Technetguy.com/blog

BING is live, the msn search is gone.

2009-06-01T21:07:27Z

Watched the live video of the BING, I really liked the idea about the categorized search result. Also found that http://search.msn.com is already redirected to the http://www.bing.com Also noticed a unofficial definition of the BING -- :-) "BING = But its not Google" It’s interesting that Microsoft is taking the social media like Twitter to do the release announcement as well....(read more)

WCF RESTFul Service

2009-04-12T04:17:04Z

Creating a RESTFul web service in WCF can be easier than a normal web service!? Some of my friends did not believe and they even argued that REST is not considered as a web service.So I figured it out by using the http://en.wikipedia.org/wiki/Web_service . Thanks wikipedia! With the installation of the WCF REST Service Start Kit , the RESTFul web service can be created by using a wizard and you can create multi types of service as well.Before you install this start kit, you have to install the visual...(read more)

LINQ to Flickr support for Photo search

2009-04-05T03:06:00Z

Long time ago, I had one post said that the NFlickr is going to support the LINQ , because my new baby was born at last Dec and I could not get any chance until yesterday. So NOW I will announce that the first version of the LINQ to Flickr is ready for the Alpha test. To search your photos can be as simple as this 1: PhotoList list = new PhotoList(service); 2: var photos = from photo in list 3: where photo.TagName == "Yang" && photo.Family == true 4: select photo; .csharpcode, .csharpcode...(read more)

The problem of JSON and how to fix it(Update with code sample)

2008-11-18T04:39:00Z

In AJAX world, JSON is one of the most popular implementation, Asp.Net AJAX is using this concept, and it really works great. But in some very special situation, JSON could cause some unexpected problems.For example, if the user want to submit a huge document on the local to the server side, you will find it takes too much time before the server-side gets requested.Why? We knew that one task for the so-call AJAX framework is to serialize the data and on the server-side to de-serialize the data. But...(read more)

Multi culture programming in Asp.Net

2008-11-16T06:08:00Z

Asp.Net provide a very powerful multi culture programming pattern.In this post , we will discuss specially under the user control level. Last weekend, one of my friends asked the question about the user control multi culture programming problem. He could not make it work at that time and I was in the middle of the NFlickr Release , so I promised to write a post about this problem. First, I did not know what exactly the problem was, but I will try to use the standard way to implement it. The user...(read more)

The first version of the NFlickr released

2008-11-16T02:41:00Z

Yesterday, the first Alpha version of the NFlickr was released.This version was fully written by the .Net 3.5 and using the provider concept.For example, if you don't like the built-in encryption or you don't like the implementation of the http communication. you can use the same interface of the provider. From the functionality part, this version will have the below support: 1. Login to the Flickr Service 2. Create the PhotoSet. 3. Upload the photo 4. Add the photo to a photoSet. 5. Search the photo...(read more)

Is silverlight application the windows or web?

2008-11-15T04:40:00Z

Before we started to develop a silverlight application, does anyone ask the question? Is silverlight the windows or web? The short answer is Windows, but wait.... All the code was executed on the client side, the code was download was downloaded to the client side.It has to be executed under the silverlight ActiveX control or browser plugin(For Firefox, Chrome, etc). But why we still call the silverlight as a web application? Because it can interact with the HTML Dom, it can access the IIS by using...(read more)

Using the DOM to control the element(TreeView) inside Silverlight

2008-11-14T03:55:49Z

In our Update the Silverlight toolkit(AutoCompleteBox) With the HTML DOM value post, we talked about the interacting between the Silverlight application and html DOM. It's about using the Silverlight to access the DOM. Today we are going to discuss using the DOM to access the Silverlight. In the post Using the Silverlight toolkit build a online TreeView , we had a demo application to demo a buddy list tree using the Silverlight toolkit.Today we will wrote a more complex application to update...(read more)

Update the Silverlight toolkit(AutoCompleteBox) With the HTML DOM value

2008-11-13T05:16:00Z

The silverlight toolkit provided the AutoCompleteBox and We already had one post talking about the usage about it.

This post will focus on the way we can update the content of the AutoCompleteBox by using the value from the HTML DOM value.

The initialized code was not changed and we only added a new button on the silverlight control which is the

<Button Height="24" HorizontalAlignment="Right" Margin="0,20,123,0" VerticalAlignment="Top" Width="76" Click="Button_Click" Content="Refresh"/>

Then we go to the web page ToolboxTestTestPage.aspx to add a new Textarea input box which will provide the content.

<textarea id="DataSource" name="DataSource" cols="10" rows="5"></textarea>

The real problem is about the interacting the silverlight application and the HTML DOM object.

