dnknormark.net

Search Twitter from C# using LINQ to XML

mn.nospam@nospam.dnknormark.net (MartinHN) — Mon, 16 Mar 2009 21:28:08 +0200

In some applications, it could be cool to have a feature that enabled the user to quickly get a glimpse of what people are saying on Twitter about the user or their product, service, company etc.

For instance, a service like GetSatisfaction.com has a feature just like that. They call it Overheard, and this is what it looks like:

There’s nothing like Twitter to give you feedback. I think MediaTemple felt the effect of unhappy customers on Twitter when their servers broke down, and stayed there for more than two days!

Anyway. I wanted to search from C#, and get back a DataTable. Here’s how it’s done:

    /// <summary>
    /// Searches Twitter for the specified query.
    /// </summary>
    /// <param name="query">The query.</param>
    /// <returns>Returns the search results as a DataTable</returns>
    public DataTable Search(string query)
    {
      DataTable dt = new DataTable();
      dt.Columns.Add("text");
      dt.Columns.Add("html");
      dt.Columns.Add("pubdate");
      dt.Columns.Add("id");
      dt.Columns.Add("link");
      dt.Columns.Add("authorname");
      dt.Columns.Add("authorlink");

      XDocument tweetResults = XDocument.Load(String.Format(

      "http://search.twitter.com/search.atom?q={0}", HttpUtility.UrlEncode(query)));

      XNamespace atomNS = "http://www.w3.org/2005/Atom";
      var q = from tweet in tweetResults.Descendants(atomNS + "entry")
              select new
              {
                Text = (string)tweet.Element(atomNS + "title"),
                Html = (string)tweet.Element(atomNS + "content"),
                DatePublished = DateTime.Parse((string)tweet.Element(atomNS + "published")),
                Id = (string)tweet.Element(atomNS + "id"),
                Link = (string)tweet.Elements(atomNS + "link")
                .Where(link => (string)link.Attribute("rel") == "alternate")
                .Select(link => (string)link.Attribute("href"))
                .First(),
                Author = (from author in tweet.Descendants(atomNS + "author")
                          select new
                          {
                            Name = (string)author.Element(atomNS + "name"),
                            Uri = (string)author.Element(atomNS + "uri"),
                          }).First()
              };
      
      foreach (var item in q)
      {
        dt.Rows.Add(item.Text, item.Html, item.DatePublished, item.Id, item.Link,

                    item.Author.Name, item.Author.Uri);
      }

      return dt;
    }

Migrate web.config to support the IIS 7 Integrated Pipeline

mn.nospam@nospam.dnknormark.net (martinhn) — Sun, 15 Mar 2009 22:13:47 +0200

Whenever you deploy a website to IIS 7 that is not compliant with the IIS 7 integrated pipeline, you will get an error like this one:

Not the great error message you get. It actually gives you the solution right away: Migrate Web.config to support the integrated pipeline. To do that, start a command prompt, and execute:

%SystemRoot%\system32\inetsrv\appcmd migrate config "test/"

After doing this, our Web.config is changed to support the IIS 7 integrated pipeline and we can see the website.

Automatically translate Global and Local Resource (resx) files

mn.nospam@nospam.dnknormark.net (MartinHN) — Fri, 13 Mar 2009 22:31:14 +0200

Yesterday, I blogged about how you can use Google Translate to translate a string in C#. To make it more useful than just a simple translator, and because I need to translate some Global Resource files for an E-commerce website that I’m working on, I wanted to create a small Windows Application in C# that could read a Global Resource file (.resx) and translate it into a selected language using the method for translating a word in C# that i blogged about yesterday.

This is how it looks so far. You simply select the resource file you want to translate. Select the current language of the resource file in the middle box, and select the language you want to translate it to in the last box. Click Translate at it should work. The new resource file will be saved in the same location as the application itself.

Translate text in C#, using Google Translate

mn.nospam@nospam.dnknormark.net (martinhn) — Thu, 12 Mar 2009 00:14:26 +0200

Sometimes, it would be great to be able to translate a text from e.g. English to Danish directly from C#. This could be useful when you want to translate a Resource file into another language.

Google Translate is awesome. There’s also Windows Live Translator, but Microsoft are far behind Google (also) in this game.

