Benoît Sarie's blog

Blog de l’équipe MCS APO

Benoît Sarie — Tue, 22 May 2012 09:25:10 GMT

Le nouveau blog de mon équipe est disponible à l’adresse:

http://blogs.msdn.com/b/frmcs-dev

Vous retrouverez des articles sur les sujets de développement suivants:

- Windows 8 Metro Application, Windows Phone 7, ASP.NET, WPF, SOA, ALM, TFS, …

Voici les autres blogs des équipes MCS France :

Bonne lecture,

Benoît

Exemple d'automatisation de livraison d'une application WPF avec MSBuild

Benoît Sarie — Tue, 15 May 2012 12:38:48 GMT

Je travaille depuis quelque temps sur un projet WPF. Mon client nous met à disposition un automate d’intégration continu sur lequel nous n’avons pas la main en terme de customisation.

Ce dernier permet d’effectuer des étapes importantes telle que l’exécution de tests unitaires, couverture de code, vérification de la qualité de code, … cependant certaines étapes à mon sens sont manquantes.

Dans cet article je vais vous présenter un exemple de processus automatisé que j’utilise en complément qui me permet de sécuriser la livraison de l’application en environnement de test.

Je pars du postulat que sur mon poste de développement, seulement le framework.net et Visual Studio 2010 sont installés.

L’idée est de créer ce script le plus vite possible avec les outils en ma possession

=>Je vais donc créer un script MSBuild

Définition des étapes de processus de livraison

Le script est structuré en sept étapes:

GetLatestVersion;
IncrementVersion;
Clean;
Compil;
Deploy;
Check;
CheckinCode;

GetLatestVersion / CheckInCode

Ces targets permettent de réccupérer la dernière version du code et de faire un check-in du code modifié par ce processus. A ce niveau là tout dépend de votre gestionnaire de configuration.

Si vous utilisez Microsoft Team Foundation Server les Tasks sont disponibleq dans Microsoft.TeamFoundation.Build.targets (C:\Program Files (x86)\MSBuild\Microsoft\VisualStudio\TeamBuild pour une machine x64).

Sinon, pour les autres gestionnaires de configuration, il suffit d’obtenir des tasks tier:

MSBuild Community Tasks : http://msbuildtasks.tigris.org/
MSBuid Extension Pack : http://msbuildextensionpack.codeplex.com/

IncrementVersion

Incrémente automatiquement le numéro de version des composants de la solution;

Soit une Custom Task peut être créée, soit les MSBuild Community Tasks offrent une tâche qui permet de modifier le fichier Assemblyinfo.

Clean

Nettoyage du bin des projets de la solutions. Tous les répertoires bin/debug et bin/release sont nettoyés avec la task suivante:

Code Snippet
 <Target Name="Clean">
  <MSBuild Projects="$(SolutionName)" Targets="Clean" />
</Target>
 

Compil

Cette simple target compile la solution dans la configuration définie:

Code Snippet
  <Target Name="Compil">
    <MSBuild Projects="$(SolutionName)" Targets="Rebuild" Properties="Configuration=$(Configuration)" />
  </Target>
 

Deploy

La target Deploy copie les fichiers, présents dans le projet de démarrage de la solution, dans le répertoire de destination. Si le répertoire n’existe pas il est créé.

Code Snippet
  
<Target Name="Deploy">
  <Copy DestinationFolder="$(DestinationFolder)" SourceFiles="@(ExeFiles)" />
  <Copy DestinationFolder="$(DestinationFolder)" SourceFiles="@(DllFiles)" />
  <Copy DestinationFolder="$(DestinationFolder)" SourceFiles="@(PdbFiles)" />
  <Copy DestinationFolder="$(DestinationFolder)" SourceFiles="@(ConfigFiles)" />
  <Copy DestinationFolder="$(DestinationFolder)" SourceFiles="@(XmlFiles)" />
  <Copy DestinationFolder="$(DestinationFolder)" SourceFiles="@(IconesFiles)" />
</Target>
 

Check

A ce niveau là, j’ai créé une custom task qui, à partir d’un fichier Xml, vérifie les fichiers déposés dans le répertoire de livraison.

Cette custom task vérifie:

le nombre de binaires présents;
la version de ces binaires;

Dans la artie suivante de ce post, je vais rentrer en détail sur la façon de faire une Custom Task MSBuild.

Voici les grandes lignes du script MSBuild =>MyBuild.proj

Code Snippet
 <Project xmlns="http://schemas.microsoft.com/developer/msbuild/2003" DefaultTargets="START">
 
  <ItemGroup>
    <SubSteps Include="GetLatestVersionCode" />
    <SubSteps Include="IncrementVersion" />
    <SubSteps Include="Clean" />
    <SubSteps Include="Compil" />
    <SubSteps Include="Deploy" />
    <SubSteps Include="Check" />
    <SubSteps Include="CheckInCode" />
  </ItemGroup>
 
  <ItemGroup>
    <SubCheckSteps Include="CheckDllVersion" />
  </ItemGroup>
  
  <PropertyGroup>
    <Version Condition="'$(Version)' == ''" >1.0.0.23</Version>
    <SolutionName Condition="'$(SolutionName)' == ''" >WPFApplication.sln</SolutionName>
    <Configuration Condition="'$(Configuration)' == ''" >Release</Configuration>
    <OutputFolder Condition="'$(OutputFolder)' == ''" >$(MSBuildProjectDirectory)\WPFApplication\bin\$(Configuration)</OutputFolder>
    <CurrentVersion Condition="'$(CurrentVersion)' == ''" >$(ApplicationPrefix)_V$(Version)_$([System.DateTime]::Now.ToString('yy-MM-dd'))</CurrentVersion>
    <VersionFile Condition="'$(VersionFile)' == ''" >$(MSBuildProjectDirectory)\iCRM.Framework\Application\ProjectInfo.cs</VersionFile>
    <DestinationRootFolder Condition="'$(DestinationRootFolder)' == ''" >c:\Livraison</DestinationRootFolder>
    <DestinationFolder Condition="'$(DestinationFolder)' == ''" >$(DestinationRootFolder)\$(CurrentVersion)</DestinationFolder>
    <TestFile Condition="'$(TestFile)' == ''">$(MSBuildProjectDirectory)\CheckVersion.xml</TestFile>
  </PropertyGroup>
 
  <ItemGroup>
    <ExeFiles Include="$(iCRMOutputFolder)\*.exe" />
    <DllFiles Include="$(iCRMOutputFolder)\*.dll" />
    <PdbFiles Include="$(iCRMOutputFolder)\*.pdb" />
    <ConfigFiles Include="$(iCRMOutputFolder)\*.config" />
    <XmlFiles Include="$(iCRMOutputFolder)\*.xml" />
    <IconesFiles Include="$(iCRMOutputFolder)\*.ico" />
  </ItemGroup>
  
  <Target Name="START" DependsOnTargets="@(SubSteps)">
 
  </Target>
 
  <Target Name="GetLatestVersionCode">
      <!-- Task to get latestet version of code -->
  </Target>
 
  <Target Name="IncrementVersion">
    <!-- Incrementation of assembly version -->
  </Target>
 
  <Target Name="Clean">
    <MSBuild Projects="$(SolutionName)" Targets="Clean" />
  </Target>
 
  <Target Name="Compil">
    <MSBuild Projects="$(SolutionName)" Targets="Rebuild" Properties="Configuration=$(Configuration)" />
  </Target>
 
  <Target Name="Deploy">
    <Message Text="Config" />
    <Message Text="$(OutputFolder)\*.exe" />
    <Message Text="@(ExeFiles)" />
    <Copy DestinationFolder="$(DestinationFolder)" SourceFiles="@(ExeFiles)" />
    <Copy DestinationFolder="$(DestinationFolder)" SourceFiles="@(DllFiles)" />
    <Copy DestinationFolder="$(DestinationFolder)" SourceFiles="@(PdbFiles)" />
    <Copy DestinationFolder="$(DestinationFolder)" SourceFiles="@(ConfigFiles)" />
    <Copy DestinationFolder="$(DestinationFolder)" SourceFiles="@(XmlFiles)" />
    <Copy DestinationFolder="$(DestinationFolder)" SourceFiles="@(IconesFiles)" />
  </Target>
 
  <Target Name="Check">
    <!-- Task to check deployment -->
  </Target>
 
  <Target Name="CheckInCode">
    <!-- Task to check-in code -->
  </Target>
 
 
</Project>
 

Dans ce script, deux patterns MSBuild sont vraiment intéressants:

l’utilisation de DependsOnTarget pour permettre une meilleure granularité de target. Comme pour du code .net, je met le minimum de logique dans une target => 1 target = 1 rôle;
la pré-initialisation des properties au début du script afin d’avoir lors de l’exécution du script MSBuild un minimum de paramettres à saisir.

Execution du script

Pour démarrer ce script, ouvrez le Visual Studio Command Prompt (2010) pour avoir MSBuild dans le path.

Naviguez jusqu’à votre fichier proj MyBuild.proj, lancez la ligne de commande suivante:

msbuild MyBuild.proj /p:Version=1.0.0.24

Cette ligne de commande lance le processus pour la version 1.0.0.24, la compilation est faite en release.

Exemple de Custom Task MSBuild

Créez un projet de classe de librairie nommé par exemple MyCustomTaskMSBuild.

Ajouter les références aux dlls Microsoft.Build.Framework.dll:

Renommez la classe Class1.cs en MyCustomTask.cs et faites la hériter de l’interface ITask.

