Question

When I compile the following code, I only see the error during Run-time which says
“Unable to cast object of type ‘Foo1’ to type ‘Foo2′”

Why is the compiler not showing this error during compile time?

public void Start()
{
     Foo1 objFoo1 = new Foo1();
     Foo2 objFoo2 = (Foo2)objFoo1;

     //objFoo1.FooA = 10;
     //Console.WriteLine(objFoo2.FooA);

}

public class Foo1    {}
public class Foo2 : Foo1    {}

Answer 1

The compiler is not smart enough to know that objFoo1 is really a simple Foo1. It doesn’t analyze your source code deeply enough to determine that. All it sees is that objFoo1 is a Foo1, Foo2 is derived from Foo1, and therefore the cast is legal. If instead you changed that to (int) objFoo1 it would bomb at compile-time since the compiler could see that there’s no possible way to turn a Foo1 into an integer.

Imagine that there were 1000 lines of code between the two declarations, many of them doing various assignments to objFoo1. The compiler would have to do a lot of hard work to attempt to determine what object exactly is in objFoo1. In general it wouldn’t always be able to do so. It’s far easier for the compiler to just take the static type information at face value and assume that objFoo1 is some type of Foo1 object but no more. That way it doesn’t reject your simple test program but accept a more complicated one.

It’s also not the compiler’s job to reject your code, in this case. It is possible, though admittedly far-fetched, that you are intentionally trying to generate a ClassCastException by doing an illegal cast. It’d be unusual but not illegal to do that.

Along these lines, there are other cases where the compiler’s static code analysis can be fooled. This will fail to compile:

return;
System.out.println("hello world!"); // compile error - unreachable code

While this will compile fine even though it is semantically identical:

if (true) return;
System.out.println("hello world!");