I was wondering how far I should take my IoC implementation in terms of loosely coupling my application. I am using constructor injection in my MVC controller but wondered if I should also be resolving all my other dependencies in the controller instead of newing up objects. For example if I was to create a new User object within a controller method would I use?
var user = new User();
var user = myContainer.Resolve<IUser>();
I like the idea of breaking my dependency on User but is that going too far and could possible make my code harder to read?
This is a very good question, when you read and hear about DI it makes sense and this is the next natural conclusion.
First Steven's point is correct you should not pass the container around. If you need to construct IUser on the fly and it needs to be abstract you should rather inject an Abstract Factory.
Sebastian also posted a good link.
I actually wrote about this in an answer I posted here.
The bottom section of the post:
Not every object and dependency needs or should be dependency injected, first consider if what you are using is actually considered a dependency:
What are dependencies?
Application Configuration
System Resources (Clock)
Third Party Libraries
Database
WCF/Network Services
External Systems (File/Email)
Any of the above objects or collaborators can be out of your control and cause side effects and difference in behavior and make it hard to test. These are the times to consider an Abstraction (Class/Interface) and use DI.
What are not dependencies, doesn't really need DI?
List
MemoryStream
Strings/Primitives
Leaf Objects/Dto's
So in your particular case it really depends on what happens when IUser is constructed also whether you actually need to substitute it with different implementations. If this is not the case and User has only simple types with no external dependencies or side effects just new it up.
Consider what happens when you call new User(), look at the graph below, if it causes other objects to be created and looks like something in the graph below consider IoC.
A cascading dependency object graph:
In this example new on the object either requires or creates a whole bunch of other dependencies. It is most likely out of your control what those dependencies are or do.
When your object is a simple dto it doesn't suffer this problem and IoC is likely not so much required.
Related
Following the guidelines I read in:
https://www.devtrends.co.uk/blog/how-not-to-do-dependency-injection-the-static-or-singleton-container
I want to try and avoid using a service locator.
But on the other hand, I don't register all the types in the startup.cs file. I don't think this is right that all these internal types are referenced in the main startup.cs
I currently have a factory class that has a collection of builder classes.
Each builder class is in charge of creating a specific object.
I don't want to create all these builder classes in advance as I might not need to use them and creating them is a bit heavy.
I saw an example of how to achieve this in the link above. However the startup.cs class needs to know all these builders. I don't think this is appropriate, I'd rather have the factory class be the only one that is exposed to them. I was trying to understand if there is some kind of func/action method that I can inject from the startup.cs file into my factory class. This func/action will be in charge of creating/registering the builders and then I can activate this func/action within the class factory. I'd like this func/action to receive the interface/class/maybe name of the builder but using generics isn't working. I searched a lot and didn't find any solution so I assume this is not possible.
Seems I have 2 options:
1. Use service locator. This way only the factory class will know the builders. However if in the future, if I want to change the DI I need to "touch" the factory class (I'm contaminating the factory class). Wanted all the DI code to be located only in the startup.cs class.
2. Register the builders in the startup.cs but now the startup.cs is aware of the builders. This kinda couples the code, not really single role of responsibility
It would have been great to inject the factory class a func/action from the startup.cs that would do the registration but the factory class itself activates it.
Is this possible?
I want to try and avoid using a service locator
Great, because the Service Locator is an anti-patttern.
don't register all the types in the startup.cs file.
You should do your registrations in one single 'area' of your application: the start-up path. This area is commonly referred to as the Composition Root (the place where object graphs are composed).
I don't think this is right that all these internal types are referenced in the main startup.cs
No matter how you design it, the startup assembly is the most volatile part of the system and it always depends on all other assemblies in the application. Either directly or transitively (through another referenced assembly). The whole idea of Dependency Injection is to minimize the coupling between components and the way to do this is to centralize coupling by moving it to the Composition Root. By making types internal however, you are decentralizing object composition and that limits your flexability. For instance, it becomes harder to apply decorators or interceptors for those registered types and control them globally. Read this question and its two top voted answers for more information.