Ensuite, à vous de jouer, après avoir surchargé les méthodes/propriétés de ITask, vous pouvez créer un code celui-ci:

Code Snippet
 using Microsoft.Build.Framework;
 
namespace MyCustomTaskMSBuild
{
    public class MyCustomTask : ITask
    {
 
        public IBuildEngine BuildEngine
        {
            get; set;
        }
        public ITaskHost HostObject
        {
            get; set;
        }
 
        [Required]
        public string MyInputParam { get; set; }
 
        public bool Execute()
        {
            // Add task logic here
            Trace(string.Format("My First Task - {0}",MyInputParam), MessageImportance.Normal);
            return true;
        }
 
        protected void Trace(string message, MessageImportance importance)
        {
            BuildMessageEventArgs args = new BuildMessageEventArgs(message, string.Empty, TaskName, importance);
            this.BuildEngine.LogMessageEvent(args);
        }
 
        public string TaskName { get { return "My Custom Task"; } }
    }
}
 

Pour tout ce qui est logging de messages, utilisez la méthode Trace.

J’ai rajouté un MyInputParam, avec l’attribut Required il devient obligatoire lors de l’exécution du script MSBuild.

Comment insérer cette nouvelle Task dans mon fichier proj

En entête du fichier proj, rajoutez la ligne suivante:

Code Snippet
 <UsingTask AssemblyFile="MyCustomTaskMSBuild.dll" TaskName="MyCustomTaskMSBuild.MyCustomTask" />
 

Cette ligne défini le lien avec la Custom Task, il suffit ensuite d’appeller cette Task dans une target comme suit:

Code Snippet
  
<Target Name="test">
  <MyCustomTask MyInputParam="!!!" />
</Target>
 

Et maintenant, testez:

Ensuite, en vie réelle, vérifiez que le ficher proj est dans le même répertoire que la dll de la Custom Task.

Comment debugger cette Custom Task

Pour debugger cetteTask, forcez le projet MyCustomTaskMSBuild à Startup Project.

Etre sûr que le fichier proj est bien dans le répertoire bin\debug du projet MyCustomTaskMSBuild.

Dans l’onglet Debug des propriétés de ce projet. Sélectionner Start exsternal program et pointez vers l’exécutable MSBuild, dans mon cas C:\Windows\Microsoft.NET\Framework\v4.0.30319\MSBuild.exe

Dans les Command Line Arguments, il suffit de mettre le nom du fichier MSBuild avec les bons paramètres.

F5 et le debug peut commencer.

Quelques liens

Documentation MSBuild sur MSDN :

Blog de l’équipe MSBuild

http://blogs.msdn.com/b/msbuild/

Benoît

Team Foundation Server 2010 tools & links

Benoît Sarie — Mon, 13 Jun 2011 13:37:04 GMT

Here are some interesting links on TFS 2010.

Tools that you can add around TFS 2010

Check-in policy

TFS 2010 Time Tracking Checkin Policy
http://visualstudiogallery.msdn.microsoft.com/58ec8802-bd48-4673-a465-4ba099cfbce1

As Project Manager, you need to track the time spent on a specific task, this check-in policy is for you.

Build

Inmeta.VisualStudio.TeamExplorer.BuildExplorer
http://visualstudiogallery.msdn.microsoft.com/35daa606-4917-43c4-98ab-38632d9dbd45

Your TFS project contains so many builds, this tool enable you to create folders in the Build Explorer.

Team Foundation Server Build Extensions Power Tool March 2011
http://visualstudiogallery.msdn.microsoft.com/2d7c8577-54b8-47ce-82a5-8649f579dcb6

You need to execute Maven 2 and Ant builds on your build server. Takes advantage of JUnit tests results.

Templates

Microsoft Visual Studio Scrum 1.0
http://visualstudiogallery.msdn.microsoft.com/59ac03e3-df99-4776-be39-1917cbfc5d8e

New process template : Scrum.

Specs & Modeling

Spec Explorer 2010 Visual Studio Power Tool
http://visualstudiogallery.msdn.microsoft.com/271d0904-f178-4ce9-956b-d9bfa4902745?SRC=Featured

Spec Explorer is a visual tool for modeling software behavior and generating test suites from those models

UML Extension For Spec Explorer 2010
http://visualstudiogallery.msdn.microsoft.com/ce73da2a-072f-44d0-ae18-600213b56520

Extension of the Spec Explorer, adding sequence diagram.

Visualization and Modeling Feature Pack Runtime
http://visualstudiogallery.msdn.microsoft.com/93ba7cc9-e145-453f-a269-d98e0576b961

Tests

Visual Studio 2010 Lab Management Feature Pack
http://msdn.microsoft.com/en-us/vstudio/ee712698
Microsoft Visual Studio 2010 Load Test Feature Pack
http://msdn.microsoft.com/en-us/library/gg701769.aspx

Integration

Microsoft Team Foundation Server 2010 and Microsoft Project Server Integration Feature Pack
http://msdn.microsoft.com/en-us/library/gg455680.aspx

Installation & connection

Microsoft Team Foundation Server 2010 MOSS Configuration Tool
http://visualstudiogallery.msdn.microsoft.com/db469790-5e3e-42f3-906e-411a73795a1b
Team Foundation Server Integration Tools (March 2011 Release)
http://visualstudiogallery.msdn.microsoft.com/f854dd59-8eeb-4673-8d9a-ae012989bfa2
Team Foundation Server MSSCCI Provider 2010
http://visualstudiogallery.msdn.microsoft.com/bce06506-be38-47a1-9f29-d3937d3d88d6
Team Foundation Server Power Tools March 2011
http://visualstudiogallery.msdn.microsoft.com/c255a1e4-04ba-4f68-8f4e-cd473d6b971f

Articles & Guides

Visual Studio TFS Branching and Merging Guidance
http://msdn.microsoft.com/en-us/magazine/gg598921.aspx
Visual Studio TFS Team Project and Collection Guidance
http://msdn.microsoft.com/en-us/magazine/gg983486.aspx
Visual Studio TFS Branching Guide 2010
http://tfsbranchingguideiii.codeplex.com/
Visual Studio TFS Upgrade Guide
http://vs2010upgradeguide.codeplex.com/

Blogs

Some interesting blogs:

My practices on FetchXML (MS Dynamics CRM 4.0)

Benoît Sarie — Thu, 09 Jun 2011 14:04:59 GMT

Since few months, I work on developing a WPF UI that directly connects to MS Dynamics CRM 4.0. In this post I will present you some of the practices I use in this project.

Restriction of the result

Reduce the attributes to return

Specify all desired attributes and not all-attributes element. all-attributes will return all fields of the desired entities:

<attribute name=”…” />

<attribute name=”…” />

instead of:

<all-attributes />

I really recommend you to manually set all the attributes you need. This will reduce the size of the response and in a same time, this reduce the SQL request generated by CRM.

Code Snippet
 <fetch distinct="true" mapping="logical">
  <entity name="contact">
    <attribute name="contactid" />
    <attribute name="gendercode" />
    <attribute name="firstname" />
    <attribute name="lastname" />
  </entity>
</fetch>
 

Pagination with Paging cookie

Take few minutes to create your own paging mechanism.

Step 1 :

Create a context object for the server calls:

Code Snippet
 /// <summary>
/// Context object to keep between all crm calls
/// </summary>
public class PaginationContext
{
    /// <summary>
    /// Id of the page
    /// </summary>
    public int Page { get; set; }
    /// <summary>
    /// Number of element to return
    /// </summary>
    public int ElementCount { get; set; }
    /// <summary>
    /// Paging cookie
    /// </summary>
    public string PagingCookie { get; set; }
    /// <summary>
    /// HAs more records to return
    /// </summary>
    public bool HasMoreRecords { get; set; }
}
 

Encapsulate the query encapsulation and parse the return of the FetchXML query:

Code Snippet
 /// <summary>
/// Generic mechanism to paginate
/// </summary>
/// <param name="query"></param>
/// <param name="context"></param>
/// <param name="paginationcontext"></param>
/// <returns></returns>
public static XElement PaginateResults(this StringBuilder query,
    CrmDataContext context,
    PaginationContext paginationcontext)
{
 
    XElement result = null;
    XDocument fetchQueryDoc = XDocument.Parse(query.ToString());
    XElement rootQueryElement = fetchQueryDoc.Element("fetch");
 
    // Initialize the number od records to return
    XAttribute countAttribute = rootQueryElement.Attribute("count");
    if (rootQueryElement.Attribute("count") != null)
        countAttribute.Value = paginationcontext.ElementCount.ToString();
    else
        rootQueryElement.Add(new XAttribute("count", paginationcontext.ElementCount.ToString()));
 
    // Initialize the page id
    XAttribute pageAttribute = rootQueryElement.Attribute("page");
    if (rootQueryElement.Attribute("page") != null)
        pageAttribute.Value = paginationcontext.Page.ToString();
    else
        rootQueryElement.Add(new XAttribute("page", paginationcontext.Page.ToString()));
 
    // Initialize the paging cookie
    XAttribute pagingCookieAttribute = rootQueryElement.Attribute("paging-cookie");
    if (rootQueryElement.Attribute("paging-cookie") != null)
        pagingCookieAttribute.Value = paginationcontext.PagingCookie.ToString();
    else
        rootQueryElement.Add(new XAttribute("paging-cookie", paginationcontext.PagingCookie.ToString()));
 
    using (var service = context.CreateService())
    {
        string fetchResult = service.Fetch(fetchQueryDoc.ToString());
        XDocument resultDoc = XDocument.Parse(fetchResult);
        if (resultDoc != null)
        {
            XElement rootElement = resultDoc.Element("resultset");
            if (rootElement.Attribute("morerecords") != null && rootElement.Attribute("morerecords").Value == "1")
                paginationcontext.HasMoreRecords = true;
            else
                paginationcontext.HasMoreRecords = false;
            if (rootElement.Attribute("paging-cookie") != null && string.IsNullOrEmpty(rootElement.Attribute("paging-cookie").Value) == false)
            {
                string pagingCookie = rootElement.Attribute("paging-cookie").ToString().TrimStart("paging-cookie=\"".ToCharArray()).TrimEnd('"');
            }
            result = rootElement;
        }
    }
    return result;
 
}
 

Test the code:

Code Snippet
 [TestMethod]
public void TestPagination()
{
    CrmDataContext context;
    // ...
    // Code to connect to CRM
    // ...
    using (var service = context.CreateService())
    {
        FetchXMLHelper.PaginationContext paginationContext = new FetchXMLHelper.PaginationContext() { ElementCount = 50, HasMoreRecords = true };
        do
        {
            StringBuilder fetchXml = new StringBuilder();
            fetchXml.Append("<fetch distinct=\"false\" mapping=\"logical\" >");
            fetchXml.Append("<entity name=\"account\">");
            fetchXml.Append("   <all-attributes />");
            fetchXml.Append("</entity>");
            fetchXml.Append("</fetch>");
 
            XElement element = fetchXml.PaginateResults(context, paginationContext);
            if (element != null)
            {
                // ...
                // Code to deserialize the result of the fetch query
          ��     // ...
            }
        } while (paginationContext.HasMoreRecords);
 
    }
}
 

Joining entities

FetchXML offers the ability to return fields from parent entity, linq queries not.

