Showing posts with label Generics. Show all posts
Showing posts with label Generics. Show all posts

Thursday, March 18, 2010

How to use strongly-typed name with INotifyPropertyChanged

You may already have read my posts about how to use INotifyPropertyChanged in a type-safe way (here and here), but sometime you don’t want to modify all your classes to use this method. All you want is avoid the use of a magic string to define the property. Your code will be refactoring proof.

For example let say you have this property:

public string FirstName
{
    get { return _firstName; }
    set 
    {
        if (_firstName == value)
            return;
        _firstName = value;
        RaisePropertyChanged("FirstName");
    }
}

Some refactoring tool, like Resharper, will be able to change the “FirstName” string but not Visual Studio itself. The solution? Replace this string with a strong type value. How? Let’s assume we can do this:

public string FirstName
{
    get { return _firstName; }
    set 
    {
        if (_firstName == value)
            return;
        _firstName = value;
        RaisePropertyChanged(this.NameOf(p => p.FirstName));
    }
}

Notice that you must specify the “this” keyword to make it work. That is exactly What this extension method let you do:

public static class ObjectExtensions
{
    public static string NameOf<T>(this T target, Expression<Func<T, object>> propertyExpression)
    {
        MemberExpression body = null;
        if (propertyExpression.Body is UnaryExpression)
        {
            var unary = propertyExpression.Body as UnaryExpression;
            if (unary.Operand is MemberExpression)
                body = unary.Operand as MemberExpression;
        }
        else if (propertyExpression.Body is MemberExpression)
        {
            body = propertyExpression.Body as MemberExpression;
        }
        if (body == null)
            throw new ArgumentException("'propertyExpression' should be a member expression");

        // Extract the right part (after "=>")
        var vmExpression = body.Expression as ConstantExpression;

        // Extract the name of the property to raise a change on
        return body.Member.Name;
    }
}

You can even use it in your property changed handler:

private void OnPropertyChanged(object sender, PropertyChangedEventArgs args)
{
    if (args.PropertyName == this.NameOf(p => p.FirstName))
    {
        // ...
    }
}

Enjoy!

Wednesday, March 25, 2009

Good practice to use Dispatcher in WPF background thread

Here is a good way to use extension method in a multi threaded context. Everybody knows that when you try to update UI from any other thread than the UI you get an “InvalidOperationException” with message “The calling thread cannot access this object because a different thread owns it.”. Look at the following sample:

Let say somewhere in your code you have this

private void Button_Click(object sender, RoutedEventArgs e)
{
    // ...
    ThreadPool.QueueUserWorkItem(DoWork, this);
    // ...
}

If you implement DoWork Like this…

private static void DoWork(object state)
{
    Window1 win = (Window1) state;
    for (int i = 0; i < 100; i++)
    {
        // do some work
        win.progress1.Value = i;
    }
    win.progress1.Value = 100;
}

…you will be in trouble.

Because you can’t update UI from a thread other than the UI one you will get the InvalidOperationException as stated before.

The solution is to use the Dispatcher object. As from microsoft documentation:

Only the thread that the Dispatcher was created on may access the DispatcherObject directly. To access a DispatcherObject from a thread other than the thread theDispatcherObject was created on, call Invoke or BeginInvoke on the Dispatcher the DispatcherObject is associated with.

Subclasses of DispatcherObject that need to enforce thread safety can do so by calling VerifyAccess on all public methods. This guarantees the calling thread is the thread that theDispatcherObject was created on.

So our previous sample should look like this:

private static void DoWork(object state)
{
    Window1 win = (Window1) state;
    for (int i = 0; i < 100; i++)
    {
        // do some work
        win.Dispatcher.Invoke(new Action<ProgressBar, int>((p, v) => p.Value = v), win.progress1, i);
    }
    win.Dispatcher.Invoke(new Action<ProgressBar>(p => p.Value = 100), win.progress1);
}

This is a little more work but not it works. Because we don’t want to call the Dispatcher object when it’s not needed we sould do this:

private static void DoWork(object state)
{
    Window1 win = (Window1) state;
    for (int i = 0; i < 100; i++)
    {
        // do some work
        if (win.Dispatcher.CheckAccess())
            // We can call on the current thread
            win.progress1.Value = i;
        else
            // we need to call Invoke
            win.Dispatcher.Invoke(new Action<ProgressBar, int>((p, v) => p.Value = v), win.progress1, i);
    }

    if (win.Dispatcher.CheckAccess())
        // We can call on the current thread
        win.progress1.Value = 100;
    else
        // we need to call Invoke
        win.Dispatcher.Invoke(new Action<ProgressBar>(p => p.Value = 100), win.progress1);
}

Ouch! This is a lot more work. We cannot do that every time. That’s when extensions method comes handy. We can replace this whole process of choosing the right implementation with a single extension method. It will make our code more readable and more managable.

