I have a legacy code, and I have a problem with reconstructor it.
At start of my application I load from WCF to property on App (this is SL application) list of users.
Then every control (for sending emails, view calendar and assigning tasks) use this property as
(App.Current as App).Users
Now, I'm trying to create Unit Test for one of controls that use this lists, and I'm stuck.
Should I make a Constructor Injection(I'm using Unity) with App as parameter? Or maybe introduce some class to hold this list?
Updated with OP's implementation as the pseudocode was incomplete.
I propose create an interface for all your application services
Inject IApplicationService to your modules.
You can use this interface for all the services the application provides(probably you will need more). Mock the interface for the unit tests
OP's implemantation
public interface IApplicationService
{
List<User> Users{get;set;}
}
public class ApplicationService : IApplicationService
{
public List<User> Users
{
get { return (App.Current as App).Users; }
set { (App.Current as App).Users = value; }
}
}
public partial class MainWindow : UserControl
{
readonly IApplicationService _applicationService
public MainWindow(IApplicationService applicationService)
{
_applicationService=applicationService;
}
}
I would create a wrapper class that will expose the list of users. In production code this class will just be a wrapper around your App.Current property and it can be injected in the constructor trough Unity.
In your Unit Tests you can easily mock the App parameter and pass it when constructing a new SUT.
Something like:
public interface IUserList
{
List<User> Users { get; }
}
public class SUT
{
private IUserList UserList { get; set; }
public SUT(IUserList userList)
{
this.UserList = userList;
}
}
public class AppUserList : IUserList
{
public List<User> Users
{
get
{
return ((App)App.Current).Users;
}
}
}
For Silverlight there is an extension model called Application Extension Services.
For infrastructure purposes that might be a better alternative than adding properties to your app class and casting App.Currentback and forth.
Downside of that model is the creation of a singleton you would have to initialize for your unit tests. It would also hide the dependency on Users in your consuming classes.
Your users seem to be just data. Making that data an ambient context which can be accessed and edited everywhere in your application will bite you. You don't know who does what with that data and when he does it. This is like a session state.
So making the dependency on your data explicit would be a first step to be able to track abuse of that data.
If it makes sense to you to create a "data holder object" that has a property for Users or directly inject that data into your consumers is up to you. If there is more data than just Usersit is tempting to put all of them into the same central data store object, even if your specific consumers don't need them.
Jimmy's answer is great, but can be provide quite a bit, and some errors fixed. Differences are explained at the bottom below the code/instructions:
Create a public interface: IUserService
public interface IUserService
{
// Implemented functionality as methods where possible for better
// extendability (like IoC)
IEnumerable<User> Users();
// Add any other user service stuff as you see fit.
void AddUser(User user);
}
Write a UserService that implements IUserService
public class UserService : IUserService
{
// If you need DI for this service, follow the same pattern of using
// fields and controller injection. I left examples in comment below.
// private readonly IRepository _repository;
// Constructor is unnecessary if you do not need DI example.
public UserService(/* IRepository repository */)
{
// _repository = repository;
}
// Methods
public IEnumerable<User> Users()
{
return ((App)App.Current).Users;
}
public void AddUser(User user)
{
((App)App.Current).Users.Add(user);
}
}
Inject IUserService into classes via their Constructor
In this case your MainWindow as an example:
public partial class MainWindow : UserControl
{
private readonly IUserService _userService;
public MainWindow(IUserService userService)
{
_userService = userService;
}
// Example method consuming the service
public IEnumerable<User> GetUsers()
{
return _userService.Users();
}
}
Differences:
Separate your User Services from a central Application Service
Better modularity. In addition I use an IApplicationService for more central/global data like Api Keys, Timeouts, cleanup, DB prepping, etc.
Return IEnumerable<T> instead of List<T>
This is just a golden rule of thumb for keeping things dry and not imposing hard instantiations on your consuming classes. Refactoring is easier/safer, and your code more extensible.
Use methods instead of properties
This is preference, but I think it smart in a service layer to use methods where possible so that you can introduce filters and overloads or continue to use dependency injection - for example, you could add GetUsers(string lastName), GetUsers(string lastName, string firstName) and maintain a clean interface for your consuming classes.