Using aliases

Using aliases for you linked entities will facilitate the execution and your deserialization. If you do not set an alias, you will not be able to return fields with same names from different entities.

Code Snippet
 <fetch distinct="true" mapping="logical">
  <entity name="contact">
    <attribute name="contactid" />
    <attribute name="gendercode" />
    <attribute name="firstname" />
    <attribute name="lastname" />
    <attribute name="createdby" />
    <attribute name="createdon" />
    <attribute name="modifiedby" />
    <attribute name="modifiedon" />
    <filter type="and">
      ...      
    </filter>
      <link-entity name="account" from="accountid" to="parentcustomerid"alias="accountAlias">
      <attribute name="accountid" />
      ...
      <attribute name="createdby" />
      <attribute name="createdon" />
      <attribute name="modifiedby" />
      <attribute name="modifiedon" />
    </link-entity>
  </entity>
</fetch>
 

Here is the result of previous FetchXML queyry:

Code Snippet
 <resultset>
  <result>
    <contactid>...</contactid>
    <gendercode>...</gendercode>
    <firstname>...</firstname>
    <lastname>...</lastname>
    <createdby>...</createdby>
    <createdon>...</createdon>
    <modifiedby>...</modifiedby>
    <modifiedon>...</modifiedon>
    <accountAlias.accountid>...</accountAlias.accountid>
    ...
    <accountAlias.createdby>...</accountAlias.createdby>
    <accountAlias.createdon>...</accountAlias.createdon>
    <accountAlias.modifiedby>...</accountAlias.modifiedby>
    <accountAlias.modifiedon>...</accountAlias.modifiedon>
  </result>
  ...
</resultset>
 

In the example, you can now returns createdby (createdon, modifiedby, modifiedon) on two different entities. After in your project, you can easily create you own deserialization mechanism having the alias as parameter.

Left Join

You cannot use left join using Linq. Fetch queries enable left join. This helps you a lot, if you need to return in each contact query specific fields of parent entity, this reduces the number of request to server.

Code Snippet
 <fetch distinct="true" mapping="logical">
  <entity name="contact">
    <attribute name="contactid" />
    <attribute name="gendercode" />
    <attribute name="firstname" />
    <attribute name="lastname" />
    <filter type="and">
      ...
    </filter>
    <link-entity name="account" from="accountid" to="parentcustomerid" alias="accountAlias" link-type="outer">
      <attribute name="accountid" />
      ...
    </link-entity>
  </entity>
</fetch>
 

Quick generation of fetch XML query

Here is in 5 points a quick way to generate your FetchXML query:

Use the advance find of MS CRM;
Use the graphical user interface to define your query;
Run the query (click on Find button) ;
Press CTRL+N, this will create a new IE window but will keep the context;
Replace the url in the address bar
- javascript:alert(resultRender.FetchXml.value);

This will open a pop-up with the expected FetchXml

Book in French: Windows 7 Avancé

Benoît Sarie — Wed, 02 Mar 2011 10:20:35 GMT

One of my collegue has just shipped the second edition of his book on Windows 7.

I really recommand you to read it ;-).

It is available at http://www.amazon.fr/Windows-7-avanc%C3%A9-Thomas-Garcia/dp/2212129955/ref=sr_1_2?ie=UTF8&s=books&qid=1299059442&sr=8-2

You can also follow him at on his blog http://blogs.msdn.com/b/lgmorand/

Debugging/profiling/optimizing a WPF application

Benoît Sarie — Fri, 18 Feb 2011 08:32:19 GMT

Mots clés Technorati : WPF,Debugging,Profiling,Optimizing

In this post I will describe and redirect you to good articles or good tools for.

I will talk about four main points:

Debugging bindings;
Checking your visual tree;
Profiling your WPF application;
Optimizing your WPF application;

Data bindings

Here are two very important posts that explicit different solutions for debugging bindings:

- This post from Bea Stollnitz;

- This post from Karl Shifflet;

After reading those post, in WPF projects I use all those techniques in this order:

I read the Visual Studio Output Window to show. In most cases, the message is enough explicit.
Usage of the DebugConverter: here is the implementation of the DebugConverter I use in projects.

Code Snippet
 public class DebuggingConverter : IValueConverter
  {
      public object Convert(object value, Type targetType, object parameter, System.Globalization.CultureInfo culture)
      {
          System.Diagnostics.Debugger.Break();
          return value;
      }
 
 
      public object ConvertBack(object value, Type targetType, object parameter, System.Globalization.CultureInfo culture)
      {
          System.Diagnostics.Debugger.Break();
          return value;
      }
  }
 

I define it in the resources of the application, and it’s done. I just have to call this converter on the binding where I encounter an issue. Automatically the debugger breaks in my converter. After solving the binding issue, I suppress the converter and that’s all.

I don’t really use the trace sources because most of times, I already found the issue before using it

Visual Tree debugging

In a previous post, I explained you my favorite tools, the two I use quickly to debug the visual tree of my application are Mole and Snoop.

Mole is useful when you set a breakpoint in Visual Studio and you want to debug the visual/logical tree of your application. It’s integrated into Visual Studio, when in your code you set a breakpoint, you can easily open Mole to understand what is happening in the visual or in the logical tree. I use Mole in Visual Studio, when debugging. I’m able to quickly find what’s happening –> incorrect owner, style not applied correctly, …

Snoop is different tool, also displays the visual tree of a WPF application, it is not working within Visual Studio. It takes the applications at a specific moment. It also displays in a tree control the visual tree to easily navigate into the elements but also allows to split the application in a 3D representation of all visual elements to better understand the graphical composition of the application. This allows to to better understand why part of your application is not docked correctly or stretched correctly.

Profiling tools

Performance Profiling Tools for Windows Presentation Foundation (also know as WPFPerf) is available in the Microsoft Windows SDK for Windows 7 and the .NET Framework 4.

WPFPerf contains 2 tools :

Perforator : analyze rendering behavior, this is used to look for rendering issues;
Visual Profiler : analyze the use of WPF services to determine which visual element if causing a performance bottleneck.

I really use them when I really want to understand what is happening on a UI point of view, how the visual tree is build, how all the events/animations are

If you need more documentation on those tools, I redirect you to WPF Performance Suite on MSDN .

Optimization reference

Here are my two main resources for optimization

MSDN Article : Optimizing WPF Application Performance.
WPF Performance and .NET Framework Client Profile blog : this blog gives a lot of tips for WPF optimizations.

Do not forget that WPF is also a part of .NET, so also use the Improving .NET Application Performance and Scalability form Pattern & Practice.

Conclusion

In this post I summarized some actions to do during a development life cycle on a WPF project.

And of course, you can follow a TechDays 2011 Session, from one of my colleague Luc Vo Van on WPF/Silverlight performance tips & tricks.

Visual Studio 2010 + sos.dll: finding memory leaks

Benoît Sarie — Mon, 17 Jan 2011 19:38:31 GMT

During development phase of my current project I try to figure out if my application has memory leaks. The client where I am working prevent me from installing software on my development computer. I don’t have installed Microsoft Windows SDK, so I cannot use WinDBG. This article is a quick post to introduce Visual Studio 2010 and sos.dll

Project configuration

Before starting your application, on the project properties> Debug tab, make sure that “Enable unmanaged code debugging” is enabled.

Starting analyzing

Start your debugging your application, execute the scenario where you found that there is a memory leak, break at the end of this execution (Break All or CTRL+ALT+DEL).

In Visual Studio, Open the immediate windows and load sos.dll.

For .net 4.0 application:

.load C:\windows\Microsoft.NET\Framework\v4.0.30319\sos.dll

or

.load sos

For .net 2.0, 3.0 or 3.5 application:

.load C:\windows\Microsoft.NET\Framework\v2.0.50727\sos.dll

Enter the following command line to get the summary of object types that are still in memory:

!dumpheap –stat

If one of the type should not be in memory, you can filter for a specific type:

!dump –type MyNamespace.MyType

This lists all objects of this specific type in the heap. Once you have the address of one specific object you can get a dump of a specific object:

!gcroot 02369d30

This dump lists all the objects that references, and that forces the object to be kept in the heap. After you should investigate deeper

Main commands to use

Command	Comment
!dumpheap –stat	Summary of classes
MyNamespace.MyType	Lists all objects in the heap of this specific type
!gcroot 02369d30	Returns the ump of a specific object

Interesting links

MSDN reference : http://msdn.microsoft.com/fr-fr/library/bb190764.aspx

Tess Ferrandez blog: http://blogs.msdn.com/b/tess/ (very interesting blog). She wrote several good post but also one dedicated on “Debugging Silverlight applications with windbg and sos.dll”

My developer toolbox

Benoît Sarie — Mon, 17 Jan 2011 18:29:00 GMT

Here a post to list my 10 more interesting tools that I use very often in my everyday job. Visual Studio and Expression Blend are the tools I use all day long, the goal of this post is to give you some interesting tools.