He is the whole extension class with all possible overload for a method called Dispatch. This method will dispatch the process only if needed:

public static class DispatcherExtensions
{
    public static TResult Dispatch<TResult>(this DispatcherObject source, Func<TResult> func)
    {
        if (source.Dispatcher.CheckAccess())
            return func();

        return (TResult) source.Dispatcher.Invoke(func);
    }

    public static TResult Dispatch<T, TResult>(this T source, Func<T, TResult> func) where T : DispatcherObject
    {
        if (source.Dispatcher.CheckAccess())
            return func(source);

        return (TResult)source.Dispatcher.Invoke(func, source);
    }

    public static TResult Dispatch<TSource, T, TResult>(this TSource source, Func<TSource, T, TResult> func, T param1) where TSource : DispatcherObject
    {
        if (source.Dispatcher.CheckAccess())
            return func(source, param1);

        return (TResult)source.Dispatcher.Invoke(func, source, param1);
    }

    public static TResult Dispatch<TSource, T1, T2, TResult>(this TSource source, Func<TSource, T1, T2, TResult> func, T1 param1, T2 param2) where TSource : DispatcherObject
    {
        if (source.Dispatcher.CheckAccess())
            return func(source, param1, param2);

        return (TResult)source.Dispatcher.Invoke(func, source, param1, param2);
    }

    public static TResult Dispatch<TSource, T1, T2, T3, TResult>(this TSource source, Func<TSource, T1, T2, T3, TResult> func, T1 param1, T2 param2, T3 param3) where TSource : DispatcherObject
    {
        if (source.Dispatcher.CheckAccess())
            return func(source, param1, param2, param3);

        return (TResult)source.Dispatcher.Invoke(func, source, param1, param2, param3);
    }

    public static void Dispatch(this DispatcherObject source, Action func)
    {
        if (source.Dispatcher.CheckAccess())
            func();
        else
            source.Dispatcher.Invoke(func);
    }

    public static void Dispatch<TSource>(this TSource source, Action<TSource> func) where TSource : DispatcherObject
    {
        if (source.Dispatcher.CheckAccess())
            func(source);
        else
            source.Dispatcher.Invoke(func, source);
    }

    public static void Dispatch<TSource, T1>(this TSource source, Action<TSource, T1> func, T1 param1) where TSource : DispatcherObject
    {
        if (source.Dispatcher.CheckAccess())
            func(source, param1);
        else
            source.Dispatcher.Invoke(func, source, param1);
    }

    public static void Dispatch<TSource, T1, T2>(this TSource source, Action<TSource, T1, T2> func, T1 param1, T2 param2) where TSource : DispatcherObject
    {
        if (source.Dispatcher.CheckAccess())
            func(source, param1, param2);
        else
            source.Dispatcher.Invoke(func, source, param1, param2);
    }

    public static void Dispatch<TSource, T1, T2, T3>(this TSource source, Action<TSource, T1, T2, T3> func,
                                                     T1 param1, T2 param2, T3 param3) where TSource : DispatcherObject
    {
        if (source.Dispatcher.CheckAccess())
            func(source, param1, param2, param3);
        else
            source.Dispatcher.Invoke(func, source, param1, param2, param3);
    }
}

That seems a lot of code to write but see how it simplifies the code when you use it:

private static void DoWork(object state)
{
    Window1 win = (Window1) state;
    for (int i = 0; i < 100; i++)
    {
        // do some work
        win.progress1.Dispatch((p, v) => p.Value = v, i);
    }

    win.progress1.Dispatch(p => p.Value = 100);
}

This is almost as simple as our first implementation of DoWork. The only difference is in this version we call “Dispatch” with a lambda expression that will always be run on the UI thread.

