Right now I am learning MVC, but have running into a problem that I can't figure out.
In a test project I have downloaded I can see that some controllers doesn't have a parameterless constructor, instead they have a constructor that has many parameters. How is that possible?
I mean, when someone visit a page a controller must be created first. The thing that creates the controller must create it with the parameterless constructor.
I guess that it is something that I have missed in the MVC. :)
This is known as constructor injection, which is used as the basis of dependency injection tools like Ninject and StructureMap. It allows for the loose coupling of objects.
Essentially whenever you have to create an instance of an object using new you are tightly coupling yourself to a concrete implementation of that object. This can make things difficult if not impossible for testing.
void MakeBacon()
{
var smokey = new Bacon();
}
By using injection methods such as constructor or dependency injection we are loosely coupling ourselves as we no longer have to create a new instance, it is injected instead.
void MakeBacon(IBacon smokey){
// Make bacon here.
}
Here is a good answer on stack exchange explaining constructor injection
What is constructor injection?
And here is a Microsoft article on constructor injection
DependencyResolver.SetResolver() when your app starts up. It will then try to resolve constructor dependencies using that resolver.Without seeing the code, my first thought is that there is some dependency injection involved (i.e. Ninject, Autofac, etc...) which will inject a constructed object (from rules defined) into the constructors. If you are unfamiliar with the Inversion of Control and Dependency Injection patterns, here is a stack over flow question that has a pretty clear explanation: What is Inversion of Control?
If something else is going on, then perhaps you could include some code so we could take a look at it.
