CODEWITHSUNDEEP
c#dependency-injectionautomationinversion-of-control
Kobek
Kobek

Reputation: 1211

C# How to avoid passing arguments when using IoC resolver

I have recently began experimenting with dependency injection and IoC.

Here is an issue that bothers me... I have a dependency resolver, that basically allows me to do the following

injection.Register<IMyInterface, MySpecificType>();

This will automatically resolve the type, whenever something in the code asks for it and provide MySpecificType as an implementation for IMyInterface.

This works great for Controllers in ASP.NET for example, where the controller constructor has no parameters and is also dynamically invoked by the framework.

But what happens if I have my own custom class (in this case MySpecificType), and I want to dynamically resolve it. That is, whenever something in my code requires IMyInterface, the resolver should pass in MySpecificType.

Here is an example :

injection.Register<IMyInterface, MySpecificType>();


public SomeClass(IMyInterface dependency)
{
   //do something
}

Then somewhere in my code I wish to create a new instance ofSomeClass.

var instance = new SomeClass(//What do i pass here)

What should I pass in the constructor. Ofcourse I could do something like new MySpecificType(), but this would make my automatic dependency resolver pointless. Should I have a separate empty constuctor and use that or is this something that cannot happen.

Upvotes: 0

Views: 347

Answers (1)

user4478810
user4478810

Reputation:

You should ask your IoC container to instantiate your class for you. Then it really depends on your IoC container.

For example, with autofac

var builder = new ContainerBuilder();
builder.RegisterType<MyComponent>().As<IService>();
var container = builder.Build();

using(var scope = container.BeginLifetimeScope())
{
    var service = scope.Resolve<IService>();
}

Upvotes: 1

Related Questions