Update: Is there a way to achieve what I'm trying to do in an IoC framework other than Windsor? Windsor will handle the controllers fine but won't resolve anything else. I'm sure it's my fault but I'm following the tutorial verbatim and objects are not resolving with ctor injection, they are still null despite doing the registers and resolves. I've since scrapped my DI code and have manual injection for now because the project is time sensitive. Hoping to get DI worked out before deadline.
I have a solution that has multiple classes that all implement the same interface
As a simple example, the Interface
public interface IMyInterface {
string GetString();
int GetInt();
...
}
The concrete classes
public class MyClassOne : IMyInterface {
public string GetString() {
....
}
public int GetInt() {
....
}
}
public class MyClassTwo : IMyInterface {
public string GetString() {
....
}
public int GetInt() {
....
}
}
Now these classes will be injected where needed into layers above them like:
public class HomeController {
private readonly IMyInterface myInterface;
public HomeController() {}
public HomeController(IMyInterface _myInterface) {
myInterface = _myInterface
}
...
}
public class OtherController {
private readonly IMyInterface myInterface;
public OtherController() {}
public OtherController(IMyInterface _myInterface) {
myInterface = _myInterface
}
...
}
Both controllers are getting injected with the same interface.
When it comes to resolving these interfaces with the proper concrete class in my IoC, how do I differentiate that HomeController needs an instance of MyClassOne and OtherController needs an instance of MyClassTwo?
How do I bind two different concrete classes to the same interface in the IoC? I don't want to create 2 different interfaces as that breaks the DRY rule and doesn't make sense anyway.
In Castle Windsor I would have 2 lines like this:
container.Register(Component.For<IMyInterface>().ImplementedBy<MyClassOne>());
container.Register(Component.For<IMyInterface>().ImplementedBy<MyClassTwo>());
This won't work because I will only ever get a copy of MyClassTwo because it's the last one registered for the interface.
Like I said, I don't get how I can do it without creating specific interfaces for each concrete, doing that breaks not only DRY rules but basic OOP as well. How do I achieve this?
Update based on Mark Polsen's answer
Here is my current IoC, where would the .Resolve statements go? I don' see anything in the Windsor docs
public class Dependency : IDependency {
private readonly WindsorContainer container = new WindsorContainer();
private IDependency() {
}
public IDependency AddWeb() {
...
container.Register(Component.For<IListItemRepository>().ImplementedBy<ProgramTypeRepository>().Named("ProgramTypeList"));
container.Register(Component.For<IListItemRepository>().ImplementedBy<IndexTypeRepository>().Named("IndexTypeList"));
return this;
}
public static IDependency Start() {
return new IDependency();
}
}
I hope you can use service overrides.
Ex.
container.Register(
Component.For<IMyService>()
.ImplementedBy<MyServiceImpl>()
.Named("myservice.default"),
Component.For<IMyService>()
.ImplementedBy<OtherServiceImpl>()
.Named("myservice.alternative"),
Component.For<ProductController>()
.ServiceOverrides(ServiceOverride.ForKey("myService").Eq("myservice.alternative"))
);
public class ProductController
{
// Will get a OtherServiceImpl for myService.
// MyServiceImpl would be given without the service override.
public ProductController(IMyService myService)
{
}
}
You should be able to accomplish it with named component registration.
container.Register(Component.For<IMyInterface>().ImplementedBy<MyClassOne>().Named("One"));
container.Register(Component.For<IMyInterface>().ImplementedBy<MyClassTwo>().Named("Two"));
and then resolve them with
kernel.Resolve<IMyInterface>("One");
or
kernel.Resolve<IMyInterface>("Two");
See: To specify a name for the component
Typically DI containers follow Register, Resolve and Release patterns. During the register phase there are two steps. The first is to specify the mapping as you are doing. The second step is to specify the rules which govern which to inject where.
This problem is very common when we try to address Cross cutting concerns using decorators. In these situations, you have multiple classes(decorators) implementing a single interface.
Briefly, we need to implement IModelInterceptorsSelector which allows you to write imperative code that decides which Interceptor to apply to which types or members.
This is elaborately described in the book Dependency Injection in .Net book by Mark Seemann. Look for chapter 9 interception or search for the above interface.
