Why C# implements integer type as a struct and not as a primitive type?

Question

Looking on how the int type is implemented as the System.Int32 struct in C#, I wonder why was implemented like this (by wrapping this into a struct type), and not opted for a primitive type implementation like in Java?

Doesn't the additional wrapping in a struct produces additional performance hits?

Answer 1

There is no "wrapping" in a struct going on here. For all practical purposes, System.Int32 struct is a built-in primitive type, in the sense that the compiler recognizes them, and generates special instructions while handling expressions of primitive types. The only place where an int (or any other struct for that matter) is wrapped is during a boxing conversion, which is required when you want to pass an int to an API that takes an object.

The biggest difference between Java's and C#'s handling of primitives is that you can use C# primitives in places where user-defined struct types can go, most notably, in C# generic arguments, while Java treats primitives as a completely separate group of types.