Cast App.Current without the as keyword
This is a good practice because using the as keyword means when the cast fails it will return null, rather than throw an exception. I prefer the exception because 99% of the time, if your cast fails, your next operations will too. :)
Enjoy!
Related
If I have a service that relies on data obtained through runtime, what is the best way to inject it into a class?
I have an Order class:
class Order {
string OrderID { get; set; }
string CustomerName { get; set; }
...
}
I want to encapsulate a lot of logic from the database, so I have a service:
class OrderService {
private readonly IOrderRepository _orderRepository;
private readonly IOrder _order;
public OrderService(IOrderRepository orderRepository, IOrder order) {
_orderRepository = orderRepository;
_order = order;
}
// some methods that compile data from the repository
public bool CheckAlreadyExists() { ... }
public string GetLatestShippingStatus() { ... }
...
...
public void Create() { ... }
}
Controller logic:
public class OrderController {
private readonly IOrderRepository _orderRepository
public OrderController(IOrderRepository orderRepository)
{
orderRepository = _orderRepository
}
public IActionResult Create(Order order)
// this is bad because now I have a dependency on IOrderRepository and OrderService
OrderService orderService = new OrderService(orderRepository, order)
if (!orderService.CheckAlreadyExists()) {
orderService.Create();
}
end
}
The two options I am aware of:
Refactor the code to pass runtime data into each of the functions instead
Create a factory OrderServiceFactory
I do not want to have to pass the parameter into every method which all rely on the same object. It seems like overkill to create a factory for every time I need this pattern, which seems like it should be a common use-case.
I think I'm fundamentally misunderstanding something.
Is there a pattern that I'm unaware of?
Could I create a service that keeps track of the runtime data?
Or am I just being stubborn and should create a factory?
I would simply comment but I don't have the reputation. Long story short, you need to be passing runtime data to OrderService. Read the link provided by Nkosi.
Having OrderService instantiated with a particular Order does not make sense. That means that you have to new up OrderService for every Order you get. Instead, OrderService should have per-lifetime scope. Ie - you can use the same instance of OrderService with multiple Orders. It's not overkill to pass runtime data to every method of a service; it's standard. You're overcomplicating things by forcing your service to rely on an instance of the object it is servicing. And your OrderRepository should not be injected in your controller at all. Use the service to call repository methods.
I'm using Repository and UoW pattern. My services look like this:
public class MyService : IService
{
private readonly IUnitOfWork<MyContext> unitOfWork;
private readonly IMyRepository myRepository;
public MyService(IUnitOfWork<MyContext> unitOfWork, IMyRepository myRepository)
{
this.unitOfWork = unitOfWork;
this.myRepository = myRepository;
}
//Methods...
}
Within services, I need to use other entities (for example to check for rights, etc).
Is it recommended to use the relevant repositories in the service or use the services directly?
Also, for each user we have rights (boolean) for each CRUD action. These rights are stored in the database.
Should checking of rights be done at the controller level or at the service level?
My golden rule is:
When you get business logic in your UI create a service, otherwise use
the repository directly.
So if you have this code in the UI:
var user = repos.Get(1);
user.FirstName = txtFirstName.Text;
repos.Save(user);
You are fine in my opinion. But if you instead have something like:
var user = userRepository.Get(1);
var accessChecker = authorizationRepository.GetForUser(id);
if (!accessChecker.MaySendEmail(user))
throw new SecurityException("You may not send emails");
var emailSender = new EmailSenderService();
emailSender.Send(user, txtDestination.Text, txtMessage.Text);
repos.Save(user);
It's likely that you should use a service instead.
