Question

Possible Duplicate:
Generic constraints, where T : struct and where T : class

Is there a particular reason that you cannot overload generic methods using mutually exclusive Type constraints in C#? For instance, take these methods:

T DoSomething<T>(T arg) where T : class
{ /* Do something */ }

T DoSomething<T>(T arg) where T : struct
{ /* Do something */ }

and try to invoke them with

DoSomething("1");
DoSomething(1);

The way I see it, the DoSomething() methods are mutually exclusive as far as the parameters that they will take – the first one takes a reference type, the second takes a value type. The compiler should be able to tell that the DoSomething call with a string argument goes to the first method and the DoSomething call with an int argument goes to the second method.

Am I missing something conceptually with generics here? Or is this just a feature that wasn’t implemented in C#?

Answer 1

Generic constraints are not part of the method signature

See this answer Generic contraints on method overloads

Jon Skeet blog post on the topic