The first code I tried is to use the common sense way which is

   1:             HtmlElement datasource= HtmlPage.Document.GetElementById("DataSource");

   2:              if (datasource != null)

   3:              {

   4:                  string content = datasource.GetProperty("innerHTML") as string;

   5:                  List<string> list=content.Split("\n".ToCharArray()).ToList();

   6:                  txtAutoComplete.ItemsSource = list;

   7:              }

But I find it's almost no way for me get the user input value, so I changed my way to using the Javascript.

   1:             string content = (string)HtmlPage.Window.Invoke("GetDataSource", null);

   2:              List<string> list = content.Split("\n".ToCharArray()).ToList();

   3:              txtAutoComplete.ItemsSource = list;

when I did this , I realized that how cool and convenient will be when we have the C# dynamic.

And the Javascript on the page is

   1:  <head runat="server">

   2:      <title>ToolboxTest</title>

   3:      <script type="text/javascript" language="javascript">

   4:          function GetDataSource() {

   5:              return $get("DataSource", document).value;

   6:          }

   7:      </script>

   8:  </head>

Keep in mind , don't use the Html.Document.Invoke because the Javascript is an instance of the html window, not the document. In the future post, I will take another way around to getting the silverlight value from the HTML Dom.

LINQ Tips #7 Create a LINQ to Anything(3)

2008-11-12T03:42:27Z

Yesterday, we implemented a simple LINQ to Objects pattern. In our post, we overwrote the data source , if you can remember, all the return value were added a tail as "LINQ".

But this seems useless, in the NFlickr project, we need to send the filter to the Flickr service and get the filtered result.So we need to implement the "Where" clause.

Before we start, can you answer the question first?

If we want to implement the new clause like "Where","Orderby", where we should implement?

The answer is the method (CreateQuery) for the interface IQueryProvider!

Yesterday, we only handled the "Select" clause, I will add a new case switch to implement the "Where" condition.

   1:              case "Where":

   2:                              StringBuilder sb = new StringBuilder();

3:

   4:                              functions = new List<KeyValuePair<string, Func<int>>>();

   5:                              filterList = new List<string>();

   6:                              LambdaExpression lambdaExpression = (LambdaExpression)(((UnaryExpression)callExpression.Arguments[1]).Operand);

   7:                              ExpandExpression(lambdaExpression.Body);

   8:                              return (IQueryable<TElement>)this;

The key point is to cast the callExpression from the Expression to LambdaExpression. After we got the LambdaExpression, we can go inside the method ExpendExpression.

   1:          List<string> filterList = null;

   2:          public void ExpandExpression(Expression lambdaExpression)

   3:          {

   4:              switch (lambdaExpression.NodeType)

   5:              {

   6:                  case ExpressionType.Equal:

   7:                      Expression right = (BinaryExpression)lambdaExpression;

   8:                      filterList.Add((string)((ConstantExpression)right).Value);

   9:                      break;

  10:              }

  11:          }

This function is to preserver the filter list and this list will be used we the code is to return the data source.

        public IEnumerator<Student> GetEnumerator()
        {
            
            IEnumerator<Student> list = (IEnumerator<Student>)this;
            if (functions == null)
                return list;

            List<Student> result = new List<Student>();

            foreach (Student student in students)
            {
                foreach (string filter in filterList)
                {
                    if (student.FirstName.Equals(filter, StringComparison.OrdinalIgnoreCase))
                    {
                        result.Add(student);
                    }
                }
            }
            return result.GetEnumerator();
            
        }

Now, we changed the behavior from the original LINQ expression from the case sensitive to not-sensitive.

   1:    var classmates = from classmate in GetClassRoom2()

   2:                               where classmate.FirstName=="tom"

   3:                               select classmate;

And

   1:    var classmates = from classmate in GetClassRoom2()

   2:                               where classmate.FirstName=="Tom"

   3:                               select classmate;

will be the same result.

From now, a simple version of the LINQ to "Anything" can be released, welcome to send me email or leave a comment about this topic.

LINQ Tips #6 Create a LINQ to Anything(2)

2008-11-11T03:35:30Z

Yesterday, we sort of created a baby step of a LINQ to anything implementation;

We stopped at the place to implement the interface IQueryProvider, the most important method in this interface for us is IQueryable<TElement> CreateQuery<TElement>(System.Linq.Expressions.Expression expression).