Don't use your UoW to just wrap your database context. Since all your repositories are directly dependent of a given context (more or less, ofc), your repositories can be included in the UoW. Something along the lines of:
public interface IUnitOfWork<TContext> : IDisposable { }
public abstract class UnitOfWork<TContext> : IUnitOfWork<TContext> {
private readonly TContext _context;
protected TContext Context { get{ return _context; } }
protected UnitOfWork(TContext context){
_context = context;
}
}
public interface IMyDbUnitOfWork : IUnitOfWork<MyContext>{
public ICarRepository Cars { get; }
public IOwnerRepository Owners { get; }
}
public class MyDbUnitOfWork : UnitOfWork<MyContext>, IMyDbUnitOfWork{
public MyDbUnitOfWork():base(new MyContext()){}
private ICarRepository _cars;
public ICarRepository Cars {
get{
return _cars ?? (_cars = new CarRepository(Context));
}
}
private ICarRepository _owners;
public IOwnerRepository Owners {
get{
return _owners ?? (_owners = new OwnerRepository(Context));
}
}
}
public class MyService : IService
{
private readonly IMyDbUnitOfWork _unitOfWork;
public MyService(IMyDbUnitOfWork unitOfWork)
{
_unitOfWork = unitOfWork;
}
//Methods...
}
Obviously you can create this more or less generic, but I believe this should be enough to pass my point.
As a note, and since I normally use IoC frameworks, my services receive an IUnitOfWorkFactory because of the diferent lifestyles.
For the permissions question, it really depends how much control you want to have and how user friendly you want your application to be. Normally is a mix of both. Your application should know if your user has access to the screen but also if you must disable buttons accordingly. Since you also must prevent that, if by any reason, the user can invoke your service method, you can't allow it.
To solve this problem I don't filter by CRUD actions but by Service actions instead, intercepting every service invocation, which makes it easy to map my permissions to the user interface since normally is a 1 to 1 relation between button action and service action.
I think using repositories is just fine. I wouldn't invent a service layer for each of the repos.
Repository is used for abstracting the data access and service layer is to encapsulate business logic, however with recent trend , I find this overkill. Having service layer is fine if they act as controllers but don't try to map one to one to each entity or repo.
I typically use services from the UI and those services in turn use the repositories. I also find it useful to have some domain objects that encapsulate reusable logic in the services.
I do this so that rather than services calling each other and getting circular references, services use a common domain object instead. This avoids circular references and people copying and pasting the same code all over the place.This domain object may then use the repositories if necessary.
I have been helping a few friends on a project and there is a class that uses Ninject. I am fairly new to C# and I have no idea what that class is doing, which is why I need to understand Ninject. Can anyone explain what Ninject is and when does one use it(with example if possible)? Or if you can point to some links that would be great too.
I tried this question: Ninject tutorials/documentations? but it didn't really help a beginner like me.
Ninject is dependency injector for .NET, practical realisation of pattern Dependency Injection (form of Inversion of Control pattern).
Suppose you have two classes DbRepository and Controller:
class Controller {
private DbRepository _repository;
// ... some methods that uses _repository
}
class DbRepository {
// ... some bussiness logic here ...
}
So, now you have two problems:
You must initialize _repository to use it. You have several options for doing this:
Manually, within the constructor. But what if the constructor of DbRepository changes? You would need to rewrite your Controller class because code it's dependent upon was changed. It's not hard if you have only one Controller, but if you have a couple of classes that have a dependency on your Repository you have a real problem.
You can use a service locator or factory. But now you have a dependency on your service locator. You have a global service locator and all code must use it. How you will you change the behavior of your service locator when you need to use it in one part of your code for activation logic but for something else in another part of your code? There is only one way - passing the service locator through constructors. But with more and more classes you will need to pass it more and more times. Anyway, it's a good thought but in the long run, it's a bad idea.
class Controller {
private DbRepository _repository;
public Controller() {
_repository = GlobalServiceLocator.Get<DbRepository>()
}
// ... some methods that uses _repository
}
You can use dependency injection. Look at the code:
class Controller {
private IRepository _repository;
public Controller(IRepository repository) {
_repository = repository;
}
}
Now when you need your controller you write: ninjectDevKernel.Get<Controller>(); or ninjectTestKernel.Get<Controller>();. You can switch beetween dependency resolvers as fast as you want. See? It's simple, you don't need to write a lot.
