Within IIS you can backup everything to a server package by opening the context menu selecting depoy, then “Export Server Package”. This will create the server package containing the IIS configuration along with all the sites and their files. This is all good however it is still a manual process that is likely to be forgotten about after a while. Likely there are a series of libraries we can use to automate this process.

The namespace Microsoft.Web.Deployment provide the functionality we need to backup IIS, this is provided by Microsoft.Web.Deployment.dll. To backup the server we need to use the DeploymentManager to create an object use to create the package, we can then use this to “sync” the server setting to a package.

After we have backed up the site package we need to upload it to Amazon S3. To do this we can use the AWSSDK .NET library. The library can be installed via NuGet by searching for “AWS SDK”. This will install the required assemblies.

We can use the AWS SDK to upload our server package zip to S3 using the TransferUtilityUploadRequest to create our request to use with the TransferUtility to upload our file.

When this is all tied together we have an app that will backup and upload the server package to S3.

I have attached a sample app that can be configured using the App.config

Download yeticode-utility.zip

/// <summary>
/// User Online Action Filter Attribute
/// </summary>
public class UserOnlineActionFilterAttribute : ActionFilterAttribute
{
    /// <summary>
    /// Called by the ASP.NET MVC framework after the action method executes.
    /// </summary>
    /// <param name="filterContext">The filter context.</param>
    public override void OnActionExecuted(ActionExecutedContext filterContext)
    {
        try
        {
            if (filterContext.HttpContext != null && filterContext.HttpContext.User != null && filterContext.HttpContext.User.Identity != null)
            {
                Membership.Provider.GetUser(filterContext.HttpContext.User.Identity.Name, true);
            }
        }
        catch
        {
            // BAD: Swallow the exception to prevent it messing with the users action
        }

        base.OnActionExecuted(filterContext);
    }
}

Add the action filter to the global list inside the global.asax Application_Start.

GlobalFilters.Filters.Add(new UserOnlineActionFilterAttribute());

401 - Unauthorized: Access is denied due to invalid credentials.

After searching around for a solution to why this was not displaying the correct page, even if it was redirecting to the login page I decided to try to set the error displaying to return more information. After adding <httpErrors errorMode=”Details” /> to try and get a detail output from IIS the page started to work as expected.

With this option enabled the login page was now correctly displayed. This can be set without having to edit the web.config on IIS7 by editing the “Error Pages” features settings for your site in IIS. Set the Error Response to “Detailed errors”

Related Reading

I am using NHibernate 2.1.2 and C# .NET 3.5 in Visual Studio 2008.

Step 1

Download Castle.Core from gitbhub
git://github.com/castleproject/Castle.Core.git

Step 2

Enable Castle to allow partially trusted callers.
Open buildscripts/CommonAssemblyInfo.cs and enable the library to enable partially trusted callers

Step 3

Disable generation of debug information for the projects Castle.Core and Castle.DynamicProxy.

Step 4

Build the projects

Step 5

Download Nhibernate source code from source forge.

http://sourceforge.net/projects/nhibernate/files/

Step 6

Open Nhibernate.Everything.sln and update the references for the NHibernate.ByteCode.Castle project so that the references for Castle.Core and Castle.DynamicProxy2 are the libraries build in step 4. Disable generation of debug information for the projects NHibernate, Iesi.Collections and NHibernate.ByteCode.Castle, similar to what was done in step 3.

Step 7

Copy the outputted library files from NHibernate.ByteCode/bin/Release to your shared libs folder of your project.

Step 8

Update the references in your project to point to the new libraries. Next turn off reflection optimization. This needs to be done in code before you create the configuration object. Configuring this in the hibernate.cfg.xml does not work. Update all your projects assemblies to set allow partially trusted callers, similar to step 2. Update your web.config to set requirePermission=”false” in the nhibernate section delceration

NHibernate.Cfg.Environment.UseReflectionOptimizer = false;
Configuration = new Configuration();
Configuration.Configure();

You can now run your site in medium trust using lay loading.

Links

• msysgit – http://code.google.com/p/msysgit/
• TortoiseGit – http://code.google.com/p/tortoisegit/
• NHibernate – http://nhforge.org/
• Castle.Core Source – http://github.com/castleproject/
• NHibernate Source – http://sourceforge.net/projects/nhibernate/files/

Related Reading

Comments

The task automatically identify comments in the following style

//TODO: First Test to do Item
//FIXME: First Fix me Item
//HACK: First Hack Item

Additional matches can be made by adding in Token elements to the NAnt task.

Sample

<ToDo source="path/to/source" output="report.xml" searchpattern="*.cs;*.txt">
 <Tokens>
 <Tokens Value="BUGFIX" />
 </Tokens>
</ToDo>

Output

<ToDo xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:xsd="http://www.w3.org/2001/XMLSchema">
 <Items>
 <Item>
 <Message>First Test to do Item</Message>
 <File>NAnt.ToDo\ToDoTask.cs</File>
 <Line>0</Line>
 <Type>TODO</Type>
 </Item>
 <Item>
 <Message>First Hack Item</Message>
 <File>NAnt.ToDo\ToDoTask.cs</File>
 <Line>0</Line>
 <Type>HACK</Type>
 </Item>
 <Item>
 <Message>First Fix me Item</Message>
 <File>NAnt.ToDo\ToDoTask.cs</File>
 <Line>0</Line>
 <Type>FIXME</Type>
 </Item>
 </Items>
</ToDo>

Cruise Control Xsl

Below is a basic xsl style sheet that can be used with CruiseControl.NET to display the output from NAnt.ToDo.