I am not an expert at this, but was searching for the exact same problem and found the ans in the above book.
Hope this helps.
Regards
Dev1
Related
If there are more than one implementation of an interface, then I would have to use #Qualifier to indicate which class I want to autowire with. Wouldn't it make more sense to autowire the class directly instead of interface?
This is what object oriented programming (especially abstraction and polymorphism) is about.
You build the classes independendent of concrete implementations and use an interface.
This allows you to change what implementation you use at any point.
Assume you have this:
public interface I{
void doSomething();
}
#Component("a")
public class A implements I{
public void doSomething(){
//...
}
public void doSomethingElse(){
//...
}
}
#Component("b")
public class B implements I{
public void doSomething(){
//...
}
}
If you use
#Autowired
private A yourVariable;
you might be tempted to use the doSomethingElse() method that is an implementation detail (by what reason whatsoever).
However, if you use
#Qualifier("a")
private I yourVariable;
this cannot happen as the method is not present in the interface.
This gives you the ability to swap the implementation at any time.
Wouldn't it make more sense to autowire the class directly instead of interface?
This pretty much depends on how exactly do you work with qualifiers but in general, the answer is "No", autowiring the class is a bad thing to do if you have an interface - you should always work by interface.
Here are the examples:
interface I { void foo(); }
class IImpl1 implements I {
void foo() {System.out.println("Impl1");
pubic void bar() {// another method - why not?}
}
class IImpl2 implements I { void foo() {System.out.println("Impl2"); }
Note, that an implementation IImpl1 has an additional method bar that doesn't belong to the interface I
Case 1:
#Component
public class Driver {
#Autowired
private IImpl1 i;
}
Here class Driver is tightly could to the concrete implementation IImpl1 in general I can call method bar from the Driver class because I have this implementation, but in this case if I'll have to switch the implementation of Driver in future you'll have to also change the code: both change the reference, and also get rid of calls to IImpl1.bar() that might be not that easy to do - its a change in logic. As a result, the whole polymorphic behavior is lost.
By far its the worst way to program.
Now, consider Case 2:
#Component
public class Driver {
#Autowired
#Qualifier("someLogicalName") // under the hood it spring will map it to IImpl1 but you don't know that here
I i;
}
In this case - the things are much better - you never couple the driver to the concrete implementation at the level of code. This means that in general its enough to change the qualifier to get another implementation. You can't call bar anymore - it won't compile.
You also do not know what the implementation will be injected. If in future the implementation of the component that actually implements the logic of "someLogicalName" will change (by that I mean that you'll create a new class IImpl3), you won't feel it here - you'll place that qualifier on that new class and all the places that use it will get that implementation automatically.
Case 3:
#Configuration
public class MyConfiguration {
#Bean
public Driver driver(#Qualifier("someLogicalName") I i) {
return new Driver(i);
}
}
public class Driver {
private final I i;
public Driver(I i) { // in real life this constructor can be generated by
// lombok or something, put it here for the sake of completeness
this.i = i;
}
}
This by far the most "clean" approach - conceptually its really similar to case 2, however in this case you don't place the qualifier at the level of Driver class at all, in fact, the class Driver is not even aware at the level of code / metadata that its related to Spring anyhow. Of course all the "advantages" of case 2 are also applicable here as well.
What is the purpose of SudentRepository in this example? Why do I need one?
public class StudentController : Controller
{
private IStudentRepository _repository;
public StudentController() : this(new StudentRepository())
{
}
public StudentController(IStudentRepository repository)
{
_repository = repository;
}
I updated to actually include a specific question that I think you're getting at. The purpose of StudentRepository is to encapsulate interactions with persisted data. The Controller need not know if its stored in a db, flat file, in memory, etc.
The reason you're injecting it in via an interface is because you may eventually have multiple implementations of that repository, and the interface is just a contract to ensure basic functionality across all implementations. This is called constructor injection (a type of dependency injection) in case you want to learn more.
I am a newbie in TDD (Asp.net MVC3 environment) and trying to adopt TDD as our better better development approach.
In our production code,we have a following scenario
In web
//Autofac used to resolve Dependency
TestController(XService xSerivice,YSerivice yService)
{_xService =xService,_YService= yService}
[HTTPPost]
ActionResult Create(A1 a1)
{
_xService.XUnitOfWork.A1.add(a1)
_xService.XUnitOfwork.SaveChanges();
}
// where X, Y are different context,Concrete class, no interface implemented!