Our code will look like:

   1:   #region IQueryProvider Members

2:

   3:          public IQueryable<TElement> CreateQuery<TElement>(System.Linq.Expressions.Expression expression)

   4:          {

   5:              IQueryable<TElement> element = null;

   6:              switch (expression.NodeType)

   7:              {

   8:                  case System.Linq.Expressions.ExpressionType.Call:

   9:                      MethodCallExpression callExpression = (MethodCallExpression)expression;

  10:                      switch (callExpression.Method.Name)

  11:                      {

  12:                          case "Select":

  13:                              element = new StudentQueryObject<TElement>(callExpression);

  14:                              break;

  15:                          default:

  16:                              throw new NotImplementedException("This method is not supported");

  17:                      }

  18:                      break;

  19:              }

  20:              return element;

  21:          }

22:

  23:          public IQueryable CreateQuery(System.Linq.Expressions.Expression expression)

  24:          {

  25:              return CreateQuery<Student>(expression);

  26:          }

27:

  28:          public TResult Execute<TResult>(System.Linq.Expressions.Expression expression)

  29:          {

  30:              MethodCallExpression callExpression = (MethodCallExpression)expression;

  31:              throw new NotImplementedException("No execution is defined");

  32:          }

33:

  34:          public object Execute(System.Linq.Expressions.Expression expression)

  35:          {

  36:              return Execute<Student>(expression);

  37:          }

38:

  39:          #endregion

Maybe you already noticed that in the "Select" method , We created a new class called StudentQueryObject. Before we went into this class, we can say that all the implementation we have now are kind of generic implementation and can be used under almost all situation.

The implementation for the StudentQueryObject class is

   1:      public class StudentQueryObject<T> : IQueryable<T>

   2:      {

   3:          private Expression _expression;

4:

   5:          public StudentQueryObject(Expression expression)

   6:          {

   7:              _expression = expression;

   8:          }

   9:          #region IEnumerable<T> Members

10:

  11:          public IEnumerator<T> GetEnumerator()

  12:          {

  13:              MethodCallExpression callExpression = (MethodCallExpression)_expression;

  14:              ClassRoom2 classroom = (ClassRoom2)(((ConstantExpression)callExpression.Arguments[0]).Value);

15:

  16:              UnaryExpression quote = (UnaryExpression)callExpression.Arguments[1];

17:

  18:              Expression<Func<Student, T>> projector = (Expression<Func<Student, T>>)quote.Operand;

  19:              Func<Student, T> func= projector.Compile();

20:

  21:              List<T> list = new List<T>();

22:

23:

  24:              foreach (Student student in classroom)

  25:              {

  26:                  list.Add(func(new Student { FirstName=student.FirstName +"Linq", LastName=student.LastName + "Linq" }));

  27:              }

  28:              return list.GetEnumerator();

  29:          }

30:

  31:          #endregion

32:

  33:          #region IEnumerable Members

34:

  35:          System.Collections.IEnumerator System.Collections.IEnumerable.GetEnumerator()

  36:          {

  37:              return GetEnumerator();

  38:          }

39:

  40:          #endregion

41:

  42:          #region IQueryable Members

43:

  44:          public Type ElementType

  45:          {

  46:              get { return typeof(T); }

  47:          }

48:

  49:          public System.Linq.Expressions.Expression Expression

  50:          {

  51:              get { return _expression; }

  52:          }

53:

  54:          public IQueryProvider Provider

  55:          {

  56:              get { return null; }

  57:          }

58:

  59:          #endregion

  60:      }

The method of the GetEnumerator can be the data provider for the LINQ query source. So basically , we

can put the data provider logic here .In this example, I just changed the original provided data.

All the data will be added a word "Linq" at the tail.

This is a fairly simple implementation, But this can be extended to very complex scenario.The NFlickr  will have

a full set of implementation and I will have a post after the first release of the NFlickr.

Using the Pex to write Unit test automatically

2008-11-10T20:55:11Z

I know this is against TDD and I am not going to discuss this :)

MSR just released a new product call Pex, you can also find a video on PDC at http://channel9.msdn.com/pdc2008/TL51/.

Today , I am going to use Pex to write unit tests for my two ugly code.

First, My original code is like

   1:  public bool RunDemo(string message, int value, DateTime actionTime)

   2:          {

3:

   4:              int length = message.Length;

5:

   6:              if (value > length)

   7:              {

   8:                  value = 50;

   9:              }

  10:              if (value < 100)

  11:              {

  12:                  for (int i = 0; i < value; i++)

  13:                  {

  14:                      if (actionTime < DateTime.Now)

  15:                      {

  16:                          Console.Write(message);

  17:                          Console.WriteLine();

  18:                          if (value > 100)

  19:                          {

  20:                              Console.WriteLine("To see what will happen");

  21:                          }

  22:                      }

  23:                  }

  24:              }

  25:              return true;

  26:          }

It's basically the ugliest code I had ever written. But how the unit test will go?

