Dagger 2 - Injecting from multiple independently scoped components into a single object - dagger

The issue that I'm facing is that I want to have two independent scopes that don't really fall into a parent-child hierarchy. In my case, I want two types of scopes:
1) "Feature" based scopes. e.g., when a user enters a feature, a scoped component is created. When the user leaves that feature, that scope is destroyed.
2) "Activity" based scopes (this is for an Android app, sorry about the terminology if you don't use Android). When an activity is created, a scoped component is created. When the activity is destroyed, that scope is destroyed.
Subcomponents nor component dependencies work for what I'm after. This is because the feature could end before the activity is destroyed. Similarly, the activity could end before the feature is finished.
I know that I can just use provision methods instead of member injection methods and hold two separate components, but I want the simplicity of being able to just inject all my dependencies in one go into a single object. Does anyone else have any thoughts on this?

This is the best solution I can think of. I felt the same way about coupling my Activity-related objects to other states in my app and implemented something like this.
Say you have three scopes: Application scope, Activity scope, and some Feature scope (eg like a User login state that has their profile info and such). The Activity and Feature scopes are children of the Application scopes, but Feature and Activity are unrelated siblings.
Do not directly inject the Featured scope objects directly into the Activity. Instead, inject a bridge that contains getters to the feature scoped side.
Example:
public interface UserManager {
#Nullable
User getLoggedInUser();
#Nullable
ShoppingCart getShoppingCart();
}
These methods are nullable because the user may or may not be logged in when the other objects attempt to access them. If they are null, the user is not logged in; the feature is not activated.
The implementation of this freely performs the dependency injection from the Feature scoped side since it does not reference anything from the Activity Scope.
public class UserManagerImpl implements UserManager {
#Inject
ShoppingCart cart;
#Inject
User user;
public UserManagerImpl(MyApplication application){
UserScopeComponent component = application.getUserComponent();
if(component != null) {
//only attempt injection if the component exists (user is logged in)
component.inject(this);
}
}
//put simple getters here. They will return null objects if the component didnt inject anything.
}
The ApplicationModule Provides this object. It is not an Application scope or a singleton; you want to initialize it again every time it is injected in case the login state has changed.
#Provides
//No Scope
UserManager provideUserManager(){
return new UserManagerImpl(context);
}
From then on you can inject the UserManager from anywhere in your app, call its getter, check if the output is not null, and away you go. Your Feature state is no longer coupled to the Activity state and you can be free.

Related

Autofac - dynamic Resolve based on registered provider

I'm having problems finding proper solution for my problem, namely:
Let's consider workflow:
Application starts
Main components are registered in Autofac
Application loads plugin assembly and registers modules within it
Container is being build
Plugin handling logic is run
Plugin can add its own controllers. To properly handle that I had to prepare interface which will provide me types of custom controllers:
interface ICustomControllerProvider
{
IEnumerable<Type> GetControllerTypes();
}
Based on the above my app knows how to integrate specified types as controllers.
All controllers are also defined as services, so Autofac deals with their creation, and so...
Problem:
I want to avoid specifying custom controller type twice
public PluginControllerProvider : ICustomControllerProvider
{
public IEnumerable<Type> GetControllerTypes()
{
// 1st type specification
// controller types are specified here, so they could be integrated with app
yield return typeof(ControllerX);
yield return typeof(ControllerY);
}
}
public class PluginModule : Module
{
protected override void Load(ContainerBuilder builder)
    {
builder.RegisterType<PluginControllerProvider>().As<ICustomControllerProvider>();
// 2nd type specification
// controllers have to be register in module as well
builder.RegisterType<ControllerX>();
builder.RegisterType<ControllerY>();
}
}
Is there any way how ControllerX and ControllerY could be managed by Autofac, where I specified them only in PluginControllerProvider?
I tried achieving that by providing custom registration source and resolving ICustomControllerProvider, however I cannot resolve ICustomControllerProvider based on arguments provided by IEnumerable<IComponentRegistration> RegistrationsFor(Service service, Func<Service, IEnumerable<ServiceRegistration>> registrationAccessor) from IRegistrationSource
Autofac does not offer the ability to inject a list of types registered with Autofac. You'll have to get that yourself by working with the lifetime scope registry.
Before I get into how you might solve this, I should probably note:
Listing the types registered is not a normal thing from a DI perspective. It'd be like listing all the controllers in your MVC application - you don't normally need to do that, and if you did, you'd likely need to build a whole metadata structure on top of it like the ApiExplorer that was built to do that on top of ASP.NET Core. That structure wouldn't be supported or involved with the DI system because you're querying about the system, not injecting live instances into the system.
If you are relying on DI to resolve controllers, you probably don't need a whole separate controller provider. Once you know what type you need for the request, you'd just resolve it.
All that's to say, while I'll answer the question, the way the design here is posed may be something you'd want to look at. What you're doing, trying to involve DI with listing metadata about the app... seems somewhat backwards (you'd feed the DI container based on the list of types, not get the list of types from the DI container).
But let's just go with it. Given:
There are controllers registered with Autofac
The controllers have no common base class or interface
There's no attributes on the controllers you could query
What I'd do is register all the controllers with some metadata. You need something to be able to locate the controllers in the list of all the types in the container.
builder.RegisterType<ControllerX>().WithMetadata("controller", true);
builder.RegisterType<ControllerY>().WithMetadata("controller", true);
Now in the plugin controller, you need to inject an ILifetimeScope because you have to query the list of stuff registered. The ILifetimeScope that gets injected into the controller will be the same scope from which the plugin controller itself was resolved.
You can use the injected scope to query things in the component registry tagged with your metadata.
public class PluginControllerProvider : ICustomControllerProvider
{
private readonly Type[] _controllerTypes;
public PluginController(ILifetimeScope scope)
{
_controllerTypes = scope
.ComponentRegistry
.Registrations
.Where(r => r.Metadata.ContainsKey("controller"))
.Select(r => r.Activator.LimitType)
.ToArray();
}
public IEnumerable<Type> GetControllerTypes()
{
return _controllerTypes;
}
}
Now, disclaimers:
If you are registering more controllers in child lifetime scopes (e.g., during a BeginLifetimeScope() call), you will need a controller provider from that scope or it won't get all the controller types. The provider needs to come from the scope that has all the registrations.
If you're using registration sources (like the AnyConcreteTypeNotAlreadyRegisteredSource), this won't capture things that come from the registration sources. It'll only capture things that come from direct registrations of a type (or lambda) on a ContainerBuilder.
But it should work.