In Business Layer
Xservice(XUnitofWork) // no interface implemented!
In DAL Layer
'XUnitofWork:DataRepostory(Generic)...
{
GenericRepository<a1Entity> A1,
GenericRepository<a2Entity> A2
}
Now I realize that we should implement interface both in our BAL and Web layer.
My question is are there any way i can mock the services(XService,YService) in our controller to test some behavior (TDD) [for example save change exception occur while saving a entity via' _xService.XUnitOfwork.SaveChanges()'?
Please help.Thanks in Advance!
If you mark members (properties, methods) in your concrete class as virtual, I think you may be able to just mock those methods / properties individually. (I think the VB equivalent of virtual is Overridable..?)
Moq works by creating a new concrete implementation of something at runtime when your test runs. This is why it works so well with interfaces and abstract classes. But if there is no interface or abstract class, it needs to override a method or property.
Reply to question author's answer:
Since you are a self-proclaimed TDD newbie, I just wanted to point out that adding a parameterless constructor to a class just for the sake of making the class testable should not be an acceptable solution.
By giving your GenericRepository class a hard dependency on Entity Framework's DbSet / IDbSet, you are creating a tight coupling between your repository implementation and EF... note the using System.Data.Entity line at the top of that file.
Any time you decide to add a constructor dependency, you should seriously consider adding it as an interface or abstract class. If you need access to members of a library which you do not control (like EF's DbContext), follow Morten's answer and wrap the functionality in your own custom interface.
In the case of DbContext, this class does more than just provide you with a UnitOfWork implementation. It also provides you a way of querying out data and adding / replacing / removing items in your repository:
public interface IUnitOfWork
{
int SaveChanges();
}
public interface IQuery
{
IQueryable<TEntity> GetQueryable<TEntity>() where TEntity : class;
}
public interface ICommand : IQuery
{
void Add(object entity);
void Replace(object entity);
void Remove(object entity);
}
You can pretty easily wrap DbContext in these 3 interfaces like so:
public class MyCustomDbContext : DbContext, IUnitOfWork, ICommand
{
// DbContext already implements int SaveChanges()
public IQueryable<TEntity> GetQueryable<TEntity>() where TEntity : class
{
return this.Set<TEntity>();
}
public void Add(object entity)
{
this.Entry(entity).State = EntityState.Added;
}
public void Replace(object entity)
{
this.Entry(entity).State = EntityState.Modified;
}
public void Remove(object entity)
{
this.Entry(entity).State = EntityState.Deleted;
}
}
Note how your interfaces take no dependencies on System.Data.Entity. They use primitives and standard .NET types like object, IQueryable<T>, and int. This way, when you give your generic repository dependencies on the interfaces, you can remove the dependency on System.Data.Entity:
// using System.Data.Entity; // no need for this dependency any more
public class GenericRepository
{
private readonly ICommand _entities;
private readonly IQueryable<TEntity> _queryable;
public GenericRepository(ICommand entities)
{
this._entities = entities;
this._queryable = entities.GetQueryable<TEntity>();
}
//public GenericRepository()
//{
// no need for a parameterless constructor!
//}
}
...and your GenericRepository is now fully unit testable, since you can easily mock any of these interface methods.
Final Notes:
Also, after seeing your answer to your own question, it looks like you have CompanyRepository as a property of your UnitOfWork class. You then inject UnitOfWork as a dependency on your CompanyInformationController. This is backwards. Instead, you should be injecting the CompanyRepository (or its interface) into the controller's constructor. The UnitOfWork pattern has nothing to do with maintaining references for your known repositories. It is about tracking multiple changes made to related items so that they can all be pushed once as a single transaction. EF does this automatically, so as long as AutoFac is providing the same DbContext instance no matter whether your app requests an IQuery, ICommand, or IUnitOfWork implementation, then the only method UnitOfWork should be concerned with is SaveChanges().
thanks for your reply. The test I was trying to do was successful after spending few hours and changes my previous code.
Changes are follows:
1) Now using UnitofWork in my controller instead of a redundant service.