.net development

Visual Studio Gallery lists many interesting tools:

Productivity Power Tools
Power Commands for Visual studio
Team Foundation
Regex Editor
tangible T4 Editor plus modeling tools for VS2010
…

Reflector is the tool used, as written on the red-gate site, “Explore and analyze .NET assemblies”.

A lot of add-ins exists –>Reflector Add-Ins. One I always use is Baml Viewer. It loads and render the BAML in XAML.

PInvoke.Net is in a same time a reference web site but also tool helping into interaction with com dll.

LinqPad helps you to create Linq queries (Linq to Objects, Linq to XML, Entity Framework, …) but also you can also write code and it’s instantly executed.

WPF development

Snoop helps you in your WPF developments, it displays the visual tree of your application (you also can see it in 3D). I use it to understand why a specific WPF control is not displayed correctly.

Mole (available for Visual Studio 2008) is a WPF visualizer, very useful when debugging your application and more precisely WPF visual tree or logical tree.

Http debugger

Fiddler is a web debugging proxy.

File and Disk utilities

When do not know why a file is already used by another process or what’s happening on your system Process Monitor is the tool. It traces in real time file, registry and process activity.

Memory debugger

WinDBG is available in the Microsoft Windows SDK ( Seven version, for other version of OS ). It helps you into finding memory leaks.

Visual Studio + sos.dll (sos.dll is i) is also an other way to find memory leaks.

Unity 2 – Configuration file and Unity (Part 2)

Benoît Sarie — Sat, 18 Sep 2010 12:31:15 GMT

I’m real fan of Unity, I already worked with other Ioc, but Unity remains my favorite.

In my previous post, I explained how to configure Unity at runtime. This one will describe how to use the configuration file.

First of all add the following references to your project:

System.Configuration.dll
Microsoft.Practices.Unity.Configuration.dll

Simple scenario

Add a using directive to get easy access to the container initializer:

Code Snippet
 using Microsoft.Practices.Unity.Configuration;
 

Update the Program class as follows:

Code Snippet
 class Program
{
    static void Main(string[] args)
    {
        IUnityContainer container;
        container = new UnityContainer();
 
        // Initialize the container with the config file
        container.LoadConfiguration();
        // Resolve IMyComponent
        IMyComponent obj = container.Resolve<IMyComponent>();
 
        Console.WriteLine(obj.Run(5));
 
        Console.ReadKey();
    }
}
 

Add a configuration file (which is a simple XML file):

Code Snippet
 <?xml version="1.0" encoding="utf-8" ?>
<configuration>
    <configSections>
        <section name="unity" type="Microsoft.Practices.Unity.Configuration.UnityConfigurationSection, Microsoft.Practices.Unity.Configuration"/>
    </configSections>
    <unity xmlns="http://schemas.microsoft.com/practices/2010/unity">
        <container>
            <register type="UnityTest.IMyComponent, UnityTest" mapTo="UnityTest.MyComponent, UnityTest"  />
        </container>
    </unity>
 
</configuration>
 

Run the example and the result is the same as in my previous post.

Singleton scenario

To use the injected object as a singleton, just define a ControledContainerLifetimeManager, which manages the lifetime of your object as a unique instance:

Code Snippet
 <?xml version="1.0" encoding="utf-8" ?>
<configuration>
    <configSections>
        <section name="unity" type="Microsoft.Practices.Unity.Configuration.UnityConfigurationSection, Microsoft.Practices.Unity.Configuration"/>
    </configSections>
    <unity xmlns="http://schemas.microsoft.com/practices/2010/unity">
        <container>
            <register type="UnityTest.IMyComponent, UnityTest" mapTo="UnityTest.MyComponent, UnityTest"  >
                <lifetime type="Microsoft.Practices.Unity.ContainerControlledLifetimeManager, Microsoft.Practices.Unity" />
            </register>
        </container>
    </unity>
 
</configuration>
 

Alias scenario

For the alias scenario, remove the previous configuration file, and change the code as follows :

Code Snippet
 class Program
{
    static void Main(string[] args)
    {
        IUnityContainer container;
        const string FRENCH = "fr";
        const string DEFAULT = "us";
        // Use the runtime configuration
        container = new UnityContainer();
 
        // Initialize the container with the config file
        container.LoadConfiguration();
 
        // Resolve IMyComponent
        IMyComponent obj = container.Resolve<IMyComponent>(DEFAULT);
        IMyComponent objFr = container.Resolve<IMyComponent>(FRENCH);
 
        Console.WriteLine(obj.Run(5));
        Console.WriteLine(objFr.Run(5));
 
        Console.ReadKey();
    }
}
 

Then use the following configuration file instead, which defines the aliases with the name attribute:

Code Snippet
 <?xml version="1.0" encoding="utf-8" ?>
<configuration>
    <configSections>
        <section name="unity" type="Microsoft.Practices.Unity.Configuration.UnityConfigurationSection, Microsoft.Practices.Unity.Configuration"/>
    </configSections>
    <unity xmlns="http://schemas.microsoft.com/practices/2010/unity">
        <container>
            <register type="UnityTest.IMyComponent, UnityTest" mapTo="UnityTest.MyComponent, UnityTest" name="us" >
                <lifetime type="Microsoft.Practices.Unity.ContainerControlledLifetimeManager, Microsoft.Practices.Unity" />
            </register>
            <register type="UnityTest.IMyComponent, UnityTest" mapTo="UnityTest.MyComponentFr, UnityTest" name="fr" >
                <lifetime type="Microsoft.Practices.Unity.ContainerControlledLifetimeManager, Microsoft.Practices.Unity" />
            </register>
        </container>
    </unity>
 
</configuration>
 

Unity 2 – Configuring Unity at Runtime (Part1)

Benoît Sarie — Sat, 18 Sep 2010 12:27:05 GMT

I’m real fan of Unity, I already worked with other IoCs, but Unity remains my favorite.

In the upcoming posts, I will explain you how to :

Configure in Runtime or in Design time;
Make the best use of it

In this post I will assume that you already know about ALL the advantages of using the IoC pattern. You can download Unity here.

Simple scenario

First of all, we will create the project that will help us for all the samples. I will name it UnityTest.

First and foremost, let’s add a reference to Microsoft.Practices.Unity.

Then create an interface named IMyComponent:

Code Snippet
 public interface IMyComponent
{
    string Run(int i);
}
 

Implement that interface in the MyComponent class:

Code Snippet
 public class MyComponent : IMyComponent
{
    #region IMyComponent Members
    int loopIndex = 0;
    public MyComponent() 
    {
        loopIndex = 0;
    }
    public string Run(int nb)
    {
        StringBuilder sb = new StringBuilder();
        sb.Append("Hello : ");
        for(int i=0; i<nb;i++)
        {
            loopIndex = loopIndex +i;
            sb.Append(loopIndex.ToString()).Append(" ");
        }
        return sb.ToString();
    }
 
    #endregion
}
 

Finally, implement Program as follows (register, then resolve the types):

Code Snippet
 class Program
{
    static void Main(string[] args)
    {
        IUnityContainer container;
        // Use the runtime configuration
        container = new UnityContainer();
        container.RegisterType<IMyComponent,MyComponent>();
 
        // Resolve IMyComponent
        IMyComponent obj = container.Resolve<IMyComponent>();
        IMyComponent objBis = container.Resolve<IMyComponent>();
 
        Console.WriteLine(obj.Run(5));
        Console.WriteLine(objBis.Run(5));
        Console.ReadKey();
 
    }
}
 

What was actually done, was to instantiate the Unity container, and then register a new couple inside it. From then on, every time I will need an implementation of IMyComponent, Unity will instantiate MyComponent through a call to Resolve.

Singleton scenario

If I want to have MyComponent as a singleton, I will have to register it as an instance:

Code Snippet
 container.RegisterInstance<IMyComponent>(new MyComponent());
 

Then, each time I will need an instance of IMyComponent, Unity will return me the registered singleton.

Code Snippet
 class Program
{
    static void Main(string[] args)
    {
        IUnityContainer container;
        // Use the runtime configuration
        container = new UnityContainer();
        container.RegisterInstance<IMyComponent>(new MyComponent());
 
        // Resolve IMyComponent
        IMyComponent obj = container.Resolve<IMyComponent>();
        IMyComponent objBis = container.Resolve<IMyComponent>();
 
        Console.WriteLine(obj.Run(5));
        Console.WriteLine(objBis.Run(5));
        Console.ReadKey();
 
    }
}
 

Aliases scenario

I will create a new class MyComponentFr that will also implement the IMyComponent interface.

Code Snippet
 public class MyComponentFr : IMyComponent
{
    #region IMyComponent Members
    int loopIndex = 0;
    public MyComponentFr()
    {
        loopIndex = 0;
    }
    public string Run(int nb)
    {
        StringBuilder sb = new StringBuilder();
        sb.Append("Bonjour : ");
        for (int i = 0; i < nb; i++)
        {
            loopIndex = loopIndex + i;
            sb.Append(loopIndex.ToString()).Append(" ");
        }
        return sb.ToString();
    }
 
    #endregion
}
 

This is exactly the same code as the initial MyComponent except that the text is in French. The following code demonstrates how to use aliases to return different implementations of IMyComponent, depending on an alias :

Code Snippet
 class Program
{
    static void Main(string[] args)
    {
        IUnityContainer container;
        const string FRENCH = "fr";
        const string DEFAULT = "us";
        // Use the runtime configuration
        container = new UnityContainer();
        container.RegisterType<IMyComponent, MyComponent>(DEFAULT);
        container.RegisterType<IMyComponent, MyComponentFr>(FRENCH);
        // Configure aliases with singleton
        // container.RegisterInstance<IMyComponent>(DEFAULT, new MyComponent());
 
        // Resolve IMyComponent
        IMyComponent obj = container.Resolve<IMyComponent>(DEFAULT);
        IMyComponent objFr = container.Resolve<IMyComponent>(FRENCH);
 
        Console.WriteLine(obj.Run(5));
        Console.WriteLine(objFr.Run(5));
 
        Console.ReadKey();
    }
}
 

As you can see, I just have to register the implementation with an alias.