I don't register all the types
The concern of having a Composition Root that is too big is not a valid one. One could easily split out the Composition Root into multiple smaller functions or classes that all reside in the startup assembly. On top of that, if you read this, you'll understand that registering all types explicitly (a.k.a. "Explicit Register") is typically pointless. In that case you're probably better off in using DI without a Container (a.k.a. Pure DI). Composition Roots where all types are registered explicitly are not very maintainable. One of the areas a DI Container becomes powerful is through its batch-registration facilities. They use reflection to load and register a complete set of types in a few lines of code. The addition of new types won't cause your Composition Root to change giving you the highest amount of maintainability.
I don't want to create all these builder classes in advance as I might not need to use them and creating them is a bit heavy
Creation of instances should never be heavy. Your injection constructors should be simple and composing object graphs should be reliable. This makes building even the biggest object graphs extremely fast. Factories should be reduced to an absolute minimum.
TLDR;
Register or compose your object graphs solely in the Composition Root.
Refrain from using the Service Locator anti-pattern; Whole applications can (and should) be built purely with Constructor Injection.
Make injection constructors simple and prevent them from doing anything else than storing their incoming dependencies.
Refrain from using factories to compose services, they are not needed in most cases.
I'd like to create a good app in ASP.NET MVC 5 using EF 6 Code first concept. I want it to be well-designed i.e. having generally speaking: Presentation, Logic and Data layers separated. I want it to be testable :)
Here's my idea and some issues related with creating application
Presentation layer: It's my whole MVC - view models(not models), views, controllers
I believe that's validation should be done somewhere else (in my opinion - it's a part of business logic) but it's quite convenient to use attributes from the DataAnnotations namespace in ViewModelds and check validation in controller.
Logic layer: Services - classes with their interfaces to rule business logic.
I put there functions like: AddNewPerson(PersonViewModel Person), SendMessageToPerson(...).
They will use DB context to make their actions (there's a chance that not all of them will be relying on context). There's a direct connection between service and db - I mean the service class have reference do context.
Where should I do mapping between ViewModel and Model? I've heard that service is a bad place for it - so maybe in controllers. I've heard that service should do the work related with db exclusively.
Is it right? Is my picture of service layer is good?
Data layer: I've read about Repository and UoW patterns a lot. There're some articles which suggest that EF6 implements these two things. I don't want to create extra code if there's no need for such a behavior. The question is: am i right to assume that i don't need them?
Here's my flow:
View<->Controllers(using ViewModels)<->Services(using Models)<->DB.
**I'm gonna use DI in my project.
What do you think about my project structure?
There is no reason to use a Unit of Work pattern with Entity Framework if you have no need to create a generic data access mechanism. You would only do this if you were:
using a data access technology that did not natively support a Unit of work pattern (EF does)
Wanted to be able to swap out data providers sometime in the future.. however, this is not as easy as it might seem as it's very hard NOT to introduce dependencies on specific data technologies even when using an Unit of Work (maybe even BECAUSE you are)... or
You need to have a way of unifying disparate data sources into an atomic transaction.
If none of those are the case, you most likely don't need a custom Unit of Work. A Repository, on the other hand can be useful... but with EF6 many of the benefits of a Repository are also available since EF6 provides mocking interfaces for testing. Regardless, stay away from a generic repository unless it's simply an implementation detail of your concrete repositories. Exposing generic repositories to your other layers is a huge abstraction leak...
I always use a Repository/Service/Façade pattern though to create a separation between my data and business (and UI and business for that matter) layers. It provides a convenient way to mock without having to mock your data access itself and it decouples your logic from the specific that are introduced by the Linq layer used by EF (Linq is relatively generic, but there are things that are specific to EF), a façade/repository/server interface decouples that).
In general, you're on the right path... However, let me point out that using Data Attributes on your view models is a good thing. This centralizes your validation on your model, rather than making you put validation logic all over the place.
You're correct that you need validation in your business logic as well, but your mistake is the assumption that you should only have it on the business logic. You need validation at all layers of your application.. And in particular, your UI validation may have different requirements than your business logic validation.
