I believe that when you add two unsigned int values together, the returned value's data type will be an unsigned int.

But the addition of two unsigned int values may return a value that is larger than an unsigned int.

So why does unsigned int + unsigned int return an unsigned int and not some other larger data type?

4 Answers

Bathsheba On

This would have truly evil consequences:

Would you really want 1 + 1 to be a long type? And (1 + 1) + (1 + 1) would become a long long type? It would wreak havoc with the type system.

It's also possible, for example, that short, int, long, and long long are all the same size, and similarly for the unsigned versions.

So the implicit type conversion rules as they stand are probably the best solution.

You could always force the issue with something like

0UL + "unsigned int" + "unsigned int"
eerorika On

Let's imagine that we have a language where adding two integers results in a bigger type. So, adding two 32 bit numbers results in a 64 bit number. What would happen in expression the following expression?

auto x = a + b + c + d + e + f + g;

a + b is 64 bits. a + b + c is 128 bits. a + b + c + d is 256 bits... This becomes unmanageable very fast. Most processors don't support operations with so wide operands.

user463035818 On

The type of a varaible does not only determine the range of values it can hold, but sloppy speaking, also how the operations are realized. If you add two unsigned values you get an unsigned result. If you want a different type as result (eg long unsigned) you could cast:

unsigned x = 42;
unsigned y = 42;
long unsigned z = static_cast<long unsigned>(x) + static_cast<long unsigned>(y);

Actually the real reason is: It is defined like that. In particular unsigned overflow is well defined in C++ to wrap around and using a wider type for the result of unsigned operators would break that behaviour.

As a contrived example, consider this loop:

for (unsigned i = i0; i != 0; ++i) {}

Note the condition! Lets assume i0 > 0, then it can only ever be false when incrementing the maximum value of unsigned results in 0. This code is obfuscated and should probably make you raise an eyebrow or two in a code-review, though it is perfectly legal. Making the result type adjust depending on the value of the result, or choosing the result type such that overflow cannot happen would break this behaviour.

Community On

Because a variable + a same type variable can be only equal to that type variable , (well in some cases it will but not in your case)


int + int = int a int plus another int cannot be equal to a float because it dont have the properties of a float. I hope this answers your question bye!