<?xml version="1.0"?>
<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform" version="1.0">
 <xsl:output method="html"/>
 <xsl:param name="applicationPath"/>
 <xsl:template match="/">
 <div id="ToDo">
 <h1>ToDo List</h1>
 <div id="Summary">
 <h3>To Do</h3>
 <table>
 <tbody>
 <xsl:for-each select="//ToDo/Items/Item">
 <tr>
 <td>
 <xsl:value-of select="Type/text()"/>
 </td>
 <td>
 <xsl:value-of select="Message/text()"/>
 </td>
 <td>
 <xsl:value-of select="File/text()"/>
 </td>
 </tr>
 </xsl:for-each>
 </tbody>
 </table>
 </div>    
 </div>
 </xsl:template>
</xsl:stylesheet>

In order so solve this I decided to experiment to find a method of getting cruise control to run our jqunit tests for me. To run the tests I looked at wrapping the jqunit tests with nunit tests.

jqRunner is designed so that you can us your existing jqunit tests in their existing location without changing them. jqRunner required that all the scripts that are needed to run are registered. The full path to the file is required. jqRunner then executes the tests and returns the results, which are then parsed by nunit as tests using the TestCaseSource attribute.

var sampleTestCase = new jqUnit.TestCase('Sample Test Case', function() {
    /*setup*/
    // this.yep(1);
}, function() {
    /*teardown*/
    // this.ok(1)
});
sampleTestCase.test('Sample Test 1', function() {
    this.ok(1);
});
sampleTestCase.test('Sample Test 2', function() {
    this.ok(0);
});

using System;
using System.Collections.Generic;
using NUnit.Framework;
using jqRunner;
using System.IO;
using System.Reflection;

namespace jqRunner.Tests
{
    [TestFixture,RequiresSTA]
    public class TestCaseSourceTests
    {
        [Test, TestCaseSource("GetTestResults"), RequiresSTA]
        public void CheckTest(ITestResult result)
        {
            Assert.IsTrue(result.Pass, result.Name);
        }

        private static ITestResult[] GetTestResults()
        {
            string jsFile = Path.GetDirectoryName(Assembly.GetExecutingAssembly().CodeBase) + @"\jqUnitTests\SampleUnitTest.js";
            TestBed target = new TestBed();
            target.RegisterScript(jsFile);
            return target.Execute().ToArray();
        }
    }
}

This is an initial development snapshot and is bound to have plenty of problems.

Download
http://static.yeticode.co.uk/blog/downloads/jqRunner-snapshot.zip

Related Reading

So I’ve been working more on the support for django within monodevelop. Heres some screenshots of it.

Reading though the sample chapter is a good way to get started with ASP.NET MVC as the chapter explains what it is doing in a easy to understand manor, but more importantly it explains why they are being done. The chapter is a comprehensive walk through of how to create the NerdDinner application. Throughout the chapter the authors provide titbits of useful information, for example the use of the repository pattern and dependency injection. Throughout the chapter the authors trying to enforce good coding practice and often make reference to test driven development which in a later chapter they go into more detail about, discussing different development techniques.

The sample chapter is more than enough to get anyone started with ASP.NET MVC and follows the development process from start to finish including a section on unit testing the application. The following chapters delve into more detail about specific areas of the ASP.NET MVC framework. The second and third chapters go into detail about the MVC pattern, detailing its background and its use on the web. The chapter also provides an overview, albeit brief, of other frameworks that are available. Although their coverage was brief they were not mentioned as either superior or inferior rather just as alternatives to the ASP.NET MVC framework. This tone was subsequently repeated in the third chapter when the authors compare ASP.NET with WebForms to ASP.NET MVC. They often mention that MVC is not a replacement for WebForms rather it is a different approach. These two chapters are extremely useful as they allow you to make an informed decision whether ASP.NET MVC is the right choice for you.

I’ve still to finish reading the rest of this book so I’ll update this post the more I read of it.

Here’s a sample nunit test that uses nmock. The nmock object mocks the calls that request specific urls and loads the data for local files.

[Test()]
public void MockLinkFinderTest()
{
	IMock mockWebClient = new DynamicMock(typeof(IWebClient));
	mockWebClient.ExpectAndReturn("GetPage",File.ReadAllText("HTML/1.htm"),new object[]{"http://domain.example/1.htm"});
	mockWebClient.ExpectAndReturn("GetPage",File.ReadAllText("HTML/2.htm"),new object[]{"http://domain.example/2.htm"});
	mockWebClient.ExpectAndReturn("GetPage",File.ReadAllText("HTML/3.htm"),new object[]{"http://domain.example/3.htm"});

	string domain = "domain.example";
	string url = "http://domain.example/1.htm";

	LinkFinder linkFinder = new LinkFinder();
	LinkItemCollection linkCollection = linkFinder.Find(url,domain,(IWebClient)mockWebClient.MockInstance);

	int expected = 3;
	Assert.AreEqual(expected,linkCollection.Count);
}

The RssNotify project aims to provide an easy way for a user to receive notifications when an rss feed has been updated. the idea is to use notify-sharp to display the new news items. The rest of the interface will be used to try to explore the features widgets that GTK# offers.