Visual Studio Scrum 1.0

Benoît Sarie — Wed, 21 Jul 2010 12:58:00 GMT

A new process template is released, after CMMI and Agile template, here is the Scrum 1.0.

Here are two interesting links on this subject:

- Aaron Bjork annnonced this new process template:

http://blogs.msdn.com/b/aaronbjork/archive/2010/07/19/announcing-microsoft-visual-studio-scrum-1-0.aspx?wa=wsignin1.0

- Brian Harry explained few weeks ago the main lines of this process template:

http://blogs.msdn.com/b/bharry/archive/2010/06/07/a-scrum-process-template-for-tfs.aspx

Visual Studio 2010 : How to deploy a Web application with Web Application Project Deployment (Part 3)

Benoît Sarie — Thu, 03 Jun 2010 10:33:43 GMT

In all project it’s a real nightmare to manage configuration files. In Silverlight projects like in web application, you can take advantage of Web Application Project Deployment features of Visual Studio 2010. In this blog, I’ll explain you :

How to create a new configuration (part 1);
How to create package definition (part 1);
How to create package (part 1);
How to publish the application (part 1);
How to use transformation with web.config files (part 2);
Some more links on this subject (part 2).
Installation of Web Deployment Tool (Part 3).
How to deploy automatically the application (part 3).
How to force the creation of the deployment package in a TFS Build (part 3).
How to deploy automatically the application in a TFS Build (part 3).

In this post, I won’t explain SQL deployment.

Installation of the Web Deployment Tool

The download is available at http://www.iis.net/download/WebDeploy.

You just need to follow the steps to install it.

If you want to install it on a server, you will need to install the “Remote Application Service”.

How to deploy automatically the application

When you generate the deployment package, you have several files.

<MyProject>.deploy.cmd : this script is really important, it contains the script to deplou the application;
<MyProject>.SetParameters.xml : this file contains the parameters of the deployment, for exemple the name of the destination application (IIS)
<MyProject>.zip : the package

All the parameters of the deployment are defined in the web application project. In Visual Studio, open the Properties Panel to update the parameters.

How to force the creation of the deployment package in a TFS Build

In your build definition, when you compile your solution, you should force MSBuild to use this parameter:

/p:DeployOnBuild=true

In this exemple, I’m using the DefaultTemplate.xaml, I just have to set the parameter in the MSBuild Arguments.

Automatically, the build will deploy in the define drop folder the generated package.

How to deploy automatically the application in a TFS Build (part 3).

First of all, you need to install the build server the Web Deployment Tool.

After that, in your build, once the build and the tests are completed, for example in the finally of “Try Compile, Test, and Associate Changesets and Work Items” , you can add the code for the deployment.

Create a new sequence for example called Deploy. Create a new variable called DeployCommandFiles with type IEnumerable<String>.

Find the project to deploy using FindMatchingFile Activity and store the result in DeployCommandFiles :

Create a Loop for each element of DeployCommandFiles and each element is a string (variable commandFile).

Excecute the script using InvokeProcess Activity:

This activity will deploy the package on the server using this command line:

<MyProject>.deploy.cmd /t /m:MyServer

Be care full, only administrator of the destination server can deploy application. You should grant admin privileges on the build service account.

Conclusion

Deployment of web application is really easy with Visual Studio 2010. It’s really helpfull to use TFS Build to deploy a web application.

You can also, when the application is deployed, run automated tests (UI Coded tests) to functionally validate the application.

Visual Studio 2010 : How to deploy a Web application with Web Application Project Deployment (Part 2)

Benoît Sarie — Mon, 10 May 2010 20:16:24 GMT

How to create a new configuration (part 1);
How to create package definition (part 1);
How to create package (part 1);
How to publish the application (part 1);
How to use transformation with web.config files (part 2);
Some more links on this subject (part 2).

In this post, I won’t explain SQL deployment.

How to use transformation with web.config files

In this post, I’ll explain how to use Web.config Transformation Syntax for Web Application Project Deployment. I’ll demonstrate this mechanism with two samples:

transformation for debug configuration;
transformation for staging configuration;

In your web.config file, you will define all parameters that you need in your application. For example, you need a DB connection string named MYDB. You need to define the connectionString add element with a specific name but without any connectionString value. You can also define, for example, compilation target to framework 4.0, debug mode. Finally, you disable custom errors.

Code Snippet
       <?xml version="1.0"?> 
<configuration> 
  <connectionStrings> 
      <add name="MyDB"connectionString="..." /> 
    </connectionStrings> 
    <system.web> 
      <compilation debug="true" targetFramework="4.0" /> 
      <customErrors mode="Off"></customErrors> 
    </system.web> 
   
    <system.webServer> 
      <modules runAllManagedModulesForAllRequests="true"/> 
    </system.webServer> 
  
</configuration> 
     

This Web.Config file will be used/transformed to be compliant with the desired configuration.

Web.Debug.config

In the Web.Debug.Config file, you’ll create some transformation with a specific syntax. In this example, I’ll just add a specific connectionString dedicated to debug environment. In the Add element, you just need to add xdt:Transform and wdt:Locator=”Match(name)”. Those two attributes will inform that the transformation will be applied on an “add” xml element where the attribute name=”MyDB”, connectionString attribute will contains the specific connection string.

Code Snippet
       <?xml version="1.0"?> 
<configuration xmlns:xdt="http://schemas.microsoft.com/XML-Document-Transform"> 
  <connectionStrings> 
    <add name="MyDB" 
      connectionString="Data Source=MySQLServer;Initial Catalog=MyDevDB;Integrated Security=True" 
      xdt:Transform="SetAttributes" xdt:Locator="Match(name)"/> 
  </connectionStrings> 
  <system.web> 
   
  </system.web> 
</configuration> 
     

Here is the result of the translation after compiling and publishing the application.

Code Snippet
       <?xml version="1.0"?> 
<configuration> 
  <connectionStrings> 
      <add name="MyDB"connectionString="Data Source=MySQLServer;Initial Catalog=MyDevDB;Integrated Security=True" /> 
    </connectionStrings> 
    <system.web> 
      <compilation debug="true" targetFramework="4.0" /> 
      <customErrors mode="Off"></customErrors> 
    </system.web> 
   
    <system.webServer> 
      <modules runAllManagedModulesForAllRequests="true"/> 
    </system.webServer> 
  
</configuration> 
     

Web.Staging.config

You could also make some replacement or element removal and of course another DB connectionString replacement:

Code Snippet
       <?xml version="1.0"?> 
<configuration xmlns:xdt="http://schemas.microsoft.com/XML-Document-Transform"> 
  <connectionStrings> 
    <add name="MyDB" 
      connectionString="Data Source=MySQLServer;Initial Catalog=MyStagingDB;Integrated Security=True" 
      xdt:Transform="SetAttributes" xdt:Locator="Match(name)"/> 
  </connectionStrings> 
  <system.web> 
    <customErrors defaultRedirect="GenericError.htm" mode="RemoteOnly" xdt:Transform="Replace"> 
      <error statusCode="500" redirect="InternalError.htm"/> 
    </customErrors> 
    <compilation xdt:Transform="RemoveAttributes(debug)" /> 
  </system.web> 
</configuration> 
     

Here is the result of the translation.

Code Snippet
       <?xml version="1.0"?> 
<configuration> 
  <connectionStrings> 
      <add name="MyDB"connectionString="Data Source=MySQLServer;Initial Catalog=MyStagingDB;Integrated Security=True" /> 
    </connectionStrings> 
    <system.web> 
      <compilation targetFramework="4.0" /> 
      <customErrors defaultRedirect="GenericError.htm" mode="RemoteOnly"> 
          <error statusCode="500" redirect="InternalError.htm"/> 
      </customErrors> 
    </system.web> 
   
    <system.webServer> 
      <modules runAllManagedModulesForAllRequests="true"/> 
    </system.webServer> 
  
</configuration> 
     

This mechanism is very helpfull, hoping that this helps.

Some more links on this subject

ASP.NET Web Application Project Deployment Overview
http://msdn.microsoft.com/en-us/library/dd394698(VS.100).aspx

Web.config Transformation Syntax for Web Application Project Deployment
http://msdn.microsoft.com/en-us/library/dd465326(VS.100).aspx

Web Deployment: Web.Config Transformation
http://blogs.msdn.com/webdevtools/archive/2009/05/04/web-deployment-web-config-transformation.aspx

Visual Studio 2010 : How to deploy a Web application with Web Application Project Deployment (Part 1)

Benoît Sarie — Mon, 29 Mar 2010 20:04:49 GMT

In all project it’s a real nightmare to manage configuration files. In Silverlight projects like in web applications, you can take advantage of Web Application Project Deployment features of Visual Studio 2010. In this blog, I’ll explain you :

How to create a new configuration (part 1);
How to create package definition (part 1);
How to create package (part 1);
How to publish the application (part 1);
How to use transformation with web.config files (part 2);
Some more links on this subject (part 2).

In this post, I won’t explain SQL deployment.

Creation of a new solution configuration

Right click on your solution file in Visual Studio, select Properties, on the left tree view, select configuration.

After clicking on the button “Configuration Manager …”, another pop-up appears. In the left drop down list, select “<New …>”. The “<Edit …>” button is used to rename or remove a configuration.

In the new pop-up, enter the name of the desired new configuration. In this case, “Staging”. You can use the “Copy settings from:” to duplicate settings from a specific configuration.

