I am trying to create an IOC container in Castle Windsor that's configuration is shared across assemblies.
(What follows is an example of how this works in Unity. What I want to do is to make it work the same way using Castle Windsor)
I have the following project configuration...
TestCompany.Services.Host
(Web project hosting a number of .svc files)
PrintService.svc
Web.Config
Unity.Config
TestCompany.Services.PrintService
IPrintService.cs
PrintService.cs
The actual implementation of my "PrintService" is not implemented inside my Services.Host but in the TestCompany.Services.PrintService assembly.
As part of my shared project code (not shown) I have a container helper which is responsible for loading the unity configuration...
public static IUnityContainer GetContainer()
{
// Checks for existance of container (_container == null) ommitted.
var section = ConfigurationManager.GetSection("unity") as UnityConfigurationSection;
section.Configure(_container, name);
...
...
}
This method loads the unity configuration section from the Unity.Config and uses it to configure the container.
The advantage of this method is that one Unity.Config loaded inside (I presume) the AppDomain can service a number of assemblies. Simply calling GetContainer() from any of the assemblies consumed by my service host will return a container populated with the same type resolution's etc.
I really want to use the fluent configuration in Castle Windsor but I dont see how without this "shared" configuration file that can be acheived. PrintService and any future services will all need to resolve the same dependencies and I dont want to have to repeat my fluent configuration between these services.
Ideally I need some sort of container configured in the service host app that can "flow" into all of the assemblies that it makes use of.
Thanks.
I think I may not be understanding your question but I think I understand your scenario and here is how I do something similar, if it helps at all...
My Philosophy:
Each part of the application should be in charge of registering what
it knows about and nothing more, so there is no need for a single
central configuration file and things that are shared between
components are registered in one place and their interfaces are
available everywhere via a common library.
So let's take an example...
First of all, let us just say (for the purposes of my example) that IPrintService is something that you want to register an implementation of once and use throughout the application and that we have some other component that needs to be implemented by some external module from the main application. We, therefore, create an assembly called Common like so:
Common
public interface IPrintService
{
void Print();
}
public interface IMyService
{
void DoSomething();
}
Now let us think about the main part of the application (maybe it is an ASP .NET application, maybe justa console application, does not really matter). Here we construct the container and ask it to find all the possible components. We can do that like so:
Main Application
// Could be the Global.asax code behind but for simplicity this is
// just a console application
class Program
{
private static readonly IWindsorContainer Mycontainer
= BootstrapContainer();
// Allow access to the raw container - this is probably a bad idea but
// in the rare case that you need it you can get it from here
public static IWindsorContainer Container { get { return Mycontainer; } }
private static IWindsorContainer BootstrapContainer()
{
// Here we will just install every IWindsorInstaller found in any
// assembly in the same folder as the application (so no need for
// references or anything).
var c = new WindsorContainer();
string folder = Path.GetDirectoryName(
Assembly.GetExecutingAssembly().Location);
c.Install(FromAssembly.InDirectory(new AssemblyFilter(folder)));
return c;
}
}
// Here is the print service implementation
public class MyPrintService : IPrintService
{
public void Print()
{
// Print!
}
}
// This is the installer for the main module - here we are saying exactly
// what is implementing the interface
public class MainApplicationInstaller : IWindsorInstaller
{
public void Install(IWindsorContainer container,
IConfigurationStore store)
{
container
.Register(Component
.For<IPrintService>()
.ImplementedBy<MyPrintService>());
}
}
So now we have a common library with our shared inetrfaces and a main application that will register an implementation for our shared interface and also load up any other modules in the system.
The only thing, therefore, left to do is to consume that print service and use it. We can do this anywhere that is using the container so let's create a third assembly that references only Common (we will call it test module.
Test Module
// This installer installs just the things inside this module since that
// is all it knows about but those things can use things that are
// registered in the container by anybody.
public class TestModuleInstaller : IWindsorInstaller
{
public void Install(IWindsorContainer container,
IConfigurationStore store)
{
container
.Register(Component
.For<IMyService>()
.ImplementedBy<MyServiceThatDoesSomething>());
}
}
public class MyServiceThatDoesSomething : IMyService
{
private readonly IPrintService _printService;
public MyServiceThatDoesSomething(IPrintService printService)
{
_printService = printService;
}
public void DoSomething()
{
// Use the print service!
_printService.Print();
}
}
Finally compile everything and copy the test module to the same folder as the main application and then from the main you can do this:
Container.Resolve<IMyService>().DoSomething();
And then the magic happens! Well, some code runs and you find that the print service is called by the class from the module even though it knows nothing about it.
Anyway, maybe that helps a little bit, maybe not, good luck!
Related
I have the following :
a class library with connection classes such as connection, command, parameter
a DAL with entities, mapper, interface, services as well as a static class that holds hard coded connectionString and InvariantName.
an Asp.Net Core project
References :
DAL has a reference to the class library to make use of its connection class to which it provides connectionString and InvariantName thanks to its static class etc..
Asp.Net has a reference to the DAL.
What I want :
I now want to use the User Secrets to store hard coded sensitive data connections and get rid off the static class.
I know I can use the the Asp.Net Core startup.cs to read the settings from Configuration and make use of binding to store them into a class and use DI.
My guess :
DI seems "easy" when used inside an Asp controller. But I need the settings values (connectionString and InvariantName) outside the Asp.Net Core to be injected into a constructor of a class somewhere in my DAL.
I guess I would then need to have to reference the Asp.Net Core project to my DAL. But then I would end up with a circular reference (DAL to Asp.Net Core and the opposite).
So what's the solution?
Have an intermediate library class into which I would retreive the settings values from Asp.Net Core and then pass them to my DAL (to prevent circular reference)?
Manually recreate the "Configuration process" inside the DAL and get settings there directly
Or something else that I don't know?
Ps : I am new in development and only have a few projects'experience in Asp.Net Framework so far..and it's my first Asp.Net Core project
I know I can use the the Asp.Net Core startup.cs to read the settings from Configuration and make use of binding to store them into a class and use DI
You already answered your own question with this. This is the correct and recommended behavior to setup DI for 3rd party libs and configurations. If you want to avoid clutter in Startup class, create an extension method:
namespace Microsoft.Extensions.DependencyInjetion
{
public static MyLibraryCollectionExtensions
{
public static IServiceCollection AddMyLibrary(this IServiceCollection services)
{
services.AddDbContext<MyDbContext>(...);
}
}
}
to register your classes. Alternatively, extend the method to accept a parameter delegate to configure it
namespace Microsoft.Extensions.DependencyInjetion
{
public static MyLibraryCollectionExtensions
{
public static IServiceCollection AddMyLibrary(this IServiceCollection services, Action<MyOptions> setup)
{
var defaultOptions = ... // i.e. new MyOptions();
// pass default options to be modified by the delegate
setup?.Invoke(defaultOptions);
// your registrations
services.AddDbContext<MyDbContext>(...);
}
}
}
And all the user has to do in your library is add
services.AddMyLibrary();
// or with setup
services.AddMyLibrary(config =>
{
config.MyConnectionString = Configuration.GetConnectionString("MyContext");
});
and store the connection string in the appsettings.json.
{
"ConnectionStrings":
{
"MyContext" : "MyConnectionString here"
}
}
I finally used the ConfigurationBuilder to get values from the appsettings.json file.
It's probably not the right way to do it but it is working with my DAL and Connection dlls.
In case it helps anyone else :
I have .net core console application, which is hosted as windows service.
I want to catch an event if the user logs off/shutdown the computer.
I have found ways to catch this event in .net framework (here & here).
But I cant figure out how to achieve this in .net core.
To create service I am using "ServiceBase" class. Sample code is as given below:
public class MyService : ServiceBase
{
readonly string LogPath = "D:\\TestAppService.txt";
#region Constructors
public MyService()
{
this.CanShutdown = true;
}
#endregion
#region Protected Functions
protected override void OnStart(string[] args)
{
//your code here
// call the base class so it has a chance
// to perform any work it needs to
base.OnStart(args);
}
protected override void OnStop()
{
//your code here
// Call the base class
base.OnStop();
}
protected override void OnShutdown()
{
using (StreamWriter sw = File.AppendText(LogPath))
{
sw.WriteLine("shutdown == true");
}
//your code here
base.OnShutdown();
}
#endregion
}
The OnStop and OnStart methods are being called.
but when I shutdown the computer my OnShutdown method is not called.
According to aspisof.net, you should be able to use the SessionEnding API. This is because it is listed as being exposed in the windows Compatibility Pack - available on NuGet here.
This article on learn.microsoft.com shows how you can include it in a .NET Core application.
tl;dr
Add the NuGet package
Target Windows only
One thing to note: this was originally designed to be a temporary fix for porting Windows specific .NET code over to .NET Core.
The more accepted way to implement Windows only features is to move as much code to .NET Standard libraries as possible, and to use conditional compilation directives to include platform specific code when building for that platform.
By design dotnet core is not "friendly" with platform specific stuff
(like listening to log off event seems to me).
The solution I use in one of Windows-hosted services is described here.
When application domain is forced to close by operating system on shutdown - there is a room for using AppDomain event handlers.
I have MVC4 website project and WCF project, both using Ninject.
I want to use class from WCF project in website project. I add reference to project and get both NinjectWebCommon.Start() executing (with "The static container already has a kernel associated with it!" error).
Is there way to make what I want?
Solved this using this startup in referenced project
public class Global : NinjectHttpApplication
{
protected override IKernel CreateKernel()
{
return new StandardKernel(new ServiceModule());
}
}
I have a WCF web service hosted on my Local IIS (not Express). I've included a Global.asax in its root directory, where it is supposed to be. Since I'm using Ninject with WCF extensions, the class Global extends NinjectHttpApplication instead of HttpApplication (as seen here).
Also, I'm using the AutoMapper library in order to circumvent writing boring boilerplate code.
The problem arises because a static method I defined for configuring AutoMapper isn't being called causing AutoMapper to throw exceptions when I call Mapper.Map(). That static method's call is defined in Global.asax's Application_Start() method since I want these mappings to be performed once per the web service's lifetime.
Ninject's CreateKernel() method gets called just fine, by the way.
Am I missing something here? I've tried debugging it, it doesn't hit the breakpoint even though I've attached the debugger to w3wp.exe and also tried putting an explicit Debugger.Break() call in its body.
This is how it looks like so far:
Global.asax
<%# Application Codebehind="Global.asax.cs" Inherits="MyApp.WebHost.Global" Language="C#" %>
Global.asax.cs
public class Global : NinjectHttpApplication
{
protected override IKernel CreateKernel()
{
IKernel kernel = new StandardKernel();
/* various bindings */
return kernel;
}
protected void Application_Start(object sender, EventArgs e)
{
AutoMapperConfig.RegisterMappings();
}
/* rest of Global.asax methods (Session_Start, Application_BeginRequest, etc.) with empty bodies */
RegisterMappings method
public static class AutoMapperConfig
{
public static void RegisterMappings()
{
/* multiple calls to Mapper.CreateMap() */
Mapper.AssertConfigurationIsValid();
}
}
Svc file markup
<%# ServiceHost Language="C#"
Debug="true"
Service="MyApp.Services.MyAppService"
Factory="Ninject.Extensions.Wcf.NinjectServiceHostFactory" %>
Everything else works, I've already created a test client (a simple console app) and added a service reference. Service methods get called just fine, it is just that these mappings are a bit problematic since AutoMapper keeps throwing AutoMapperMappingException exceptions ("Missing type map configuration or unsupported mapping.") for the obvious reasons.
The application's app pool is DefaultAppPool. Should I create a separate one?
I really don't understand the problem here. Thank you in advance.
