure b >= 0 so INT_MAX - b is never > INT_MAX.

or you can avoid it by not using my compiler. ha ha.

[btripp]

herbert I see what you are saying. I tried that test case so I had to add the if b<0 code in line 19. Because if b is negative and we are only working with ints that means that a can be at the greatest INT_MAX, so go ahead and add the two numbers if and only if b is negative.

Thanks all for the Input.

Firedraco that is super smart. I dont know if my teacher would like that though. I think we were supposed to only use ints but im not sure. I see exactly how yours would work though.

Here is the actual assignment though if you guys were interested.

Your new job as a Computational Physicist involves calculations using very large integers. However, some of the results are incorrect. For example, when adding very large positive integers, sometimes the results become a negative number. After consulting with a Computer Science professor (for a large fee :) ), you have determined in some cases the results are too large to fit into an integer. Write a function that adds two integers and returns the results. If the result is too large to fit into INT_MAX, throw an error message. Test the function in the main using a try and catch.

Hint: The results of an addition can be negative.

[btripp]

@MaikCae

I just tryed your code. At first I was thinking this way and I ran into problems. It seems to work sometimes but not for all cases. For example using your code. If you entered
add(-100,-100) it would throw an exception
alslo if you added
add(INT_MAX,-5) it would throw an exception

both of the above cases should work. Because -100 + (-100) is -200 and that should fit into INT_MAX

and INT_MAX + (-5) should fit because it is simply INT_MAX - 5 < INT_MAX

I think your code works if you are not using negative numbers.

Thats why I had to pick off each case in my code because it is a harder problem than it seems.