For instance, you may implement creating a new account as a multi-step wizard in your UI, this would require different validation than your business layer because each step has only a subset of the validation of the total object. Or you might require that your mobile interface has different validation requirements from your web site (one might use a captcha, while the other might use a touch based human validation for instance).
Either way, it's important to keep in mind that validation is important both at the client, server, and various layers...
Ok, let’s clarify a few things...
The notion of ViewModel (or the actual wording of ViewModel) is something introduced by Microsoft Martin Fowler. In fact, a ViewModel is nothing more than a simple class.
In reality, your Views are strongly typed to classes. Period. To avoid confusion, the wording ViewModel came up to help people understand that
“this class, will be used by your View”
hence why we call them ViewModel.
In addition, although many books, articles and examples use the word ViewModel, let's not forget that it's nothing more than just a Model.
In fact, did you ever noticed why there is a Models folder inside an MVC application and not a ViewModels folder?
Also, ever noticed how at the top of a View you have #model directive and not # viewmodel directive?
That's because everything could be a model.
By the way, for clarity, you are more than welcomed to delete (or rename) the Models folder and create a new one called ViewModels if that helps.
Regardless of what you do, you’ll ultimately call #model and not #viewmodel at the top of your pages.
Another similar example would be DTO classes. DTO classes are nothing more than regular classes but they are suffixed with DTO to help people (programmers) differentiate between all the other classes (including View Models).
In a recent project I’ve worked on, that notion wasn’t fully grasped by the team so instead of having their Views strongly typed to Models, they would have their Views strongly typed to DTO classes. In theory and in practice everything was working but they soon found out that they had properties such as IsVisible inside their DTO’s when in fact; these kind of properties should belongs to your ViewModel classes since they are used for UI logic.
So far, I haven’t answered your question but I do have a similar post regarding a quick architecture. You can read the post here
Another thing I’d like to point out is that if and only if your Service Layer plans on servicing other things such as a Winform application, Mobile web site, etc...then your Service Layer should not be receiving ViewModels.
Your Service Layer should not have the notion of what is a ViewModel. It should only accept, receive, send, etc... POCO classes.
This means that from your Controller, inside your ActionResult, once the ModelState is Valid, you need to transform your ViewModel into a POCO which in turn, will be sent to the method inside your Service Layer.
In other words, I’d use/install the Automapper nugget package and create some extension methods that would convert a ViewModel into a POCO and vice-versa (POCO into a ViewModel).
This way, your AddNewPerson() method would receive a Person object for its parameter instead of receiving a PersonViewModel parameter.
Remember, this is only valid if and only if your Service Layer plans on servicing other things...
If that's not the case, then feel free to have your Service Layer receive, send, add, etc...ViewModels instead of POCOs. This is up to you and your team.
Remember, there are many ways to skin a cat.
Hope this helps.
I have an MVC application with a typical architecture...
ASP.NET MVC Controller -> Person Service -> Person Repository -> Entity Framework DB Context
I am using Castle Windsor and I can see the benefit of using this along with a ControllerFactory to create controller with the right dependencies. Using this approach the Controller gets a Service injected, which in turn knows how to construct the right Repository, which in turn knows the correct DbContext to use.
The windsor config is something like this...
dicontainer = new WindsorContainer();
dicontainer.Register(Component.For<IPersonService>().ImplementedBy<PersonService>());
dicontainer.Register(
Component.For<IPersonRepository>().UsingFactoryMethod(
() => new PersonRepository(new HrContext("connectionString"))));
It this the right way to do it? I don't like the UsingFactoryMethod, but can't think of another way.
Also, what if the Repository needed a dependency (say ILogger) that was not needed by the service layer? Does this mean I have to pass the ILogger into the service layer and not use it. This seems like a poor design. I'd appreciate some pointers here. I have read loads of articles, but not found a concrete example to verify whether I am doing this right. Thanks.