By default, the selected configuration is the one just created (i.e. Staging).

To make appears the dedicated web.config in the web application, right click on the web.config and select “Add Config Transforms”, and the Web.Staging.config will be created.

How to create the package definition

Right click on the web project and select “Properties”. Select “Package/Publish” tab.

This screen is used to configure you deployment. You can specifies what you really want to package : pdb files, App_Data folder, database, …

For Silverlight application, “Only files needed to run this application” in the Items to deploy drop down list also deploys xap files.

Than, you can define the Web Deployment Package settings.

How to create the package

Rebuild your application to view the resulting package, right click on the web application and select “Create Package”.

As in previous paragraph, we have kept the default settings, you can see that in the obj folder, Visual Studio has created a new folder named Staging with a lot of files.

A zip file (folder obj/Staging/Package) named like your web project is present. This files contains :

Configuration files used for deployment;
The code to deploy.

In the same folder, a cmd file is present.

How to publish the application

To deploy the package there is several way:

Import the package with IIS Manager (http://learn.iis.net/page.aspx/515/import-a-package-through-iis-manager/);
Using the cmd file that is present in the package folder;
Use Visual Studio “Publish …” (right click on the web project) function. It provides file system, MSDeploy, FTP and FPSE publish mechanism.

Visual Studio provides easy to use screens for deployment. Using Profiles is very great to store all publication locations depending on your server.

In this sample, I used a file system deployment on the c:\temp folder.

WCF 4.0 – Routing

Benoît Sarie — Wed, 24 Mar 2010 20:09:54 GMT

One of the great new features of WCF is the Routing Service. Here is a little post to show you how simple this is to put in place a WCF Routing.

Sample

Code Snippet
 <?xml version="1.0"?>
<configuration>
  <system.web>
    <compilation debug="true" targetFramework="4.0"/>
  </system.web>
  <system.serviceModel>
    <behaviors>
      <serviceBehaviors>
        <behavior>
          <serviceMetadata httpGetEnabled="true"/>
          <serviceDebug includeExceptionDetailInFaults="false"/>
        </behavior>
        <behavior name="RoutingBehavior">
          <routing routeOnHeadersOnly="false" filterTableName="filterTable1" />
          <serviceDebug includeExceptionDetailInFaults="true"/>
          <serviceMetadata httpGetEnabled="true" />
        </behavior>
      </serviceBehaviors>
    </behaviors>
    <serviceHostingEnvironment multipleSiteBindingsEnabled="true" />
    <services>
      <service behaviorConfiguration="RoutingBehavior" name="System.ServiceModel.Routing.RoutingService">
          <endpoint address="" binding="basicHttpBinding" bindingConfiguration="" name="RouterEndpoint1" contract="System.ServiceModel.Routing.IRequestReplyRouter"/>
      </service>
    </services> 
    <routing>
      <filters>
        <filter name="ServiceTestFilter1" filterType="EndpointAddress" filterData="http://localhost/MyRouting/RouterService.svc/Test" />
      </filters>
      <filterTables>
        <filterTable name="filterTable1">
          <add filterName="ServiceTestFilter1" endpointName="BasicHttpBinding_IService1" priority="0" />
        </filterTable>
      </filterTables>
    </routing>
 
    <client>
      <endpoint address="http://localhost/MyApplication/Service1.svc" binding="basicHttpBinding" bindingConfiguration="" contract="*" name="BasicHttpBinding_IService1" />
    </client>
 
  </system.serviceModel>
  <system.webServer>
    <modules runAllManagedModulesForAllRequests="true"/>
  </system.webServer>
</configuration>
 

Creation of a svc for filters

First of all, create a svc file, named RouterService.svc, that implements the RoutingService.

Code Snippet
 <%@ ServiceHost Language="C#" Debug="true" 
  Service="System.ServiceModel.Routing.RoutingService,System.ServiceModel.Routing, version=4.0.0.0, Culture=neutral,PublicKeyToken=31bf3856ad364e35"%>
 

Defining your client EndPoint

First of all, in your web.config application, the final endpoints have to be configured like in common WCF scenarios.

Code Snippet
 <client>
  <endpoint address="http://localhost/MyApplication/Service1.svc" binding="basicHttpBinding" bindingConfiguration="" contract="*" name="BasicHttpBinding_IService1" />
</client>
 

Defining youe server EndPoint

Code Snippet
   <behaviors>
    <serviceBehaviors>
      <behavior>
        <serviceMetadata httpGetEnabled="true"/>
        <serviceDebug includeExceptionDetailInFaults="false"/>
      </behavior>
      <behavior name="RoutingBehavior">
        <routing routeOnHeadersOnly="false" filterTableName="filterTable1" />
        <serviceDebug includeExceptionDetailInFaults="true"/>
        <serviceMetadata httpGetEnabled="true" />
      </behavior>
    </serviceBehaviors>
  </behaviors>
  <serviceHostingEnvironment multipleSiteBindingsEnabled="true" />
  <services>
    <service behaviorConfiguration="RoutingBehavior" name="System.ServiceModel.Routing.RoutingService">
        <endpoint address="" binding="basicHttpBinding" bindingConfiguration="" name="RouterEndpoint1" contract="System.ServiceModel.Routing.IRequestReplyRouter"/>
    </service>
  </services>
 

Defining your filters and table filters

Code Snippet
 <routing>
  <filters>
    <filter name="ServiceTestFilter1" filterType="EndpointAddress" filterData="http://localhost/MyRouting/RouterService.svc/Test" />
  </filters>
  <filterTables>
    <filterTable name="filterTable1">
      <add filterName="ServiceTestFilter1" endpointName="BasicHttpBinding_IService1" priority="0" />
    </filterTable>
  </filterTables>
</routing>
 

In <system.serviceModel>, create a <routing> tag, it contains:

<Filters>: list of filters that will be applied. Name, filterType and filterData are required;
<FilterTables>: mapping of filters with EndPoints

Silverlight et HttpHandler

Benoît Sarie — Wed, 24 Mar 2010 08:47:20 GMT

Http handlers are very interesting for several scenarios:

Security: when you want to check custom authentication or authorization;
Logging: when you want to trace requests to a specific page, resource, …;

In a Silverlight project based on Composite Application Guidance for Silverlight (also known as Prism) you can take advantage of ModuleCatalog and HttpHandler.

I’ll to explain you the advantage of using those 2 technologies in a dummy example.

Scenario

For example, my application is built on Composite Application Guidance for Silverlight. When a client connect to the Silverlight application I really want him to get just the code (code in xap ) he needs. For example, if the client belongs to the “user” role, I don’t want him to download the xaps needed by “administrator” role.

In Composite Application Guidance for Silverlight, you can take advantage of Module Catalog to load xap files from different locations.

Code Snippet
 <Modularity:ModuleCatalogxmlns="http://schemas.microsoft.com/winfx/2006/xaml/presentation" xmlns:x="http://schemas.microsoft.com/winfx/2006/xaml" xmlns:sys="clr-namespace:System;assembly=mscorlib" xmlns:Modularity="clr-namespace:Microsoft.Practices.Composite.Modularity;assembly=Microsoft.Practices.Composite">
  <Modularity:ModuleInfo Ref="ModuleZ.Silverlight.xap" ModuleName="ModuleZ" ModuleType="ModuleZ.ModuleZ, ModuleZ.Silverlight, Version=1.0.0.0" />
</Modularity:ModuleCatalog>
 

In fact, in the Ref attribute of ModuleInfo element, you can put an other location,

for example “../secured/ModuleZ.Silverlight.xap”. Instead of getting it from ClientBin folder, you will get the xap from “secured” directory.

For more information on Module Catalog.

Step 1 : Creation of the HttpHandler

In a web application, insert a new class named for example MyHandler. This class have to implement IHttpHandler interface.

Code Snippet

public class MyHandler : IHttpHandler

    {

        /// <summary>

        /// Constructor

        /// </summary>

        public MyHandler()

        {

        }

        /// <summary>

        /// Process request

        /// </summary>

        /// <param name="context"></param>

        public void ProcessRequest(HttpContext context)

        {

            // Here is the code of your implementation

            ...



        }

        /// <summary>

        /// Enable pooling

        /// </summary>

        public bool IsReusable

        {

            get { return true; }

        }

    }

This interface contains 2 methods :

ProcessRequest : method containing your logic;

IsReusable : method informing that the pooling is enabled;

Step 2 : Implementation of your logic

You just have to develop the logic.

Code Snippet
  public class MyHandler : IHttpHandler
    {
        public MyHandler()
        {
        }
        public void ProcessRequest(HttpContext context)
        {
            string path;
            // Get the requested xap file name
            string expectedFile = context.Request.Url.Segments[context.Request.Url.Segments.Length - 1];
 
            // Get the desired xap file
            if (HttpContext.Current.User.IsInRole("Admin"))
                path = string.Format(@"c:\MyApplication\Admin\{0}",expectedFile);
            else
                path = string.Format(@"c:\MyApplication\User\{0}", expectedFile);
 
 
            // Content Type
            context.Response.ContentType = "application/x-silverlight-2";
 
            // Return the xap file
            using (Stream data = File.OpenRead(path))
            {
 
               BinaryReader reader = new BinaryReader (data);
 
                const int size = 4096;
                byte[] bytes = new byte[4096];
                int numBytes;
                while((numBytes = reader.Read(bytes, 0, size)) > 0)
                    context.Response.OutputStream.Write(bytes, 0, numBytes);                  
               
            }
            context.Response.OutputStream.Flush();
     
        }
        public bool IsReusable
        {
            get { return true; }
        }
    }
 

In the ProcessRequest method, first of all, get the name of the desired xap.

Depending on the user role, define the correct path and than return the xap to the client application.

“C:\MyAplication\Admin\” -> folder containing admin xap

“C:\MyAplication\User\” -> folder containing user xap

Do not forget to set the ContentType to "application/x-silverlight-2".