How to create a library exposing named configured singletons generated at runtime with Dagger 2?

I'm considering migrating to Dagger 2 some libraries. This library expose a configurable client, each configuration can be named and later retrieved in a singleton fashion.
Let me show a pseudo-code of how the library works from the user perspective:
// initialization
ClientSDK clientA = new ClientSDK.Builder()
.configuration().attributes().here()
.apiKey("someString") // set api key / credentials
.build();
LibraryAuthenticationManager customAuthManager = new MyCustomAuthenticationManager();
ClientSDK clientB = new ClientSDK.Builder()
.configuration().attributes().here()
.apiKey("someStringElse")
.customAuthManager(customAuthManager) // override some default
.baseApiUrl("https://custom.domain.com/and/path") // override some default setting
.build();
ClientSDK.setSingleton("clientA", clientA);
ClientSDK.setSingleton("clientB", clientB);
And when I need an instance elsewhere:
// usage everywhere else
ClientSDK clientB = ClientSDK.singleton("clientB");
clientB.userManager(); // "singleton" using the configuration of clientB
clientB.subscriptionsManager(); // "singleton" using the configuration of clientB
clientB.currentCachedUser(); // for clientB
clientB.doSomething(); // action on this instance of the ClientSDK
ClientSDK instances are created by the user of the library and the ClientSDK statically keep a map of singletons associated to the name.
(The actual behavior of the SDK is slightly different: the naming is automatic and based on a mandatory configuration parameter.)
It's like I have lot of singleton classes with a single point of entry (the ClientSDK) but since I can have multiple configuration of the ClientSDK each with his own singletons instances this are not really singletons.
If I would try write a library like that with Dagger 2 I would do something like:
class ClientSDK {
#Inject SDKConfiguration configuration;
#Inject LibraryAuthenticationManager authManager;
...
}
The problem is that I need each instance of the ClientSDK to have its own configuration and authManager (and many other services) injected. And they need to be definable (the configuration) and overridable (the actual implementation) from the library user.
Can I do something like this with Dagger 2? How?
I've seen I can create custom Scopes but they are defined at compile time and the library user should be the one defining them.
(the library is an Android Library, but this shouldn't be relevant)
Thanks
It sounds like you should be creating stateful/configurable Module instances and then generating separate Components or Subcomponents for each ClientSDK you build.
public class ClientSDK {
#Inject SDKConfiguration configuration;
#Inject LibraryAuthenticationManager authManager;
// ...
public static class Builder {
// ...
public ClientSDK build() {
return DaggerClientSDKComponent.builder()
.configurationModule(new ConfigurationModule(
apiKey, customAuthManager, baseApiUrl)
.build()
.getClientSdk();
}
}
}
...where your ConfigurationModule is a #Module you write that takes all of those configuration parameters and makes them accessible through properly-qualified #Provides methods, your ClientSDKComponent is a #Component you define that refers to the ConfigurationModule (among others) and defines a #Component.Builder inner interface. The Builder is important because you're telling Dagger it can no longer use its modules statically, or through instances it creates itself: You have to call a constructor or otherwise procure an instance, which the Component can then consume to provide instances.
Dagger won't get into the business of saving your named singletons, but it doesn't need to: you can save them yourself in a static Map, or save the ClientSDKComponent instance as an entry point. For that matter, if you're comfortable letting go of some of the control of ClientSDK, you could even make ClientSDK itself the Component; however, I'd advise against it, because you'll have less control of the static methods you want, and will lose the opportunity to write arbitrary methods or throw exceptions as needed.
You don't have to worry yourself about scopes, unless you want to: Dagger 2 tracks scope lifetime via component instance lifetime, so scopes are very easy to add for clarity but are not strictly necessary if you're comfortable with "unscoped" objects. If you have an object graph of true singleton objects, you can also store that component as a conventional (static final field) singleton and generate your ClientSDKComponent as a subcomponent of that longer-lived component. If it's important to your build dependency graph, you can also phrase it the other way, and have your ClientSDKComponent as a standalone component that depends on another #Component.