The test result will be like

After we take a deeper look, we will find out that the failed unit test case is

   1:  [TestMethod]

   2:  [PexRaisedException(typeof(NullReferenceException))]

   3:  [PexGeneratedBy(typeof(Class1Test))]

   4:  public void RunDemo03()

   5:  {

   6:      bool b;

   7:      Class1 class1 = new Class1();

   8:      b = this.RunDemo(class1, (string)null, 2, default(DateTime));

   9:  }

So we know in out RunDemo code, we need to have a argument safety check first., So I neet to write the code like

   1:             // <pex>

   2:              if (string.IsNullOrEmpty(message))

   3:                  return false;

   4:              // </pex>

The second example is we are trying to write a Fibonacci code like this

   1:         public int Fibonacci(int n)

   2:          {

   3:               return Fibonacci(n - 1) + Fibonacci(n - 2);

   4:          }

The logic is so simple and we think we are done, but Pex dose not think so.

And we need to fix the code as

   1:   public int Fibonacci(int n)

   2:          {

   3:              if (n==1 || n==0)

   4:              {

   5:                  return n;

   6:              }

   7:               return Fibonacci(n - 1) + Fibonacci(n - 2);

   8:          }

In this two stupid simple code, we will have a quick snapshot for the Pex tools.

Also, the Pex has some limitation and we need to know that Pex is code, not a real developer.
Pex assumes that the analyzed program is deterministic. If it is not, Pex will go in cycles until it hits exploration bounds.
Pex does not handle multithreaded programs.
Pex does not understand native code, e.g. x86 instructions called through P/Invoke,
At this time, we only support Pex on the X86, 32-bit .NET framework.
In principle, Pex can analyze arbitrary .NET programs, written in any .NET language.
However, the VS AddIn and the code generation are only fully implemented for C#.
Pex injects generic callbacks when rewriting the IL method bodies. This has a side effect that it cannot instrument v1.0 and v1.1 .net module because their metadata models do not support generics.

There is one thing I don't like for the Pex is that Pex code generation will only generate the MsTest project, I hope it can generate a NUnit or MBUnite project in future. (Removed after I read the resource on htp://www.codeplex.com/Pex)

In the coming days, I will introduce the coverage test done by the Pex.

LINQ Tips #5 Creating a LINQ To Anything

2008-11-10T20:55:00Z

We knew there are at least 3 implementation in .Net 3.5, they are LINQ to Objects, LINQ to XML and LINQ to SQL. I don't believe Microsoft will create different pattern for those.

In my current open source project NFlickr, because I have thousands of the fotos at flickr.com , so I don't want to get everything and then filter it using the LINQ to Objects. I want to filter it on the server side and still use the same syntax.

It's not easy, so I figured out that we need to start from something simple.In the previous post LINQ Tips #3--Why you can not use LINQ? , I already had a Student class and a Classroom class.

What I need to do is to change the class Classroom inherited from IQueryable<Student>, not IEnumerable<Student> (We could make it generic, but we will do it later), The new Classroom class will be :

   1:  public class ClassRoom2 : IQueryable<Student>, IEnumerator<Student>

   2:      {

   3:          private List<Student> students = new List<Student>();

   4:          private int _index = 0;

   5:          public void Add(Student student)

   6:          {

   7:              students.Add(student);

   8:          }

   9:          #region IEnumerable<Student> Members

  10:          public IEnumerator<Student> GetEnumerator()

  11:          {

  12:              return (IEnumerator<Student>)this;

  13:          }

  14:          #endregion

  15:          #region IEnumerable Members

  16:          IEnumerator IEnumerable.GetEnumerator()

  17:          {

  18:              return GetEnumerator();

  19:          }

  20:          #endregion

  21:          #region IEnumerator<Student> Members

  22:          public Student Current

  23:          {

  24:              get { return students.ElementAt(_index-1); }

  25:          }

  26:          #endregion

  27:          #region IDisposable Members

  28:          public void Dispose()

  29:          {

30:

  31:          }

  32:          #endregion

  33:          #region IEnumerator Members

  34:          object IEnumerator.Current

  35:          {

  36:              get { return Current; }

  37:          }

  38:          public bool MoveNext()

  39:          {

  40:              _index++;

  41:              if (_index > students.Count)

  42:              {

  43:                  return false;

  44:              }

  45:              return true;

  46:          }

  47:          public void Reset()

  48:          {

  49:              _index = 0;