2) Added a parameter less constructor to the GenericRepository Class.(with out any DBContext!),because it will requied a DBContext as a parameter in Constructor,which can not be substituted by supplying a Mocked DBContext.
GenericRepository:
public class GenericRepository where TEntity : class
{
internal DbContext _context;
internal DbSet<TEntity> dbSet;
public GenericRepository(DbContext context)
{
this._context = context;
this.dbSet = context.Set<TEntity>();
}
public GenericRepository() //newly added!
{
}
...............
Complete Test
[TestMethod]
public void Index_Return_OneModel_WhenCalling()
{
//arrange
AutoMapperExtension automapper = new AutoMapperExtension();
var moqentities = new Mock<SetupEntities>();
List<CompanyInformation> list =new List<CompanyInformation>();
list.Add(new CompanyInformation{ CompanyName = "a", CompanyAddress = "aa", Id = 1});
list.Add(new CompanyInformation { CompanyName = "b", CompanyAddress = "b", Id = 2 });
var unitOfWork = new Mock<UnitOfWork>(moqentities.Object);
unitOfWork.Setup(d => d.CompanyRepository).Returns(new GenericRepository<CompanyInformation>());
unitOfWork.Setup(d => d.CompanyRepository.GetAll()).Returns(list.AsQueryable());
var controller = new CompanyInformationController(unitOfWork.Object);
//Act
var result =(ViewResult) controller.Index();
var model =(CompanyInformationViewModel) result.ViewData.Model;
//Assert
Assert.AreEqual(1, model.Id);
}
The best way is to create an interface for XService. If that is not possible for some reason (if XService is a third party class that doesn't implement an interface), then consider wrapping the functionality in a wrapperclass that does have an interface.
I have a MVC3 project that uses the Entity Framework and Ninject v2.2, and follows the Unit of Work pattern with a Service Layer wrapping my repositories.
After looking at the code below, hopefully its apparent that Ninject is using constructor chaining to inject the correct classes. It currently works prefectly in my application, however I am at the point that I need to bind an instance of IDatabase to MyDatabase with a different scope such as InSingletonScope() or InNamedScope(), not InRequestScope(). I know that I can use the [Named("MyDatabaseScope")] Attribute to customize which IDatabase object is injected, however it seems that with my code structure, if I wanted to inject my SingletonScoped instance, I would have to recreate a new Abstract and Concrete Implementation of my Unit of Work, my Service and all my Repositories, that will then chain down.
Basically my application currently goes
Controller -> Unit of Work -> Database, (Repositories -> Database)
If I have to change my Database Binding, I will now have to create another chain in addition to the current one:
Controller -> New Unit of Work -> SingletonDatabase, (New Repositories-> SingletonDatabase)
This seems to completely defeat the DRY principal. Is there a way to, from the Controller Constructor, inform Ninject that when doing constructor chaining it should use my singleton (or named binding) rather than my request scope binding, without having to recreate all my classes with a Named attribute, or a new Interface?
Sorry for the long text, I wasnt sure if I could get the point across without my code snippets and my somewhat rambling explaination.
Ninject Module Load Function:
..snip..
Bind<IUserServices>().To<UserServices>();
Bind<IBaseServices>().To<BaseServices>();
Bind<IUserRepository>().To<UserRepository>();
Bind(typeof (IRepository<>)).To(typeof (RepositoryBase<>));
Bind<IUnitOfWork>().To<UnitOfWork>();
Bind<IDatabase>().To<MyDatabase>().InRequestScope();
//This is my problem:
//Bind<IDatabase>().To<MySingletonDatabase>().InSingletonScope();
Unit of Work Implementation Constructor:
public class UnitOfWork : IUnitOfWork
{
private IDatabase _database;
public UnitOfWork(IDatabase database,
IUserRepository userRepository,
IPeopleRepository peopleRepository,
)
{
this._database = database;
this.UserRepository = userRepository;
this.PeopleRepository = peopleRepository;
}
protected IDatabase Database
{
get { return _database; }
}
...snip...
}
User Service Layer Implementation Constructor:
public class UserServices : BaseServices, IUserServices
{
private IUnitOfWork _uow;
public UserServices(IUnitOfWork uow)
: base(uow)
{
_uow = uow;
}
...snip...
