Aug 20, 2008

C# Generics with examples

Parametric Polymorphism is a well-established programming language feature. Generics offers this feature to C#.
The best way to understand generics is to study some C# code that would benefit from generics. The code stated below is about a simple Stack class with two methods:

Push () and Pop ().
First, without using generics example you can get a clear idea about two issues:
a) Boxing and unboxing overhead and
b) No strong type information at compile type.
After that the same Stack class with the use of generics explains how these two issues are solved.
Example Code
Code without using generics:

public class Stack

object[] store;
int size;
public void Push(object x)
public object Pop()
Boxing and unboxing overhead:
You can push a value of any type onto a stack. To retrieve, the result of the Pop method must be explicitly cast back. For example if an integer passed to the Push method, it is automatically boxed. While retrieving, it must be unboxed with an explicit type cast.

Stack stack = new Stack();
int i = (int)stack.Pop(); //unboxing with explicit int casting

Such boxing and unboxing operations add performance overhead since they involve dynamic memory allocations and run-time type checks.
No strong Type information at Compile Time

Another issue with the Stack class:
It is not possible to enforce the kind of data placed on a stack.
For example, a string can be pushed on a stack and then accidentally cast to the wrong type like integer after it is retrieved:

Stack stack = new Stack();
//pushing the string
int i = (int)stack.Pop();
//run-time exception will be thrown at this point
The above code is technically correct and you will not get any compile time error. The problem does not become visible until the code is executed; at that point an InvalidCastException is thrown.

Code with generics
In C# with generics, you declare class Stack {...}, where T is the type parameter.

Within class Stack you can use T as if it were a type. You can create a Stack as Integer by declaring
Stack or Stack as Customer object by declaring
Simply your type arguments get substituted for the type parameter. All of the Ts become ints or Customers, you don't have to downcast, and there is strong type checking everywhere.
public class Stack
// items are of type T, which is kown when you create the object
T[] items;
int count;
public void Push(T item) {...}
//type of method pop will be decided when you creat the object
public T Pop() {...}
In the following example, int is given as the type argument for T:

Stack OpenTag Int CloseTag stack = new Stack OpenTag Int CloseTag ();
int i = stack.Pop();

The Stack type is called a constructed type. In the Stack type, every occurrence of T is replaced with the type argument int. The Push and Pop methods of a Stack operate on int values, making it a compile-time error to push values of other types onto the stack, and eliminating the need to explicitly cast values back to their original type when they are retrieved.
You can use parameterization not only for classes but also for interfaces, structs, methods and delegates.
//For Interfaces interface IComparable OpenTag T CloseTag

//for structs struct HashBucket OpenTag T CloseTag
//for methods static void Reverse OpenTag T CloseTag (T[] arr)
//for delegates delegate void Action OpenTag T CloseTag (T arg)

Inside the CLR
When you compile StackOpenTag T CloseTag , or any other generic type, it compiles down to IL and metadata just like any normal type. The IL and metadata contains additional information that knows there's a type parameter. This means you have the type information at compile time.
Implementation of parametric polymorphism can be done in two ways

1. Code Specialization: Specializing the code for each instantiation
2. Code sharing: Generating common code for all instantiations.

The C# implementation of generics uses both code specialization and code sharing as explained below.
At runtime, when your application makes its first reference to StackOpenTag T CloseTag , the system looks to see if anyone already asked for Stack . If not, it feeds into the JIT the IL and metadata for Stack and the type argument int.

The .NET Common Language Runtime creates a specialized copy of the native code for each generic type instantiation with a value type, but shares a single copy of the native code for all reference types (since, at the native code level, references are just pointers with the same representation).

In other words, for instantiations those are value types: such as Stack OpenTag int CloseTag , Stack OpenTag Long CloseTag , Stack OpenTag Double CloseTag , Stack OpenTag Float CloseTag, CLR creates a unique copy of the executable native code.

So StackOpenTag int CloseTag gets its own code. Stack OpenTag Long CloseTag gets its own code. Stack OpenTag Float CloseTag gets its own code. Stack OpenTag int CloseTag uses 32 bits and Stack OpenTag Long CloseTag uses 64 bits. While reference types, Stack OpenTag Dog CloseTag is different from Stack OpenTag Cat CloseTag , but they actually share all the same method code and both are 32-bit pointers.

This code sharing avoids code bloat and gives better performance.
To support generics, Microsoft did some changes to CLR, metadata, type-loader,language compilers, IL instructions and so on for the next release of Visual Studio.NET(code named Whidbey).

What you can get with Generics
Generics can make the C# code more efficient, type-safe and maintainable.
Efficiency: Following points states that how performance is boosted.
Instantiations of parameterized classes are loaded dynamically and the code for their methods is generated on demand [Just in Time].
Where ever possible, compiled code and data representations are shared between different instantiations.
Due to type specialization, the implementation never needs to box values of primitive types.

Safety: Strong type checking at compile time, hence more bugs caught at compile time itself.
Maintainability: Maintainability is achieved with fewer explicit conversions between data types and code with generics improves clarity and expressively.

Generics gives better performance, type safety and clarity to the C# programs. Generics will increase program reliability by adding strong type checking. Learning how to use generics is straightforward, hopefully this article has inspired you to look deeper into how you can use them.

No comments: