I have an application designed using Microsoft's Composite Application Library. My shell has several regions defined so that I can inject content from separate modules. I'm looking for a design pattern that will reduce the coupling that these regions introduce.
In all examples I have seen, regions are defined and accessed using a string in a static class in the infrastructure project.:
<ItemsControl cal:RegionManager.RegionName="{x:Static inf:RegionNames.TabRegion}">
public static class RegionNames
{
public const string TabRegion = "TabRegion";
}
This introduces an dependency on the shell from the infrastructure project, because part of the infrastructure project must now match the shell. The CAL RegionManager throws an exception if you attempt to access a region which is not defined, so I must ensure that the infrastructure and shell projects are kept in sync.
Is there a way to isolate the shell's regions so that they are defined only within the shell (no region names in the infrastructure project)?
Is there a way to make regions optional, so that shells can be swapped out even if they don't have all the same regions? (An example: One shell has menu and toolbar regions, another only has the menu... modules should be able to inject into the toolbar if it's available, without failing when it's not)
Update - More details on my architecture
In response to depictureboy's answer below, I wanted to describe the way my system is set up... perhaps there will be more good feedback on it.
I am treating the Infrastructure and Shell projects as generic libraries, and I have several applications which use them. The Infrastructure project provides "framework" code and resources (like MVVM stuff, reflection, icons), and my Shell is a generic host window, with the basic window layout (menus, toolbars, status bar, main content area). The applications all share a common look and behave similarly because they share the shell.
My applications get their individual functionality from the modules which get loaded, so I have a bootstrapper project per application which pulls everything together (infra, shell, modules).
I imagine if I ever need to develop a brand new application that is very different from the current ones, I will be able to re-use the infrastructure project, but not the shell. That is why I am curious about decoupling the infrastructure project and the shell.
I think you have your logic backwards. Your shell is the glue that binds everything together. In my mind you want the infrastructure and shell tightly coupled because they are the application. Your modules are the parts of the application that will be changing and switching around dynamically. You want your shell regions to be static so that, for example, another developer could write a module for your application knowing where his different views were going to be placed and how the application should behave with his module attached. The Infrastructure project is there to be the go between between your shell and its modules...thats just a fact of life at least in my book. One of the WPF gurus may come up with something that absolutely blows that out of the water tomorrow....
Related
I'm very new to modules in Prism and am struggling to understand how to use them properly. I want to have a navigation panel like in the image below, which I stole from Google.
Let's say that I have 3 modules for 3 separate concerns: NavigationModule, HelpModule, and FeedbackModule. If I want to show "Help" when the app first loads, my understanding is that I would do something like the following:
Define Regions for navigation and content in my Shell (main WPF project)
Upon execution, load the navigation view into the navigation Region
Upon initialization of the NavigationModule, load a view from HelpModule into the main content Region
This leaves me with a few questions about what modules should know about each other:
Should the Shell's project have a reference to the NavigationModule in order to load its view?
Should the NavigationModule have a reference to the HelpModule in order to load its view? And a reference to FeedbackModule in order to load its views on demand?
If the answer to these questions is "no," then what's the best way for modules to be aware of each other's views? I could create a shared class library with constants for view names, but it seems a bit troublesome to maintain a bunch of strings that way whereas with references I could use nameof(). I would appreciate any direction. Thanks.
Modules should not "know" each other in the sense of having a project reference from one module to another, because that kind of defeats anything gained from having modules in the first place, that is, to have a modular application (built of components that can be swapped out independent of each other).
Modules should interact through shared interfaces, which are define outside of modules, that is either in the framework or in assemblies that "are" no modules themselves. Depending on your requirements, you should define upfront (and enforce in the build process) which modules there are and which interface-assemblies and which module is allowed to reference which interface-assembly.
So how to show the Help-view when the application loads? Send a message (e.g. via IEventAggregator or any other communication mechanism) when it's time to load the initial view (whatever that may be). The Help-module listens for the message and navigates to the Help-view. Important: the SessionResume-module might also listen for our message. It's the responsibility of whomever choses which modules to deploy to only deploy modules that are compatible with each other.
I'm new to .Net and trying to learn things. I'm trying to develop a Prism4 WPF app with
Visual Studio CSharp 2010 Express Edition,
Prism v4,
Unity as IoC,
SQL Server CE as data store.
I've studied a lot(?) and infuenced by this and this among others, and decided to implement MVVM, Repository and UnitofWork patterns. This application will be a Desktop application with a single user (me:-)
So, I've created a solution with the following projects:
Shell (Application Layout and startup logic)
Common (Application Infrastructure and Workflow Logic)
BusinessModuleA (Views and ViewModels)
BusinessModuleA.Model (Business Entities - POCO)
BusinessModuleA.Data (Repositories, Data Access (EF?) )
BusinessModuleB (Views and ViewModels)
BusinessModuleB.Model (Business Entities - POCO)
BusinessModuleB.Data (Repositories, Data Access (EF?) )
My questions are:
Which projects should reference which projects ?
If I implement Repositories in 'BusinessModuleX.Data', which is
obvious, where should I define IRepositories ?
Where should I define IUnitOfWork and where should I implement UnitOfWork ?
Is it ok if I consume UnitOfWork and Repositories in my ViewModels ?
Instict says it is bad design.
If (4) above is bad, then ViewModel should get data via a Service
Layer (another project ?). Then, how can we track changes to the
entities so as to call the relevant CRUD methods on those objects at the Service Layer?
Is any of this making any sense or am I missing the big picture ?
Ok, may be I've not made myself clear on what I wanted exactly in my first post. There are not many answers coming up. I'm still looking for answers because while what #Rachel suggested may be effective for the immediate requirements, I want to be careful not to paint myself into a corner. I've an Access Db that I developed for my personal use at Office, and which became kind of a success and now being used by 50+ users and growing. Maintaining and modifying the access code base has been fairly simple at the beginning, but as the app evolved, began to fall apart. That's why I have chosen to re-write everything in .Net/Wpf/Prism and want to make sure that I get the basic design right.
Please discuss.
Meanwhile, I came up with this...
First off, I would simplify your project list a bit to just Shell, Common, ModuleA, and ModuleB. Inside each Project I'd have sub-folders to specify where everything is. For example, ModuleA might be separated into folders for Views, ViewModels, and Models
I would put all interfaces and global shared objects such as IUnitOfWork in your Common project, since it will be used by all modules.
How you implement IUnitOfWork and your Repositories probably depends on what your Modules are.
If your entire application links to one database, or shares database objects, then I would probably create two more projects for the DataAccessLayer. One would contain public interfaces/classes that can be used by your modules, and the other would contain the actual implementation of the Data Access Layer, such as Entity Framework.
If each Module has it's own database, or its own set of objects in the database (ie. Customer objects don't exist unless you have the Customer Module installed), then I would implement IUnitOfWork in the modules and have them handle their own data access. I would probably still have some generic interfaces in the Common library for the modules to build from though.
Ideally, all your modules and your Shell can access the Common library. Modules should not access each other unless they build on them. For example, a Customer Statistics module that builds on the base Customer module should access the Customer module.
As for if your ViewModels should consume a UnitOfWork orRepository, I would have them use a Repository only. Ideally your Repository should be like a black box - ViewModels can Get/Save data using the Repository, but should have no idea how it's implemented. Repositories can get the data from a service, entity framework, direct data access, or wherever, and the ViewModel won't care.
I'm no expert on design architecture, however that's how I'd build it :)
I would highly recommend you to get the Introduction to PRISM and Repository pattern inside Design Patterns Library training videos. They are great ones. Hope it helps
Can anyone give a high level description of what is going on in the Composite C1 core? In particular I am interested in knowing how the plugin architecture works and what the core components are of the system i.e. when a request arrives what is happening in the architecture. The description doesn't have to be too verbose just a list of steps and the classes involved.
Hopefully one of the core development team would enlighten me... and maybe publish some more API (hint hint more class documentation please).
From request to rendered page
The concrete path a request takes depends on the version of C1 you're using, since it was changed to use Routing in version 2.1.2. So lets see
< 2.1.2
Composite.Core.WebClient.Renderings.RequestInterceptorHttpModule will intercept all incoming requests and figure out if the requested path correspond to a valid C1 page. If it does, the url will be rewritten to the C1 page handler ~/Rendererings/Page.aspx
2.1.1
Composite.Core.Routing.Routes.Register() adds a C1 page route (Composite.Core.Routing.Pages.C1PageRoute) to the Routes-collection that looks at the incoming path, figures out if its a valid C1 page. If it is, it returns an instance of ~/Rendererings/Page.aspx ready to be executed.
Okay, so now we have an instance of a IHttpHandler ready to make up the page to be returned to the client. The actual code for the IHttpHandler is easy to see since its located in ~/Renderers/Page.aspx.cs.
OnPreInit
Here we're figuring out which Page Id and which language that was requested and looking at whether we're in preview mode or not, which datascope etc.
OnInit
Now we're fetching the content from each Content Placeholder of our page, and excuting its functions it may contain. Its done by calling Composite.Core.WebClient.Renderings.Page.PageRenderer.Render passing the current page and our placeholders. Internally it will call the method ExecuteFunctions which will run through the content and recursively resolve C1 function elements (<f:function />), execute them and replace the element with the functions output. This will be done until there are no more function elements in the content in case functions them selves output other functions.
Now the whole content is wrapped in a Asp.Net WebForms control, and inserted into our WebForms page. Since C1 functions can return WebForms controls like UserControl etc., this is necessary for them to work correctly and trigger the Event Lifecycle of WebForms.
And, that's basically it. Rendering of a requested page is very simple and very extendable. For instance is there an extension that enables the usage of MasterPages which simply hooks into this rendering flow very elegantly. And because we're using Routing to map which handler to use, its also possible to forget about ~/Rendering/Page.aspx and just return a MvcHandler if your a Mvc fanatic.
API
Now, when it comes to the more core API's there are many, depending on what you want to do. But you can be pretty sure, no matter what there is the necessary ones to get the job done.
At the deep end we have the Data Layer which most other API's and facades are centered around. This means you can do most things working with the raw data, instead of going through facades all the time. This is possible since most configuration of C1 is done by using its own data layer to store configuration.
The Composite C1 core group have yet to validate/refactor and document all the API's in the system and hence operate with the concept of 'a public API' and what can become an API when the demand is there. The latter is a pretty darn stable API, but without guarantees.
The public API documentation is online at http://api.composite.net/
Functions
Functions is a fundamental part of C1 and is a technique to abstract logic from execution. Basically everything that either performs a action or returns some data/string/values can be candidates for functions. At the lowest level a function is a .Net class implementing the IFunction interface, but luckily there are many easier ways to work with it. Out of the box C1 supports functions defined as XSLT templates, C# methods or Sql. There are also community support for writing functions using Razor or having ASP.Net UserControls (.ascx files) to be functions.
Since all functions are registered in C1 during system startup, we use the Composite.Functions.FunctionFacade to execute whatever function we know the name of. Use the GetFunction to get a reference to a function, and then Execute to execute it and get a return value. Functions can take parameters which are passed as real .Net objects when executing a function. There is also full support for calling functions with Xml markup using the <f:function /> element, meaning that editors, designers, template makers etc. easily can access a wealth of functionality without having to know how to write .Net code.
Read more about functions here http://users.composite.net/C1/Functions.aspx and how to use ie Razor to make functions here http://docs.composite.net/C1/ASP-NET/Razor-Functions.aspx
Globalization and Localization
C1 has full multi-language support in the core. Composite.Core.Localization.LocalizationFacade is used for managing the installed locales in the system; querying, adding and removing. Locales can be whatever CultureInfo object is known by your system.
Composite.Core.ResourceSystem.StringResourceSystemFacade is used for getting strings at runtime that matches the CultureInfo your request is running in. Use this, instead of hardcoding strings on your pages or in your templates.
Read more about Localization here http://docs.composite.net/C1/HTML/C1-Localization.aspx
Global events
Composite.C1Console.Events.GlobalEventSystemFacade is important to know if you need to keep track on when the system is shutting down so you can make last-minute changes. Since C1 is highly multithreaded its easy to write extensions and modules for C1 that are multithreaded as well, taking advantage of multi core systems and parallelization and therefor its also crucial to shut down ones threads in a proper manner. The GlobalEventSystemFacade helps you do that.
Startup events
If you write plug-ins these can have a custom factory. Other code can use the ApplicationStartupAttribute attribute to get called by the Composite C1 core when the web app start up.
Data events
You can subscribe to data add, edit and delete events (pre and post) using the static methods on Composite.Data.DataEvents<T>. To attach to these events when the system start up, use the ApplicationStartupAttribute attribute.
Data
Composite.Core.Threading.ThreadDataManager is important if your accessing the Data Layer outside of a corresponding C1 Page request. This could be a custom handler that just has to feed all newest news as a Rss feed, or your maybe writing a console application. In these cases, always remember to wrap your code that accesses the data like this
using(Composite.Core.Threading.ThreadDataManager.EnsureInitialize())
{
// Code that works with C1 data layer goes here
}
For accessing and manipulating data its recommended NOT to use the DataFacade class, but wrap all code that gets or updates or deletes or adds data like this
using(var data = new DataConnection())
{
// Do things with data
}
IO
When working with files and directories its important to use the C1 equivalent classes Composite.Core.IO.C1File and Composite.Core.IO.C1Directory to .Net's File and Directory. This is due to the nature where C1 can be hosted on Azure, where you might not have access to the filesystem in the same way as you have on a normal Windows Server. By using the C1's File and Directory wrappers you can be sure that code you write will be able to run on Azure as well.
C1 Console
The console is a whole subject on itself and has many many API's.
You can create your own trees using Composite.C1Console.Trees.TreeFacade or Composite.C1Console.Elements.ElementFacade and implementing a Composite.C1Console.Elements.Plugins.ElementProvider.IElementProvider.
You can use the Composite.C1Console.Events.ConsoleMessageQueueFacade to send messages from the server to the client to make it do things like open a message box, refreshing a tree, set focus on a specific element, open a new tab etc. etc.
Composite.C1Console.Workflow.WorkflowFacade is used for getting instances of specific workflows and interacting with them. Workflows is a very fundamental part of C1 and is the way multi-step operations are defined and executed. This makes it possible to save state of operation so ie. a 10 step wizard is persisted even if the server restarts or anything else unexpected happens. Workflows are build using Windows Workflow Foundation, so are you familiar with this, you should be feeling at home
There is also a wealth of javascript facades and methods you can hook into when writing extensions to the Console. Much more than i could ever cover here so i will refrain myself from even getting started on that subject here.
composite.config
A fundamental part of C1 is providers, almost everything is made up of providers, even much of the core functionality. Everything in the console from Perspectives to Trees and elements and actions are feeded into C1 with providers. All the standard functions, the datalayer and all the widgets for use with the Function Call editor is feeded into C1 with providers. All the localisation strings for use with the Resources, users and permissions, url formatters etc. is all providers.
Composite.Data.Plugins.DataProviderConfiguration
Here all providers that can respond to the methods on DataFacade, Get, Update, Delete, Add etc. are registered. Every provider informs the system which interfaces it can interact with and C1 makes sure to route all requests for specific interfaces to their respective dataproviders.
Composite.C1Console.Elements.Plugins.ElementProviderConfiguration
Here we're defining the perspectives and the trees inside the Console. All the standard perspectives you see when you start the Console the first time are configured here, no magic or black box involved.
Composite.C1Console.Elements.Plugins.ElementActionProviderConfiguration
Action providers are able to add new menuitems to all elements in the system, based on their EntityToken. This is very powerful when you want to add new functionality to existing content like versioning, extranet security, custom cut/paste and the list goes on.
Composite.C1Console.Security.Plugins.LoginProviderConfiguration
A LoginProvider is what the C1 console will use to authenticate a user and let you log in or not. Unfortunately this isn't very open but with some reflection you should be all set.
Composite.Functions.Plugins.FunctionProviderConfiguration
Composite C1 will use all the registered FunctionProviders to populate its internal list of functions on system startup.
Composite.Functions.Plugins.WidgetFunctionProviderConfiguration
WidgetProviders are used in things like the Function Call Editor or in Forms Markup to render custom UI for selecting data.
Composite.Functions.Plugins.XslExtensionsProviderConfiguration
Custom extensions for use in XSLT templates are registered here
And then we have a few sections for pure configuration, like caching or what to to parallelize but its not as interesting as the providers.
Defining and using sections
Sections in composite.config, and other related .config files are completely standard .Net configuration and obeys the rules thereof. That means that to be able to use a custom element, like ie. Composite.Functions.Plugins.WidgetFunctionProviderConfiguration it has to be defined as a section. A section has a name and refers to a type that would inherit from System.Configuration.ConfigurationSection. Composite uses the Microsoft Enterprise Libraries for handling most of these common things like configuration and logging and validation and therefor all Composites sections inherit from Microsoft.Practices.EnterpriseLibrary.Common.Configuration.SerializableConfigurationSection. Now, this type just has to have properties for all the elements we want to be able to define in the .config-file, and .Net will automatically make sure to wire things up for us.
If you want to access configuration for a particular section you would call Composite.Core.Configuration.ConfigurationServices.ConfigurationSource.GetSection(".. section name") and cast it to your specific type and your good to go.
Adding extra properties to already defined sections
Normally .Net would complain if you write elements or attributes in the .config files that aren't recognized by the type responsible for the section or for the element. This makes it hard to write a truly flexible module-system where external authors can add specific configuration options to their providers, and therefor we have the notion of a Assembler. Its a ConfigurationElement class with a Microsoft.Practices.EnterpriseLibrary.Common.Configuration.ObjectBuilder.AssemblerAttribute attribute assigned to it that in turns takes a Microsoft.Practices.EnterpriseLibrary.Common.Configuration.ObjectBuilder.IAssembler interface as argument that is responsible for getting these custom attributes and values from the element in the .config file and emit usable object from it. This way .Net won't complain about an invalid .config file, since we inject a ConfigurationElement object that has properties for all our custom attributes, and we can get hold of them when reading the configuration through the IAssembler
Slides
Some overview slides can be found on these lins
Overview
Extensibility points
Page request handling
Function system
Data system
Data type system
Inspiration and examples
The C1Contrib project on GitHub is a very good introduction how to interact with the different parts of C1. Its a collection of small packages, that can be used as it is, or for inspiration. There are packages that manipulates with dynamic types to enable interface inheritance. Other packages uses the javascript api in the console, while others show how to make Function Providers, Trees and hook commands unto existing elements. There is even examples of how to manipulate with the Soap webservice communication going on between client and server so you can make it do things the way you want it. And the list goes on.
I have a WPF application that loads modules into my application via prism / MEF. Each module will reference an assembly I created to house NHibernate functionality (Unit of Work & repository pattern). Now that I started on the first module, the following questions came up:
Is it common place to have an nhibernate session open for the lifetime of a WPF / windows application?
Would it be a good idea to share the nhibernate session amongst all modules or have each module create its own session?
It's my understanding that the session is lightweight, so opening and closing a session is not a problem... however, if I close the session then I lose the change tracking... when I call Save, and the object is the same as when it was retrieve, I do not want a query to be executed.
Do not have the session open for a the lifetime of a wpf/windows application. You will have performance and memory problems if you do.
It really depends how these modules interact. Are they separate systems, or do they need to work in tandem to do work. If it's the former, I would say that they shouldn't share sessions. If it's the latter, I would use some kind of session manager that can pass out the current session.
I have a WPF application where I'm using the MVVM pattern.
I get the VM activated for actions that require user input and thus need to activate views from the VM.
I've started out separating the VMs into a separate component/assembly, partly because I see them as the unit testable part, partly because views should rely on VM, not the other way round. But when I then need to bring up a window, the window is not known by the VM.
All introductions I find place the VM in the WPF/App component, thus eliminating the problem.
This article recommends keeping them in separate layers : http://waf.codeplex.com/wikipage?title=Architecture%20-%20Get%20The%20Big%20Picture&referringTitle=Home
As I see it, I have the following choices
Move VMs to the WPF/App assembly to allow VMs to access the windows directly.
Place interfaces of views in VM-assembly, implement views in WPF/App assembly and register the connection through IOC or alternative ways.
File a 'request' from the VM into some mechanism/bus that routes the request (but which mechanism!? E.g something in Prism?!)
What's the recommendation?
Thanks for any comments,
Anders, Denmark
Don't pick option 1. You'll be adding an unwanted dependency from VM to V.
Options 2 and 3 are both valid and being used. Picking between these is sometimes a matter of taste. I think that IOC enables/allows mocking better whereas a messagebus works fine for small applications.
Keep your ViewModels in a separate assembly from the Views.
If you look into Cinch and MEFedMVVM you'll see a very powerful mechanisms for connecting views and viewmodels using MEF. Keeping them separate facilitates running your application headless (no UI), which is great for testing and exposing an API.