Well, it required some additional searching but I found the answer here - https://groups.google.com/forum/#!topic/ninject/wRy3ELSV4bU
The problem was that NinjectHttpApplication class itself implements the Application_Startup method so it is impossible to implement it in your own derived class (Global class).
To simulate such behavior one needs to override the OnApplicationStarted Ninject's method.
This is how it looks like regarding my particular problem:
protected override void OnApplicationStarted()
{
AutoMapperConfig.RegisterMappings();
}
I'm trying to do some testing with castle windsor involved, in one of my tests I want to check the windsor installers, so I check that the container can resolve my components given its interface.
So far, so good, the problem starts when the component has PerWebRequest lifestyle in its installer, at first it complained about HttpContext.Current is null, having that one solved creating a fake Context in test setup I'm now having this exception in nunit test
System.Exception : Looks like you forgot to register the http module Castle.MicroKernel.Lifestyle.PerWebRequestLifestyleModule
Add '' to the section on your web.config. If you're running IIS7 in Integrated Mode you will need to add it to section under
As I'm running this from NUnit, how I can register the module or class in windsor so it works, or how can be mocked, as in this test is not really a web request, just checking that the container resolve the type.
And also this same thing will happen if I make any integration tests with this component outside a real webrequest, is there any way to make this work or really mock a web request so this tests can be run?
Tranks in advance
Fer
In your test you could subscribe to the ComponentModelCreated event and change the lifestyle of your per-web-request components to something else. (example).
If you're writing an integration test with the scope of a single request, singleton should do.
If you're writing an integration test that spans multiple requests, you could use a contextual lifestyle to simulate the scope of requests.
Edit: including code from example (which is no longer available):
container.Kernel.ComponentModelCreated += Kernel_ComponentModelCreated;
…
void Kernel_ComponentModelCreated(Castle.Core.ComponentModel model)
{
if (model.LifestyleType == LifestyleType.Undefined)
model.LifestyleType = LifestyleType.Transient;
}
From version 5 of Windsor the accepted answer doesn't work if you are using Castle.Facilities.AspNet.SystemWeb.WebRequestScopeAccessor because the PerWebRequest lifestyle is already a scoped lifestyle.
I got it to work by changing the the ComponentModelCreated delegate to the following:
void Kernel_ComponentModelCreated(Castle.Core.ComponentModel model)
{
const string CastleScopeAccessorType = "castle.scope-accessor-type";
if (model.ExtendedProperties.Contains(CastleScopeAccessorType))
{
model.ExtendedProperties.Remove(CastleScopeAccessorType);
}
}
I ended up implementing this extension. ATTN: Must call before loading components with the PerWebRequest lifestyle:
public static class WindsorContainerExtensions
{
public static IWindsorContainer OverridePerWebRequestLifestyle(this IWindsorContainer container)
{
container.Kernel.ComponentModelCreated += model =>
{
if (model.IsPerWebRequestLifestyle())
{
model.LifestyleType = LifestyleType.Transient;
}
};
return container;
}
private static bool IsPerWebRequestLifestyle(this ComponentModel model)
{
return model.LifestyleType == LifestyleType.Scoped
&& model.HasAccessorType(typeof(WebRequestScopeAccessor));
}
private static bool HasAccessorType(this ComponentModel model, Type type)
=> model.HasExtendedProperty("castle.scope-accessor-type", type);
private static bool HasExtendedProperty<T>(this ComponentModel model, object key, T expected)
{
return model.ExtendedProperties[key] is T actual
&& EqualityComparer<T>.Default.Equals(actual, expected);
}
}
Requires these imports:
using System;
using System.Collections.Generic;
using Castle.Core;
using Castle.Facilities.AspNet.SystemWeb;
using Castle.Windsor;
If you also want to check if the type of scope is per web request you could also do this
var isPerWebRequestScope = JsonConvert.SerializeObject(model.ExtendedProperties).Contains("Castle.Facilities.AspNet.SystemWeb.WebRequestScopeAccessor")