Question

In Wagner’s “Effective C#,” item 23, he explains that

interface methods are not
virtual…they are a declaration of a
concrete implementation.

I find that to be a conundrum, because it means that interface methods represent early binding, and yet they have the behavior of late-binding. It rouses curiosity of how they work under the covers. In C++ this would turn into a discussion of vtables. In C#, I don’t know what it turns into. Can someone spell it out?

p.s. This question has a cousin, but this question focuses on interfaces.
p.p.s. Please don’t worry about “you don’t need to know how it works.” Again, this is about curiosity.

Answer 1

Right, they are not virtual from the language point of view. But they actually are as far as the CLR is concerned. This sample code:

class Example : IDisposable {
    public void Dispose() {}
}

Produces this IL for the Dispose() method:

.method public hidebysig newslot virtual final    // <=== here
        instance void  Dispose() cil managed
{
  // Unimportant
} // end of method Example::Dispose

Note the attributes on the method: virtual and final. The final is what ensures that you can’t override the method in a derived class. Making the interface method implementation behave like a non-virtual method in the language but a virtual one at runtime.

This then also answers your question about early/late binding. It’s early, the v-table slot is filled in when the class is loaded.