I try to avoid using factory methods (as you mentioned you felt this smelled funny). To avoid this, you could create a database session object that creates a new DbContext. Then your repositories just need to get an instance of IDbSession and use its dbContext property. Then, you can also easily control the scope of the IDbSession object (don't use singleton because it's not thread safe).
I wanted to make that point so that I could make this more important point... Make your constructors take in only objects that are registered in the DI container (no options or configurations in constructors). Options and configurations should be read/writen in classes whose sole purposes it is to read/write those values. If all classes follow this model, then your DI registration becomes easy and classes can just add whatever dependencies they need in their constructors.
If you are trying to use a third party library that has options in constructors, wrap that class in your own class that has an easy to consume constructor and uses a configuration class to read the values needed to pass to the third party library. This design also introduces a layer of abstraction between your code and the third party library which can then be used to more easily swap (or stub) third party libraries when necessary.
most of the time in the service code I would have something like this:
public SomeService : ISomeService
{
ISomeRepository someRepository;
public Do(int id)
{
someRepository.Do(id);
}
}
so it's kinda redundant
so I started to use the repositories directly in the controller
is this ok ? is there some architecture that is doing like this ?
You lose the ability to have business logic in between.
I disagree with this one.
If business logic is where it should be - in domain model, then calling repo in controller (or better - use model binder for that) to get aggregate root and call method on it seems perfectly fine to me.
Application services should be used when there's too much technical details involved what would mess up controllers.
I've seen several people mention using model binders to call into a repo lately. Where is this crazy idea coming from?
I believe we are talking about 2 different things here. I suspect that Your 'model binder' means using model simultaneously as a view model too and binding changed values from UI directly right back to it (which is not a bad thing per se and in some cases I would go that road).
My 'model binder' is a class that implements 'IModelBinder', that takes repository in constructor (which is injected and therefore - can be extended if we need caching with some basic composition) and uses it before action is called to retrieve aggregate root and replace int id or Guid id or string slug or whatever action argument with real domain object. Combining that with input view model argument lets us to write less code. Something like this:
public ActionResult ChangeCustomerAddress
(Customer c, ChangeCustomerAddressInput inp){
c.ChangeCustomerAddress(inp.NewAddress);
return RedirectToAction("Details", new{inp.Id});
}
In my actual code it's a bit more complex cause it includes ModelState validation and some exception handling that might be thrown from inside of domain model (extracted into Controller extension method for reuse). But not much more. So far - longest controller action is ~10 lines long.
You can see working implementation (quite sophisticated and (for me) unnecessary complex) here.
Are you just doing CRUD apps with Linq To Sql or trying something with real domain logic?
As You can (hopefully) see, this kind of approach actually almost forces us to move towards task based app instead of CRUD based one.
By doing all data access in your service layer and using IOC you can gain lots of benefits of AOP like invisible caching, transaction management, and easy composition of components that I can't imagine you get with model binders.
...and having new abstraction layer that invites us to mix infrastructure with domain logic and lose isolation of domain model.
Please enlighten me.
I'm not sure if I did. I don't think that I'm enlightened myself. :)
Here is my current model binder base class. Here's one of controller actions from my current project. And here's "lack" of business logic.
If you use repositories in your controllers, you are going straight from the Data Layer to the Presentation Layer. You lose the ability to have business logic in between.
Now, if you say you will only use Services when you need business logic, and use Repositories everywhere else, your code becomes a nightmare. The Presentation Layer is now calling both the Business and Data Layer, and you don't have a nice separation of concerns.
I would always go this route: Repositories -> Services -> UI. As soon as you don't think you need a business layer, the requirements change, and you will have to rewrite EVERYTHING.
My own rough practices for DDD/MVC:
controllers are application-specific, hence they should only contain application-specific methods, and call Services methods.
all public Service methods are usually atomic transactions or queries
only Services instantiate & call Repositories
my Domain defines an IContextFactory and an IContext (massive leaky abstraction as IContext members are IDBSet)
each application has a sort-of Composition Root, which is mainly instantiating a Context Factory to pass to the Services (you could use DI container but not a big deal)
This forces me to keep my business code, and data-access out of my controllers. I find it a good discipline, given how loose I am when I don't follow the above!