Difference between #Singleton & static #Provides in dagger2

May I know the difference between #Singleton VS static Provides in dagger2?
#Provides static User currentUser(AuthManager authManager) {
return authManager.currentUser();
}
#Provides #Singleton User currentUser(AuthManager authManager) {
return authManager.currentUser();
}
These are very different attributes, and you can have one or the other independently. All of these are valid:
#Provides User currentUser(...) {}
#Provides static User currentUser(...) {}
#Provides #Singleton User currentUser(...) {}
#Provides #Singleton static User currentUser(...) {}
To set the stage, a #Provides User method says "for this Component or its dependencies, call this #Provides method every time you need a User". Typically the method will return a new instance every time, and Dagger won't save or cache the instance.
#Singleton is an example of a scope, which is a fancy way to say lifecycle policy or policy for how often to create a new instance. #Provides #Singleton User says "for this Component or dependencies, just call this #Provides method once, and save the result". #Singleton happens to be a built-in common case, but you could also imagine creating a #UserScope (always return the same instance for this User), or in Android a #FragmentScope or #ActivityScope.
For your specific case, you probably don't want #Singleton, because it would instruct your component to save or cache the value from AuthManager. If the User value may change across your application's lifetime, the Component wouldn't reflect that. (In that case you would also want to make sure to inject Provider<User>, which would update, rather than User which would not.)
Leaving scopes behind for a moment, static behaves exactly the way you would expect it to in Java: If a method doesn't require any instance state, you can make it static, and your virtual machine can call it without preparing any instance state. In your generated Component implementation, Dagger will automatically call static methods statically, and instance methods on the Module instance you pass into your Component; in Android this results in a sizable performance increase. Because you don't use any instance state in your currentUser method, it can easily be made static.
Further reading:
SO: Scopes in Dagger 2
Dagger docs: Component (see heading "Scope")
With #Singleton annotation only one instance of User object will be created throughout the application lifecycle.
static on #Provides methods introduced recently to make the invocation of method faster by 15 to 20% as mentioned here. There will be multiple instances of User object if we call this method multiple times.