  50:          }

  51:          #endregion

  52:          #region IQueryable Members

  53:          public Type ElementType

  54:          {

  55:              get { throw new NotImplementedException(); }

  56:          }

  57:          public System.Linq.Expressions.Expression Expression

  58:          {

  59:              get {

  60:                  throw new NotImplementedException();

  61:              }

  62:          }

  63:          public IQueryProvider Provider

  64:          {

  65:              get {

  66:                  throw new NotImplementedException();

  67:              }

  68:          }

  69:          #endregion

  70:      }

We will notice that there are three new properties which are ElementType,Expression,Provider.What's those three property means, and how can we get the right implementation? The answer is that we need to have the Classroom inherited from another interface--IQueryProvider.
And we will have the three property implementation as :

   1:  public class ClassRoom2 : IQueryable<Student>, IEnumerator<Student>, IQueryProvider

   2:      {

   3:          #region IQueryable Members

4:

   5:          public Type ElementType

   6:          {

   7:              get { return typeof(Student); }

   8:          }

9:

  10:          public System.Linq.Expressions.Expression Expression

  11:          {

  12:              get {

  13:                  return System.Linq.Expressions.Expression.Constant(this);

  14:              }

  15:          }

16:

  17:          public IQueryProvider Provider

  18:          {

  19:              get {

  20:                  return this;

  21:              }

  22:          }

23:

  24:          #endregion

  25:      }

Then we need to implement the new four methods for the interface IQueryProvider:

The key method for us is the Method IQueryable<TElement> CreateQuery<TElement>(System.Linq.Expressions.Expression expression).To understand this method, I will suggest you to watch the video at Channel 9.

Tomorrow, I will continue on this topic and finish the full implementation and you will know how powerful the LINQ can be!

Asp.Net Ajax #1 Show/Hide ModalPopup using Javscript

2008-11-10T20:54:33Z

We I first started to use the Asp.Net Ajax toolkit, I found there is a very fundamental requirement from a lot developers which we want to have some kinds of control on the client side, no need to trigger the service side event.

After I did some investigation, I found a very easy built-in way to do it.

The asp.net markup is like

   1:  <asp:Panel runat="server" ID="pnlDemo" style="display:none">

   2:          <h1>This panel is a test demo</h1>

   3:      </asp:Panel>

4:

   5:  <cc1:ModalPopupExtender ID="ModalPopupExtender" runat="server"  TargetControlID="linkButton1"

   6:              PopupControlID="pnlDemo"

   7:              BackgroundCssClass="modalBlockBackground" BehaviorID="ModalPopupExtender"

   8:              DynamicServicePath="" Enabled="True">

   9:      </cc1:ModalPopupExtender>

If we want to use the javascript to control the popup, we can just simply to write the code like this:

   1:   if(!show)

   2:    {

   3:       $find('ModalPopupExtender').hide();

   4:    }

   5:    else

   6:    {

   7:        $find("ModalPopupExtender").show();

   8:    }

There is one thing need to be taken care which is the BehavirID, the code in the Javasctipt will look for the BehaviorID property for the ModalPopupExtender.

I hope this can be helpful for you.

Generic Event

2008-11-10T20:53:53Z

In C#, when we need to have an event, we suppose to have a delegate first. Most of the delegate will be defined as the below one if it served for an event;

   1:      public delegate void SignatureRequiredEventHandler(object sender, SignatureRequiredEventArgs args);

   2:      public delegate void SendRequestCompletedEventHandler(object sender, SendRequestCompletedEventArgs args);

   3:      public delegate void HttpRequestdEventHandler(object sender, HttpRequestEventArgs args);

This code is copied from the the NFlickr opensource project.

In my previous post, Software development principle , there is one principle called DRY(Don't Repeat Yourself), it's very obvious this kind of code is already not followed the good practice.

.Net now offer a generic event,you can find the MSDN document at http://msdn.microsoft.com/en-us/library/db0etb8x.aspx

And we simply don't need the code we just presented.When we declare the event, we just need write the code like

   1:   event EventHandler<SendRequestCompletedEventArgs> Complete;

   2:   event EventHandler<HttpRequestCompletedEventArgs> RequestComplete;

   3:   event EventHandler<HttpRequestErrordEventArgs> RequestComplete;

   4:   event EventHandler<SignatureRequiredEventArgs> SignatureRequired;

When we register the event, the code should be changed to

channel.Complete +=new EventHandler<SendRequestCompletedEventArgs>(channel_Complete);

There are some limitation here.

This generic even only can only replace the delegate which has two parameter , and the first parameter is object and second one is inherited from the EventArgs, basically the generic event only works well when we are going to use the standard event pattern.
This generic event can only replace the delegate which the return value is void.