You can't create unit tests for it. Your Controller has a dependency on DbRepository and if you want to test some method that uses repository, your code will go to the database and ask it for data. That's slow, very slow. If your code in DbRepository changes, your unit test on Controller will fall. Only integration test must warn you of 'problems' in this case. What you need in unit tests - is to isolate your classes and test only one class in one test (in ideal - only one method). If your DbRepository code fails, you will think that Controller code failed - and that's bad (even if you have tests for DbRepository and Controller - they both will fail and you can start from the wrong place). It takes a lot of time to determine where the error really is. You need to know that class A is ok, and it was class B where something failed.
When you want to replace DbRepository with something else in all your classes, you have to do a lot of work.
You can't easily control the lifetime of DbRepository. An object of this class is created on initialization of Controller and deleted when Controller is deleted. There is no sharing between different instances of the Controller class and there is no sharing between other classes. With Ninject you can simply write:
kernel.Bind<IRepository>().To<DbRepository>().InSingletonScope();
A special feature of dependency injection - agile development! You describe that your controller uses a repository with interface IRepository. You don't need to write DbRepository, you can simply create a MemoryRepository class and develop Controller while another person develops DbRepository. When work on DbRepository is finished, you just rebind in your dependency resolver that default IRepository is now DbRepository. Have a lot of controllers? All of them will now use DbRepository. That's cool.
Read more:
Inversion of control (wiki)
Dependency injection (wiki)
Inversion of Control Containers and the Dependency Injection pattern (Martin Fowler)
Ninject is an Inversion of Control container.
What does it do?
Suppose you have a Car class that depends on a Driver class.
public class Car
{
public Car(IDriver driver)
{
///
}
}
In order to use the Car class you build it like so:
IDriver driver = new Driver();
var car = new Car(driver);
A IoC containter centralizes the knowledge about how to build classes. It is a central repository that knows a few things. For example, it knows that the concrete class that you need to use to build a car is a Driver and not any other IDriver.
For example, if you are developing a MVC application, you can tell Ninject how to build your controllers. You do so by registering which concrete classes satisfy specific interfaces. At run time Ninject will figure out which classes are needed to build the required controller, and all behind the scenes.
// Syntax for binding
Bind<IDriver>().To<Driver>();
This is beneficial because it lets you build systems that are more easily unit testable. Suppose that Driver encapsulates all database access for Car. In a unit test for Car you can do this:
IDriver driver = new TestDriver(); // a fake driver that does not go to the db
var car = new Car(driver);
There are entire frameworks that take care of automatically creating testing classes for you and they are called mocking frameworks.
For more information:
GitHub/Ninject Home
Inversion of Control
Inversion of Control Containers and the Dependency Injection pattern
Mock Object
Other answers are great but I would also like to point out this Implementing Dependency Injection using Ninject article.
This is one of the best articles I ever read which explains Dependency Injection and Ninject with a very elegant example.
Here's the snippet from the article:
Below Interface will be implemented by our (SMSService) and (MockSMSService), basically the new Interface (ISMSService) will expose the same behaviors of both services as the code below:
public interface ISMSService
{
void SendSMS(string phoneNumber, string body);
}
(SMSService) implementation to implement the (ISMSService) interface:
public class SMSService : ISMSService
{
public void SendSMS(string mobileNumber, string body)
{
SendSMSUsingGateway(mobileNumber, body);
}
private void SendSMSUsingGateway(string mobileNumber, string body)
{
/*implementation for sending SMS using gateway*/
Console.WriteLine("Sending SMS using gateway to mobile:
{0}. SMS body: {1}", mobileNumber, body);
}
}
(MockSMSService) with totally different implementation using the same interface:
public class MockSMSService :ISMSService
{
public void SendSMS(string phoneNumber, string body)
{
SaveSMSToFile(phoneNumber,body);
}
private void SaveSMSToFile(string mobileNumber, string body)
{
/*implementation for saving SMS to a file*/
Console.WriteLine("Mocking SMS using file to mobile:
{0}. SMS body: {1}", mobileNumber, body);
}
}
we need to implement a change to our (UIHandler) class constructor to pass the dependency through it, by doing this, the code which uses the (UIHandler) can determine which concrete implementation of (ISMSService) to use:
public class UIHandler
{
private readonly ISMSService _SMSService;
public UIHandler(ISMSService SMSService)
{
_SMSService = SMSService;
}
public void SendConfirmationMsg(string mobileNumber) {
_SMSService.SendSMS(mobileNumber, "Your order has been shipped successfully!");
}
}
Now, we have to create a separate class (NinjectBindings) which inherits from (NinjectModule). This class will be responsible to resolve dependencies at run time, then we’ll override the load event which is used to configure the binding in it. The nice thing about Ninject is that we do not need to change our code in (ISMSService), (SMSService), and (MockSMSService).