Of course you can optimize the xap with a caching mechanism, but you have to define a mechanism to refresh it. It can be interesting when you deploy a new xap ;-)

Step 3 : Register your http handler in your web application & Test

In the web.config file you just need to register your handler (sample for IIS 7.0 running in Integrated mode):

Code Snippet
 <system.webServer>
    <handlers>
        <add verb="*" path="Secured/*.xap" name="MyHandler" type="MyNamespace.MyHandler, MyAssembly"/>
    </handlers>
</system.webServer>
 

In this case, the handler will be executed for each verb and every time I’ll request a xap in Secured directory. No need to create the Secured folder.

For other IIS, I suggest you to read this article http://msdn.microsoft.com/en-us/library/ms228090(VS.100).aspx.

To test your handler, you just have to open IE and to execute your request:

http://mycomputer/MyApplication/Secured/test.xap

Finally, you just have to run your Silverlight application. :-)

Conclusion

This scenario can be solve with several other methods. My goal is just to present you the collaboration of those 2 technos.

Adding Silverlight into your blog

Benoît Sarie — Thu, 03 Dec 2009 22:41:26 GMT

Luc Vo Van, one of my colleges has recently created the Windows Live Writer Silverlight App Plugin. Here is the link, really easy to use :

http://blogs.msdn.com/luc/archive/2009/11/19/windows-live-writer-silverlight-app-plugin.aspx

Many Thanks Luc :-)

ADO.NET Data Services : Tips & Tricks

Benoît Sarie — Thu, 03 Dec 2009 22:18:47 GMT

Introduction

In this post, I’ll give you some tips interesting that helped me in my several projects.

For more information on Data Services, I suggest you to read my previous posts:

TIPS 1 : ADO.NET Data Services CTP 1.5 CTP 2

Accessible here, this new release targeting .net 3.5 & Silverlight 3 provides some great new features (some more details about this release):

Projections;
Better data binding;
Row count;
Server Page Driven;
…

-> One of the major feature of this version is to allow cross domain access to your Data Services from your Silverlight application.

Here you can access to a description of this feature written in the ADO.NET Data Services Team Blog.

TIPS 2 : Returning hierarchical data

A great advantage of returning hierarchical in advance to reduce the number of server calls especially when you know that the hierarchical data is needed.

In the following example, the request will return a specific product:

Code Snippet
 public DataServiceQuery<Product> ReturnHierarchicalData(int productId) 
{   // Create a hierachical data query
    DataServiceQuery<Product> request = (from product in _adventureWorksProxy.Products
         .Expand("ProductSubcategory")
         .Expand("ProductSubcategory/ProductCategory")
    where product.ProductID == productId
    select product) as DataServiceQuery<Product>;
 
    return request;
}
 

This request will returns the specific ProductSubCategory entity of the product and the ProductCategory of this ProductSubcategory.

TIPS 3 : Deferred loading

In opposition to hierarchical data return in a single call, when the size of data to return is too heavy, you should use deferred loading.

In this example, I will return the ProductSubcategory of the product only when needed.

Code Snippet
 void ReturnDeferredData(Product product)
{   
    // Deferred load
    _adventureWorksProxy.BeginLoadProperty(product, "ProductSubcategory", ReturnDeferredDataCallBack, product);
}
void ReturnDeferredDataCallBack(IAsyncResult ar) 
{
    _adventureWorksProxy.EndLoadProperty(ar);
    Product product = ar.AsyncState as Product;
}
 

BeginLoadProperty requires four parameters, the entity used for deferred loading, the name of the property used in the deferred mechanism, the callback ant the context.

The callback method will end the asynchronous system.

TIPS 4 : Batching requests

One other great feature of Data Services is batching. it allows you to request in a same time several requests to the server and to return the result only in one time.

This example shows how to execute two requests on the server with only one server access. You can check using fiddler to confirm :-)

Code Snippet
 // Creation of two requests
var request1 = new DataServiceRequest<Product>(new Uri("Products",UriKind.Relative));
var request2 = new DataServiceRequest<ProductSubcategory>(new Uri("ProductSubcategory", UriKind.Relative));
 
// Execution of those two requests on the server
DataServiceResponse responses = adventureWorksProxy.ExecuteBatch(request1, request2);
foreach (QueryOperationResponse response in responses)
{
    if(response.Query.ElementType == typeof(Product))
    {
        // ...
    }
    else if (response.Query.ElementType == typeof(Product))
    {
        // ...
    }
}
 

Here is an interesting post from ADO.NET Data Services Team Blog.

Visual Studio 2010 Beta 2 : How to add an external tool (MSBuild script) to a Visual Studio button

Benoît Sarie — Tue, 24 Nov 2009 23:13:11 GMT

I’m a real fan of MSBuild Scripts, I always use a lot of MSBuild Tools in my projects (for example scripts that automatically get latest version of the code, compile, check-in, …).

The easier way I found to execute those scripts is to run them automatically while I am clicking on a button into my Visual Studio.

STEP 1: Creation of an external tool

In Visual Studio 2010, Click on “Tools” menu, “External Tools …”.

A pop-up appears, click Add button and enter the desired information.

“Title” textbox is used as the name of the button that will be created.

If you use MSBuild on a X64 operating System, you have to select x86 MSBuild version:

C:\Windows\Microsoft.NET\Framework\v4.0.21006\MSBuild.exe

Click on “Use Output window” checkbox to redirect the result of your MSBuild script into Visual Studio Output window.

Step2 : Creation of your Toolbar and affectation of the external tool

On the menu bar, right click and select “Customize …”.

A pop-up appear, create your desired toolbar. Name it MyToolbar and click OK.

Click on the “Commands” tab and click on ToolBar radio button to select the newly created external tool.

Select MyToolBar and click on “Add Command …” button.

On the left lit item, select “Tools” and than select the “External Command x” item where X is the index of your tool that appears in the “Tool” menu (starting index => 1). In my example “External Command 6”.

Click “Ok” and than “Close”.

And here is the result :-)

ADO.NET Data Services : How to consume my ADO.NET Data Services in WPF

Benoît Sarie — Tue, 17 Nov 2009 22:22:10 GMT

Introduction

In this post I will talk about how you can consume your ADO.NET Data Services. My examples will be based on WPF.

For more information on Data Services, I suggest you to read my previous posts:

How to create an ADO.NET Data Services with Visual Studio 2010 beta2
How to request your ADO.NET DataServices
Operations and Interceptors Consuming ADO.NET Data Services in WPF

Consuming in WPF

Creates a WPF application named MyWPFApplication.

Add a reference to your Data Service, right click on MyWPFApplication project, and select Add service reference. A windows appears, define the address, rename the namespace with MyDataServiceProxy and click on OK button.

In MainWindow.xaml.cs, write the following code:

Code Snippet
 public partial class MainWindow : Window
{
    public MainWindow()
    {
        InitializeComponent();
        this.Loaded += new RoutedEventHandler(MainWindow_Loaded);
    }
 
    void MainWindow_Loaded(object sender, RoutedEventArgs e)
    {
 
        // Define the URI of the Data Service Query
        Uri uri = new Uri(@"http://localhost:34570/AdventureWorksService.svc/");
        // Create an instance of the Data Service context
        MyDataServiceProxy.AdventureWorksModel.AdventureWorksEntities adventureWorksProxy = new MyDataServiceProxy.AdventureWorksModel.AdventureWorksEntities(uri);
        // Create your query
        var query = (from product in adventureWorksProxy.Products
                    select product).Take(10);
        // Set the qury result into the DataContext
        this.DataContext = query;
    }
}
 

In the loaded event, instantiate your DataService. Create a simple request using Linq syntax and put this query into the DataContext.

In MainWindow.xaml, bind the data grid with the DataContext :

Code Snippet
 <Window x:Class="MyWpfApplication.MainWindow"
        xmlns="http://schemas.microsoft.com/winfx/2006/xaml/presentation"
        xmlns:x="http://schemas.microsoft.com/winfx/2006/xaml"
        Title="MainWindow" Height="350" Width="525">
    <Grid>
        <DataGrid AutoGenerateColumns="true" Height="287" HorizontalAlignment="Left" Margin="12,12,0,0" Name="dataGrid1" VerticalAlignment="Top" Width="479"
                  ItemsSource="{Binding}">
        </DataGrid>
    </Grid>
</Window>
 

Press F5 and here is the result:

This post shows how easy it is to consume Data Services with WPF. Just few code lines are needed ;-)

ADO.NET Data Services : Operations and Interceptors

Benoît Sarie — Wed, 11 Nov 2009 15:18:56 GMT

Introduction

In this post I will talk about Operations & Interceptors functionalities included into ADO.NET Data Services.

For more information on Data Services, I suggest you to read my previous posts:

Operations

If you need to aggregate an operation, instead of creating the rule on the client side (on each sort of client application), you can centralize it easily on the server side using Operators.

In the following example, GetProductsByColor returns all products of a specific color.

Code Snippet
 [WebGet]
public IQueryable<Product> GetProductsByColor(string color)
{
    return from product in this.CurrentDataSource.Products
           where string.Compare(color, product.Color) == 0
           select product;
}
 

The WebGetAttibute is used to request your service via REST.

Inside the method, you just have to write your Linq query on the CurrentDataSource to get the desired result.

You also need to define the Service Operation Access Control. You have to grant read access to the operation. Here is the code to add in the InitializeService method:

Code Snippet
  
config.SetServiceOperationAccessRule("GetProductsByColor", ServiceOperationRights.AllRead);

InitializeService is the method where you already have granted access to the Data Service.

And finally, you can execute the code with this URL:

http://localhost:34570/AdventureWorksService.svc/GetProductsByColor?color='Red'

Interceptors

Interceptors is the Data Services mechanism to customize logic on the server side. With this mechanism, you can intercept queries and add specific restriction rules.