}
User Repository Constructor:
public class UserRepository : RepositoryBase<User>, IUserRepository
{
public UserRepository(IDatabase database)
: base(database)
{
}
...snip...
}
Controller Constructor:
public IUserServices _userServices { get; set; }
public ActivityController(IUserServices userServices)
{
_userServices = userServices;
}
}
Using Ninject 3.0.0 you can use WhenAnyAncestrorNamed("Some name") But if you need to run asyncronous things you should thing about splitting your application into a web frontend and a server backend. This could make many things easier.
I have the following code:
public abstract class BaseController : Controller
{
public IUserService UserService { get; set; }
}
All my controllers inherit from this base controller. I started out by configuring it in Ninject using the following code:
kernel.Bind<BaseController>()
.ToSelf()
.WithPropertyValue("UserService", x => x.Kernel.GetService(typeof(IUserService)));
This did not work. I assume it is because of the fact that the BaseController is an abstract class (please confirm my assumption). So I moved on to modify the configuration to:
kernel.Bind<HomeController>()
.ToSelf()
.WithPropertyValue("UserService", x => x.Kernel.GetService(typeof(IUserService)));
This does work. The minor downside is that I now have to configure every controller the same way.
Since I also have DependencyResolver setup in my ASP.NET MVC 3 project I could also remove the above Ninject configuration and modify my base controller to look like:
public IUserService UserService
{
get
{
return DependencyResolver.Current.GetService<IUserService>();
}
}
Is there any benefit to using the fluent configuration as opposed to using the DependencyResolver approach? Is one better than the other? Which approach would be considered a better practice?
It is worth mentioning that I did not want to do constructor injection in my base controller.
A better practice in MVC it is to use constructor injection over property injection. Why did you make your choice like this ?
Using Constructor Injection you states that all dependencies in constructor are necessary for the class to do its job.
Property injection means that the dependencies are optional or that there are the local defaults implementations, so all will work even if you don't provide necessary implementations yourself.
You should really know what you're doing using Property injection or you have no other choice, so the safer approach is to rely on constructor injection.
Now I'll give you my point of view. Other may have other opinions.
DependencyResolver was introduced in MVC 3 for "convenient" service location but for me it's a regular Service locator which for me is also an anti-pattern http://blog.ploeh.dk/2010/02/03/ServiceLocatorIsAnAntiPattern.aspx. I don't use it because I don't like it and there is no benefit in using it.
I prefer to user my controller factory like before and pass the dependencies through constructor.
More the IDependencyResolver has somme issues with some IoC containers (I don't know if it's the case with Ninject). You can read more here : http://mikehadlow.blogspot.com/2011/02/mvc-30-idependencyresolver-interface-is.html
If you need the same dependency in each controller then there seems to be something wrong in your design. Most likely you are handling some kind of cross cutting concern in your base controller. In this case Doing property injection is just treating sympthoms instead of cureing the disease. This should rather be handled by an aspect (e.g. a filter or an interceptor) so that you do not have to pollute your controller with something that does not belong there.
There are many ways to skin the cat they say. You could use conventions-based bindings with .WithPropertyValue() or with .OnActivaction() (as described here).
public class ControllerModule : NinjectModule
{
public override void Load()
{
// Get all controller types derived from the base controller.
IEnumerable<Type> controllerTypes = // ...
foreach (var controllerType in controllerTypes)
{
Bind(controllerType).ToSelf().InRequestScope()
.WithPropertyValue(...);
}
}
}
You could create your own custom implementation of the IInjectionHeuristic interface as described here or your own custom implementation of the IControllerActivator interface.
public class CustomNinjectControllerActivator : IControllerActivator
{
private readonly IKernel kernel;
public CustomNinjectControllerActivator(IKernel kernel)
{
this.kernel = kernel;
}
public IController Create(RequestContext context, Type controllerType)
{
var baseController = kernel.TryGet(controllerType) as BaseController;
if (baseController == null)
{
return null;
}
baseController.UserService = kernel.Get<IUserService>();
return baseController;
}
}
Heck, you could even use the service locator pattern if you are comfortable using it.
public IUserService UserService
{
get { return DependencyResolver.Current.GetService<IUserService>(); }
}
You should choose whichever solution is easiest to implement, test and maintain, and of course provides the desired behavior.