Let me know if you find this helpful or if you think of something better.

Monday, January 7, 2008

Volta: Redefining Web Development

Volta: Redefining Web Development

via Yet Another Language Geek by wesdyer on 12/5/07

Anyone who writes web applications knows that web development is not easy. Developers wrangle with a soup of technologies distributed across multiple tiers. We live in a world where programmers accept the fact that they need to know four or five different languages, tools, and environments just to get a site up and running. In ancient times, the Egyptians built marvelous structures despite the primitive tools that the workmen used. Building a pyramid took most of the national resources of wealth and labor. Today, we build structures which are vastly more complicated and yet require only a tiny fraction of the resources. The difference is in the tools and the infrastructure.

In a similar way, Volta significantly improves web development. Programmers write web applications using familiar .NET languages, libraries, and tools. Volta splits the application into multiple parts potentially running on different tiers, say, client and server. Client code needs only a minimal, JavaScript-enabled browser, though Volta will take advantage of additional runtimes that may be present.

Programmers simply refactor classes that need to run on tiers other than the client and Volta injects the boilerplate code for communication, serialization, synchronization -- all the remoting code. The developer enjoys all of the benefits of .NET: a great debugger, test tools, profiling, intellisense, refactorings, etc.

Just how simple is Volta? Let's write an application that uses a button to query the server for a string and displays that string to the client: the hello world web application.

image

Now, let's write the code for the web page. We need a Div for the output and an Input for the interaction. Of course, we could have constructed the page elements with HTML/CSS instead.

using System;
using Microsoft.LiveLabs.Volta.Html;
using Microsoft.LiveLabs.Volta.Xml;

namespace HelloWorld
{
    public partial class VoltaPage1 : Page
    {
        public VoltaPage1()
        {
            var output = new Div();
            var b = new Input();
            b.Type = "button";
            b.Value = "Get Message";
            b.Click += () => output.InnerHtml = C.GetMessage();
            Document.Body.AppendChild(output);
            Document.Body.AppendChild(b);
        }
    }

    internal class C
    {
        public static string GetMessage()
        {
            return "Hello, World";
        }
    }
}

But we want to produce the message on the server. Time to refactor.

image

Browser clients call the server, the "Origin", because this ensures the message will come from the same server that supplied the HTML.

[RunAtOrigin]
internal class C
{
    public static string GetMessage()
    {
        return "Hello, World";
    }
}

That is it. Try it out.

image

Now, click "Get Message".

image

Great. But is it cross-browser compatible?

image

Yes. And you can debug it.

image

You can even debug across tiers.

image

There is a lot more to Volta in the first technology preview which was made publicly available at 11 am PST today and there will be a lot more to come.

Still skeptical? Try it out for yourself.

http://labs.live.com/volta

image

Saturday, November 24, 2007

Passing anonymous to and from method's call

As posted by Alex James on Meta-Me blog.

T CastByExample(object o, T example)

So earlier today I was lamenting that an anonymous type can't be
shared between functions with
Wes Dyer, when he said "Well actually they can..."

Cue me learning something cool.

The first step is to create a seemingly innocent method:

    public static T CastByExample<T>(this object o, T example)
    {
        return (T) o;
    }
Seems innocent enough right? Well it is until you start using it with anonymous types. Imagine you had this function, that returns an anonymous type as object, because that is your only choice:
static object GetAnonymousType()
{
    return new { FullName = "Cosmo Kramer" };
}
Normally if you called this function anywhere you wouldn't be able to get at the anonymous type without using reflection... This is where CastByExample<T> comes to the rescue. If you know the shape of the anonymous type, you use that to do a CastByExample...
        object o = GetAnonymousType();

        //get the original anonymous type back again
        var v = o.CastByExample(new { FullName = "" });

        //Use the properties of the anonymous type initialized in another
        //function directly !!
        Console.WriteLine(v.FullName);

This works because when an anonymous type is used the compiler first checks that one with the same signature (i.e. all fields are the same name and type) hasn't already been used. If one has the same CLR type is used.

Hence if you pass in an example that is the same shape as the original anonymous type to the CastByExample(..) method will get you back to the original anonymous type... and var magic does the rest.

Nifty huh?