Why does some methods work while some do not on null values of nullable structs?

Question

Straight to the point:

int? i = null;
i.ToString(); //happy
i.GetType(); //not happy

I get a very related question which actually deals on why does i.ToString() work fine.

Edit: Just found out this corner case has been the most voted one in this SO thread!

Answer 1

It is because ToString is virtual while GetType is not. Nullables have special boxing behavior in the CLR. When you call GetType on a nullable it is boxed (see MSDN, actually this happens for any struct). However, with a nullable, the the underlying value is boxed, not the actual nullable. ToString on the other hand calls the override of ToString method on Nullable<T>. Also, I would note that int? i = null is syntax sugar for Nullable<int> i = new Nullable<int>(). So, there really is an object in your variable i.

Answer 2

Why does i.ToString() not fail?

Use any .NET decompiler and look at the Nullable<T> class. You'll see that ToString() is overriden as follows:

public override string ToString()
{
  if (!this.HasValue)
    return "";
  else
    return this.value.ToString();
}

---

Why does i.GetType() fail?

See mike z's answer.