public class NinjectBindings : Ninject.Modules.NinjectModule
{
public override void Load()
{
Bind<ISMSService>().To<MockSMSService>();
}
}
Now in UI form code, we’ll use the binding for Ninject which will determine which implementation to use:
class Program
{
static void Main(string[] args)
{
IKernel _Kernal = new StandardKernel();
_Kernal.Load(Assembly.GetExecutingAssembly());
ISMSService _SMSService = _Kernal.Get<ISMSService>();
UIHandler _UIHandler = new UIHandler(_SMSService);
_UIHandler.SendConfirmationMsg("96279544480");
Console.ReadLine();
}
}
Now the code is using the Ninject Kernal to resolve all chain of dependencies, if we want to use the real service (SMSService) in Release mode (on production environment) instead of the mock one, we need to change on the Ninject binding class (NinjectBindings) only to use the right implementation or by using the #if DEBUG directive as below:
public class NinjectBindings : Ninject.Modules.NinjectModule
{
public override void Load()
{
#if DEBUG
Bind<ISMSService>().To<MockSMSService>();
#else
Bind<ISMSService>().To<SMSService>();
#endif
}
}
Now our binding class (NinjectBindings) is living on the top of all our execution code and we can control the configuration easily in once place.
Also, see What is Inversion of Control? some very simple examples are mentioned to understand IoC.
You have to understand the Dependency Injection(DI) first. Notice here,
public interface IService
{
void Serve();
}
public class Service1 : IService
{
public void Serve() {
Console.WriteLine("Service1 Called");
}
}
public class Service2 : IService
{
public void Serve() {
Console.WriteLine("Service2 Called");
}
}
public class Service3 : IService
{
public void Serve() {
Console.WriteLine("Service3 Called");
}
}
public class Client
{
private IService service;
public Client(IService _service) //Constructor injection
{
service = _service;
}
public void ServeMethod() {
service.Serve(); //Notice here, this Serve() method has no idea what to do.
} // runtime will assign the object, that is Ninject
}
class Program
{
static void Main(string[] args)
{
IService s1 = new Service1(); //N.B. Ninject assigns object with interface
Client c1 = new Client(s1);
c1.ServeMethod();
IService s2 = new Service2(); //N.B. Ninject assigns object with interface
c1 = new Client(s2);
c1.ServeMethod();
IService s3 = new Service3(); //N.B. Ninject assigns object with interface
c1 = new Client(s3);
c1.ServeMethod();
Console.ReadKey();
}
}
// Ninject creates object in runtime for interface in runtime in ASP.NET MVC project.
/*
Output:
Service1 Called
Service2 Called
Service3 Called
*/
I'm putting together a REST service using ASP.NET Web API & Ninject, though I suspect this might be a more general IoC question than anything specific to my IoC framework. I have a number of objects that need to access a simple cache of User entities:
public class UserCache
{
private IList<User> users;
private IUserRepositoryFactory factory;
[Inject]
public UserCache(IUserRepositoryFactory factory)
{
this.factory = factory;
this.users = new List<User>();
}
public void Add(int id)
{
IUserRepository repo = factory.Create(new TestContext());
this.users.Add(repo.Get(id));
}
public int Count { get { return this.users.Count; } }
}
In practice, the cache is read-through, and will fill itself with User entities using a UserRepository (and associated IUserRepository interface):
public class UserRepository : IUserRepository
{
private readonly TestContext context;
public UserRepository(TestContext context)
{
this.context = context;
}
public User Get(int id)
{
return new User() { Name = "Test User" };
}
}
The cache is long-lived and shared across the entire application. My question is this: I want to use my UserRepository to pull User entities from my database. This repository needs to be injected into the cache somehow, or instantiated using a factory.