Here is the thing.
"Business Logic" should reside in your entities and value objects.
Repositories deal with AggregateRoots only. So, using your repositories directly in your Controllers kinda feels like you are treating that action as your "service". Also, since your AggregateRoot may only refer to other ARs by its ID, you may have to call one more than one repo. It really gets nasty very quickly.
If you are going to have services, make sure you expose POCOs and not the actual AggregateRoot and its members.
Your repo shouldn't have to do any operations other than creating, retrieving, updating and deleting stuff. You may have some customized retrieve based on specific conditions, but that's about it. Therefore, having a method in your repo that matches one in your service... code smell right there.
Your service are API oriented. Think about this... if you were to pack that service in a .dll for me to use, how would you create your methods in a way that is easy for me to know what your service can do? Service.Update(object) doesn't make much sense.
And I haven't even talked about CQRS... where things get even more interesting.
Your Web Api is just a CLIENT of your Service. Your Service can be used by another service, right? So, think about it. You most likely will need a Service to encapsulate operations on AggregateRoots, usually by creating them, or retrieving them from a repo, do something about it, then returning a result. Usually.
Makes sense?
Even with "rich domain model" you will still need a domain service for handling business logic which involves several entities. I have also never seen CRUD without some business logic, but in simple sample code. I'd always like to go Martin's route to keep my code straightforward.
I'm building an application that roughly follows the repository pattern with a service layer on top, similar to early versions of Conery's MVC Storefront.
I need to implement a page that returns all users except for the current user. I already have GetUsers() methods on the repository and service layers, so the question is where to apply the "except for the current user."
Should the service layer be aware of HttpContext, thus applying this rule? I am tempted to just pass in the current user (id) from the controller to this service method, but it seems cleaner if the service layer was HttpContext-aware and could do this on its own.
One obvious alternative is to apply this rule directly within the Controller, but I'm just not hot on that idea...
Edit - just to comment on the answers: I see the issues with the reverse dependency problem, something I was completely overlooking. I'm marking Mehrdad's as the answer due votes, but everyone has really provided a valuable response worth reading!
Absolutely not. My mindset in designing these kind of things is like this: I assume I need to write a Windows based app along with the Web application and try to minimize dependency on Web specific stuff. Passing HttpContext directly will increase coupling of your service layer to your Web UI layer which is not ideal.
The answer is, no.
Not only should the Service Layer have no dependency on the current Presentation Layer, in my opinion it should have no dependency on the current application. For instance, I would not use a custom AppContext class as JonoW suggested here.
Instead, pass the current user as a parameter to the GetAllUsersExceptForTheCurrentUser method.That way, the service can be used by any application that needs to process users, not only the current application.
You should not create a reverse dependency between your service layer and the web tier. Consider what happens when you want to extend your service layer to work with a forms-based application or windows service. Now you've got to work around the web dependency to get your same methods to work or duplicate some (perhaps, small, but still duplicate) code. You would be better served to extract the user's identifier into something useful in the context of the service layer and use that value with the service layer. Handling the filtering on the web site is also acceptable, though if you do it more than once it would need to be refactored into a common place and the service layer is the natural spot for it.
I find it good practice to build a custom AppContext class which contains my current user object (as well as other contextual data). This class has no references to System.Web. Any service methods that need to be context aware should have an AppContext parameter (e.g. for checking security rights, or getting the current user as in your case). Populate this object in the web-tier and keep it in session if you need to. This way your service layer knows nothing about System.Web.
No. Doing so will make your code harder to test and re-use.
I tend to build an infrastructure interface for this sort of thing (call it IAuthentication or something) and expose a CurrentUser property on it. Then I'd inject this into my service via its a constructor. i.e. public MyService(IAuthentication auth)
Finally you'd can build an HttpContext aware implementation of IAuthentication (say WebAuthentication).
Now when you create your service you create its dependencies as well:
var myService = new MyService(new WebAuthentication());
var otherUsers = myService.GetAllOtherUsers();
If you are using an IoC container the ugly dependency can even go away!