Example 1 : Interceptor on a query request

In the following example, you will create :

- a filter to check if the person requesting the DataService service on the Products entity is member of the Admin role (defined for example with ASP.NET role provider), else an exception is thrown;

- only Products where SellEndDate property is not null;

Open you DataService class. Don't forget to add at the top of you file:

Code Snippet
  
using System.Linq.Expressions;

Create the filter method on the product entity. You need to create a function prefixed with QueryInterceptor attribute. This attribute expects the name of the entity on which the interceptor will be applied (as defined in the Data Service Workspace http://localhost:34570/AdventureWorksService.svc/). The method has to follow a specific signature.

Code Snippet
 [QueryInterceptor("Products")]
public Expression<Func<Product, bool>> FilterProducts()
{
    if (HttpContext.Current.User.IsInRole("Admin"))
        return (p => p.SellEndDate !=null);
    else
        throw new UnauthorizedAccessException("You don't have access to this information");
}
 

Example 2 : Interceptor on a change query (Add, Change, Delete, … operation)

In this example, I’m just checking that only Admins have granted privileges for adding,updating or deleting Product entity set. The signature of the method is a little bit different from previous example.

Code Snippet
 [ChangeInterceptor("Products")]
public void onChangeProducts(Product product, UpdateOperations operations)
{
    if (!HttpContext.Current.User.IsInRole("Admin"))
        throw new UnauthorizedAccessException("You don't have access to this information");
}
 

For more information, I suggest you to read MSDN documentation.

ADO.NET Data Services : How to request your ADO.NET Data Services

Benoît Sarie — Tue, 27 Oct 2009 19:04:58 GMT

Introduction

In my previous post, I explained how to create a ADO.NET Data Service with Visual Studio 2010. In this post, I will explain how to request this Data Service using Internet Explorer (but this will also work in any other browser). The database used in previous post is AdventureWorks.

Listing and filtering entities

By default when accessing the service, the list of all entities is returned.

http://localhost:34570/AdventureWorksService.svc

http://localhost:34570/AdventureWorksService.svc/Products
This request lists all products existing in the database.

Let’s go further, you would like to return articles 6 to 7 order by ProductNumber:

http://localhost:34570/AdventureWorksService.svc/Products?$orderby=ProductNumber&$skip=5&$top=2

REST Syntax	Description
$orderby=ProductNumber	order Products by ProductNumber
&$skip=5	and skips the first 5 rows
&$top=2	and returns the two first rows

You could also order the results by ProductNumber descending:

http://localhost:34570/AdventureWorksService.svc/Products?$orderby=ProductNumber%20desc&$skip=5&$top=2

REST Syntax	Description
$orderby=ProductNumber desc	order Products by ProductNumber descending
&$skip=5	and skips the first 5 rows
&$top=2	and returns the two first rows

Here is the syntax to filter a product in particular:

http://localhost:34570/AdventureWorksService.svc/Products?$filter=ProductID%20eq%201

REST Syntax	Description
$filter=ProductID eq 1	where ProductID is equal to 1

http://localhost:34570/AdventureWorksService.svc/Products?$filter=ProductID%20ge%20900%20and%20ProductID%20lt%20950

REST Syntax	Description
$filter=ProductID ge 900	where ProductID is greater or equal to 900
and ProductID lt 950	and ProductID less than 950

Other expression syntaxes are available here.

Expanding SubEntities

One of the great feature of Data Services is the ability to also return referenced entites.

For example, if you want to request a specific Product, and also return the information defined in its ProductSubCategory property, the request is as follows:

http://localhost:34570/AdventureWorksService.svc/Products?$filter=ProductID%20eq%20771&$expand=ProductSubcategory

REST Syntax	Description
$filter=ProductID eq 771	where ProductID equals to 771
&$expand=ProductSubcategory	and expands ProductSubcategory for this specific Product. ProductSubcategory is the name of the navigation property for ProductSubcategory in the Product entity defined in your emdx file

And if you also need to retrieve the Catagory entity, which is two levels above :

http://localhost:34570/AdventureWorksService.svc/Products?$filter=ProductID%20eq%20771&$expand=ProductSubcategory,ProductSubcategory/ProductCategory

REST Syntax	Description
$filter=ProductID eq 771	where ProductID equals to 771
&$expand=ProductSubcategory, ProductSubcategory/ProductCategory	expands ProductSubcategory for this specific Product and ProductSubcategory/ProductCategory

New features of Visual Studio 2010 beta 2

When your Data Service project is running with debugger inside VS2010, all the requests executed are listed into Visual Studio. Really great feature :-)

ADO.NET Data Services : How to create an ADO.NET Data Services with Visual Studio 2010 beta2

Benoît Sarie — Tue, 27 Oct 2009 17:48:22 GMT

Introduction:

This post is the first of a series of articles based on Visual Studio 2010. In this one I will introduce you to publishing your data through ADO.NET Data Services. Here are the main interest for me into using Data Services:

Exposing your data as a Data Service allows clients to consume it easily from using the technology you want (WPF, Silverlight, ASP.NET, J2EE, …);
Addressing Scheme for Data with URIs;
Data Transport Format ubiquitous (AtomPub or JSON);
Data Storage independence;
Usage of REST semantic;

Pre-requisites:

Visual Studio 2010 beta2 + SQL Server 2005 Express or 2008 Express http://msdn.microsoft.com/en-us/vstudio/dd582936.aspx

AdventureWorks database http://www.codeplex.com/MSFTDBProdSamples/Release/ProjectReleases.aspx?ReleaseId=18407

Step 1 : Create your web application project

After installing all software pre-requisites, let’s begin.

Open Visual Studio 2010 beta 2 and create an empty ASP.NET Web application named MyDataServicesApplication, for example. This web project will contains the edmx (i.e. the Entity Data Model) file and the ADO.NET Data Services class.

Step 2 : Create an Entity Model of the AdventureWorks database

First of all, a edmx file has to be created.

Create a new Item, select ADO.NET Entity Data Model item template, name it AdventureWorksModel

Press the Add button and follow the configuration wizard step by step

Select Generate from database to generate the edmx file from the existing database structure.

Select the AdventureWorks database

Specify the name of the connection string used, the specified name is stored in the web.config file into the connectionString tags.

Click the checkbox next to the Tables element, in order to include all existing tables into your data model.

Check the Pluralize or singularize generated object names check box and specify AdventureWorksModel as the namespace. This will pluralize all collection property names, this is a new feature of Visual Studio 2010.

Step 3 : Create the ADO.NET Data Service

Create a new item, Select the ADO.NET Data Service template and name it AdventuWorksService.svc

Open AdventureWorksService.svc.cs and update the content to be like this:

Code Snippet
  
namespace MyDataServicesApplication
{
    public class AdventureWorksService : DataService<AdventureWorksEntities>
    {
        // This method is called only once to initialize service-wide policies.
        public static void InitializeService(DataServiceConfiguration config)
        {
            // Grant full access to all exsisting entities
            config.SetEntitySetAccessRule("*", EntitySetRights.All);
            // Grant full access for all operations
            config.SetServiceOperationAccessRule("*", ServiceOperationRights.All);
            config.DataServiceBehavior.MaxProtocolVersion = DataServiceProtocolVersion.V2;
        }
    }
}
 

Step 4 : Test

Press F5 to check the result, it should list all the available entities :

You can now request your database through REST, for example select all the products using this query (which is now just an URI!): http://localhost:34570/AdventureWorksService.svc/Products

The result is represented by default using the ATOM Publishing Protocol. Internet Explorer automatically applies a style on ATOM format to make it more readable. Here this is not the case :-). I suggest that you right click on the content of the page and select View Source, the result will be more user friendly.

Here is a link of the rest Services sementics, a usefull link : http://msdn.microsoft.com/en-us/library/cc668767.aspx

Conclusion

Here is a step by step guide for creating an ADO.NET Data Services with Visual Studio 2010 beta2.

Customer Care Framework 2009 SP1 RTM is now available

Benoît Sarie — Mon, 20 Apr 2009 09:39:00 GMT

CCF 2009 SP1 RTM has been released. This new version contains some new great features like :

64 bit client support;
Dynamic positionning in IAD;
Citric enhancement ;
Usage for a same application of several DDAs;
New platforms support;
...

You can found more details on those two new links:

MSDN - Development guide: http://msdn.microsoft.com/en-us/library/aa306213.aspx
Technet - Deployment guide: http://technet.microsoft.com/en-us/library/cc508530.aspx

Enterprise Library & Unity - Some interesting links

Benoît Sarie — Thu, 09 Apr 2009 22:53:00 GMT

Hands-On Labs for Enterprise Library 4.1 and Hands-On Labs for Unity Application Block 1.2 have been released few days ago; here is a the good opportunity to summarize some interesting links.

Enterprise Library:

MSDN - Enterprise Library site : http://msdn.microsoft.com/entlib
Enterprise Library wiki : http://www.codeplex.com/entlib/Wiki/View.aspx?title=EntLib%20FAQ
Codeplex : http://codeplex.com/entlib
HOL Enterprise Library 4.1 : http://www.microsoft.com/downloads/details.aspx?displaylang=en&FamilyID=ab3f2168-fea1-4fc2-b40c-7867d99d4b6a

Unity Application Block:

MSDN - Unity portal : http://msdn.microsoft.com/unity
Codeplex : http://codeplex.com/unity
HOL Unity Application Block 1.2 : http://www.microsoft.com/downloads/details.aspx?displaylang=en&FamilyID=93a5e18f-3211-44ef-b785-c59bcec4cd6f

Pattern & practice:

MSDN - Pattern & practice portal: http://msdn.microsoft.com/practices

Blogs:

· Gregori Melnik: http://blogs.msdn.com/agile

· J.D. Meier: http://blogs.msdn.com/jmeier