The trick is, the only way I've been able to both a) create the cache such that Ninject will inject its dependencies and b) have access to the cache throughout the same is to bind the cache in singleton scope and inject it into objects that need access to it:
kernel.Bind<TestContext>().ToSelf();
kernel.Bind<UserCache>().ToSelf().InSingletonScope();
...and then in a controller (for example):
[Inject]
public UserCache Cache { get; set; }
My question is, is this the best way to treat long-lived objects that require injection? Or is there some better way that I'm missing? I don't want to give the cache (or any other objects like it) direct access to the Ninject kernel.
Isn't this supposed to be the other way around? You should use IUserRepository in your controllers and the repository under the hood should fetch the data from cache (better if done using an interceptor) if it is already cached, otherwise should hit the database.
That way you don't have to worry about lifecycle of the long living cached objects. Remember that at the end of the day the whole WebAPI (so far) runs on the web stack, this means the application can be recycled unexpectedly based on different factors.
I have a class library:
public class SomeBL : ISomeBL
{
private IUser myUser;
public SomeBL(IUser user)
{
myUser = user;
}
public void TestMethod()
{
...some code using the user...
}
}
I also have a factory in this code library:
public class BLFactory
{
public static ISomeBL SomeBL
{
get { return ServiceLocator.Current.GetInstance<ISomeBL>(); }
}
}
Then I have a separate wcf application with one service that looks like this:
public class MyWcfService : IMyWcfService
{
public void TestMethod(User user)
{
BLFactory.SomeBL.TestMethod();
}
}
As you can see I am in need of IoC to properly resolve the IUser property on the SomeBL constructor level. I also don't want to pass it explicitly.
I was wondering if it's possible to configure Windsor in such a way that IUser will be resolved dynamically using value from the wcf service method's parameter ?
p.s.
Let's forget about wcf's inability to pass interfaces for a moment.
Edit#1
I solved it using Castle Project's Wcf Facility. Smooth as silk after I added it!
Based on the way you have things set up, and with your conditions I don't see how it's possible. There is no way for the container to "just know" the context by configuration.
However, I see a few options.
The first is to make your BLFactory a proper abstract factory, and pass the user to its Create method:
public class BLFactory
{
public ISomeBL Create(IUser user)
{
return new SomeBL(user);
}
}
You could also do this by calling Resolve<>() and passing the parameter there, or using Windsor's Typed Factory Facility. Referencing the container to directly resolve service in a factory class is generally not a good practice (see Three Calls Pattern).
Second option would be to pass the user as a method parameter (although you said you don't want to do this):
public class SomeBL : ISomeBL
{
public void TestMethod(IUser user)
{
...some code using the user...
}
}
This makes SomeBL more of a pure service (stateless), which IMHO is more along the lines of what DI and Windsor should be used for.
A final option assumes that the user represents the logged in user (if this is incorrect, ignore this option). Look at creating a service that returns the current logged-in user and inject that servce into your class. You would use some form of Ambient Context to store the user (on login or at some other point) and retrieve the user via this service.
public class SomeBL : ISomeBL
{
private IUser _userservice;
public SomeBL(IUserService userservice)
{
_userservice = userservice;
}
public void TestMethod()
{
IUser currentUser = _userService.GetCurrentUser();
}
}
public interface IUserService
{
IUser GetCurrentUser();
}
public class UserService : IUserService
{
public IUser GetCurrentUser
{
//pull user from Thread, HttpContext.CurrentRequest, cache, session, etc.
}
}
It can be done by the use of Typed Factory see http://stw.castleproject.org/Windsor.Typed-Factory-Facility-interface-based-factories.ashx