GWT MVP updating Activity state on Place change

What is the best practise to update Activity state on Place change? Imagine you have an activity with view that displays list of categories and list of items in the category. If different category is selected then app goes to new place with category ID. I want then to only refresh items and not to create new activity that also re-reads category list.
My current approach is like this:
public class AppActivityMapper implements ActivityMapper {
private ItemListActivity itemListActivity;
...
public Activity getActivity(final Place place) {
final Activity activity;
if (place instanceof ItemListPlace) {
if (itemListActivity == null) {
itemListActivity = new ItemListActivity((ItemListPlace) place, clientFactory);
} else {
itemListActivity.refresh((ItemListPlace) place);
}
activity = itemListActivity;
} else {
itemListActivity = null;
}
...
return activity;
}
...
Alternatives are:
listen to PlaceChangeEvents from within the activity (you can then use a FilteredActivityMapper and CachingActivityMapper for the caching of the activity in your ActivityMapper, so that it's reduced to only create a new activity when asked). †
have some component listen to PlaceChangeEvents and translate them to business-oriented events, the activity then listens to those events rather than PlaceChangeEvents, otherwise the same as above.
decouple the activity from the "screen", make the "screen" a singleton with a reset() method and call that method from the activity's start (possibly passing the category ID as an argument in this case). The "screen" being a singleton could then make sure to load the categories list only once.
in your case, you could also simply put the categories list in a shared cache, so that you don't have to reuse your activity by can create a new one, the categories list will be retrieved once and put in the cache, subsequent activity instances will just use what's in the cache. This is similar to the above, but simpler, and the cache could be used by other parts of the application.
I'd personally rather go with your approach though (with a small exception, see below), as it's the simplest/easiest. Decoupling the activity from the "screen" is also an option; the GWT Team started exploring this approach in the Expenses sample (decoupling the activity responsibility from the presenter responsibility with using MVP) without ever finishing it unfortunately.
Other than that, I don't think any best practice has really emerged for now.
†. I don't like coupling my activities with the places they're used with (I don't quite like the coupling for the goTo calls either, but haven't yet found a clean and simple alternative), so I'd rather not go with this option; and similarly, I'd not pass the place to the activity constructor and refresh method like you did, but rather extract the information out of the place and pass it to the activity (e.g. in your case, only give the category ID to the activity, not the ItemListPlace instance; I would then simply call setCategory in all cases, and not even pass the category ID to the constructor).
In my opinion,
The role of the ActivityMapper is to give you back an Activity from a Place.
The role of the ActivityManager is to start the Activity given back from the ActivityMapper and to stop the current one if different. In your case you would like to "update/refresh" the current Activity.
So I would modify the ActivityMapper so as it will allways give me back the same instance of Activity for a given type of Place. A good way to do so could be to use GIN and use the singleton scope ...in(Singleton.class) to inject your Activity.
If you do that, when changing the url, if the place stays the same (meaning your url has the same word after # and before :) so that the Type of your place stays the same, the ActivityMapper will give you back the same instance of Activity so the ActivityManager will do nothing on the Activity. Check l.126 of ActivityManager
if (currentActivity.equals(nextActivity)) {
return;
}
For me you have 2 options there. The first one, as Thomas said , is to listen to PlaceChangeEvent in your Activity. The new Place you will receive can have new parameters inside based on the new url given and you can "update/refresh" your Activity.
The second one, that I find more in line with the Activity/Place pattern is to modify the ActivityManager so that it calls an update(Place) method on the Activity when the Activity given back by the ActivityMapper is the same that the current Activity.
I haven't tried any of these solutions yet but I will soon ... I might be able to update that post at that time.
You can find more information in this article I wrote on my blog on that topic
Here is a little schema I made to help me understand the pattern, hope it will help :
I would not do any logic in my ActiviyMapper except returning an activity, by creating a new one or giving a previous one (or null). According to me, the mapper doesn't have to know about refresh() or what activities do.
If that, then the logic of 'refresh()' would be given to the activy through the place which holds a token. That token should be holding the information about either what is the state of the request (a new page, reload, an id, etc).
In the activity, first, it asks for the View, the one related to this activity (tip : a singleton given by a 'ClientFactory' is good practice), then it creates a presenter for that view, and bind them together.
Lastly, the activity will use the token from the place to provide any information about state to the presenter. And then, it adds the view in the page.
It's good to know by default, with places and activies, going to the same place doesn't do anything (no reload). But you can take care of it with token and activity-mapper easily.
Hope you'll find an adapted solution for you case. Goodluck.

Autofac Session Scope

I am investigating the use of Autofac in our web application having previously used Castle Windsor in the past.
The thing that I really like with Autofac is being able to express dynamic component construction through lamda expressions, as opposed to creating DependancyResolvers etc. in Windsor.
One scenario I have is that I want a particular component to be registered at ASP.NET session level scope. With Windsor I would create/source a new LifestyleManager, however with Autofac I came up with this:
//Register SessionContext at HTTP Session Level
builder.Register(c =>
{
HttpContext current = HttpContext.Current;
//HttpContext handes delivering the correct session
Pelagon.Violet.Core.Interfaces.SessionContext instance = current.Session["SessionContext"] as Pelagon.Violet.Core.Interfaces.SessionContext;
if (instance == null)
{
instance = c.Resolve<Pelagon.Violet.Core.Interfaces.SessionContext>();
current.Session["SessionContext"] = instance;
}
return instance;
})
.FactoryScoped();
Which at some point I might be able to turn into an extension method. I accept this implemtation will bomb if the HttpContext.Current.Session is null as it should only be used in a web app.
The question is:
What is the best practice for such a registration in Autofac. I have seen a lot of mention about the use of nested containers etc. but no concrete examples, and I am keen to understand what might be wrong with the above approach (only thing I can think of is the automatic disposal stuff).
Thanks.
This looks fine.
Marking the component 'ExternallyOwned()' will ensure that Autofac doesn't call Dispose() on it.
The only gotchas here are that your session-scoped component could resolve dependencies of its own via the current container, and thus hold references to things that may belong to the current request (for instance.) This should be easy to spot in testing though.