Code:

using System;
using System.Net;
using System.Text;
using System.Text.RegularExpressions;

namespace Utilities
{
  public static class Translator
  {
    /// <summary>
    /// Translates the text.
    /// </summary>
    /// <param name="input">The input.</param>
    /// <param name="languagePair">The language pair.</param>
    /// <returns></returns>
    public static string TranslateText(string input, string languagePair)
    {
      return TranslateText(input, languagePair, System.Text.Encoding.UTF7);
    }

    /// <summary>
    /// Translate Text using Google Translate
    /// </summary>
    /// <param name="input">The string you want translated</param>
    /// <param name="languagePair">2 letter Language Pair, delimited by "|". 
    /// e.g. "en|da" language pair means to translate from English to Danish</param>
    /// <param name="encoding">The encoding.</param>
    /// <returns>Translated to String</returns>
    public static string TranslateText(string input, string languagePair, Encoding encoding)
    {
      string url = String.Format("http://www.google.com/translate_t?hl=en&ie=UTF8&text={0}&langpair={1}", input, languagePair);

      string result = String.Empty;

      using (WebClient webClient = new WebClient())
      {
        webClient.Encoding = encoding;
        result = webClient.DownloadString(url);
      }

      Match m = Regex.Match(result, "(?<=<div id=result_box dir=\"ltr\">)(.*?)(?=</div>)");

      if (m.Success)
        result = m.Value;

      return result;
    }
  }
}

The translated string is fetched by the RegEx close to the bottom. This could of course change, and you have to keep it up to date.

ASP.NET Search Engine Optimization

mn.nospam@nospam.dnknormark.net (martinhn) — Wed, 11 Mar 2009 17:46:04 +0200

I made a few E-commerce websites a few years back when I was self-employed. Ranking well in the search engines is very important for E-commerce websites in order to get relevant visitors, and if your site is designed well for converting, visitors into sales, you will make money. That’s the way it goes.

A lot of naive CEO’s and alike, think that you absolutely have to hire expensive SEO consultants in order to rank well. That is *not* true. Instead, why don’t you just spent your time and money on making original, relevant and great content? That’s a true winner.

Though, there’s still a few technical details you have to get right.

Make sure to set a unique page title on all your pages. Use the same text as a <h1> tag on the page.

Don’t screw up your URLs. You will get punished having the same content on those URLs: www.example.com/producs/computers?sortorder=price&page=2 and www.example.com/producs/computers?page=2&sortorder=price. This is called duplicate content, and search engines doesn’t like that. You can use the new canonical tag for telling search engines which one is original.

Make internal links absolute. Yes. You shouldn’t do this: <a href=”/page1.htm”>Page 1</a>. Do this instead: <a href=”http://www.example.com/page1.htm”>Page 1</a>.

301 permanent redirect example.com to www.example.com or vice versa. Allowing visitors to access your page on both URLs, will also be treated as duplicate content.

If you’re moving your website to a new platform, server, technology and your URLs will change. Make sure not to return 404 on the old URLs. You have to 301 permanent redirect the old URLs to the new ones. Otherwise you will lose all your current search engine carma.

I use those two methods to set page information, and 301 redirect:

  private void SetMetaInformation()
  {
    this.Title = PageTitle;

    HtmlMeta metaKeywords = new HtmlMeta();
    metaKeywords.Name = "keywords";
    metaKeywords.Content = this.MetaKeywords;
    this.Header.Controls.Add(metaKeywords);

    HtmlMeta metaDescription = new HtmlMeta();
    metaDescription.Name = "description";
    metaDescription.Content = this.MetaDescription;
    this.Header.Controls.Add(metaDescription);

    HtmlMeta metaRobots = new HtmlMeta();
    metaRobots.Name = "robots";
    metaRobots.Content = this.MetaRobots;
    this.Header.Controls.Add(metaRobots);
  }

  public void PermanentRedirect(string newUrl)
  {
    Response.Status = "301 Moved Permanently";
    Response.StatusCode = 301;

    Response.AddHeader("Location", newUrl);
  }

You also want to take a look at the Google SiteMap.

Create a generic handler in ASP.NET and call it sitemap.ashx. Generate and XML string like this:

<?xml version="1.0" encoding="UTF-8"?>
<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9">
   <url>
      <loc>http://www.example.com/</loc>
      <lastmod>2005-01-01</lastmod>
      <changefreq>monthly</changefreq>
      <priority>0.8</priority>
   </url>
</urlset>

Take a further look at the importance of sitemaps, if you want to know why you should use one.

C# TwitPic API client

mn.nospam@nospam.dnknormark.net (martinhn) — Tue, 10 Mar 2009 22:51:38 +0200

I’ve spent some time lately, playing around with the Twitter API. And along with that belongs the TwitPic’s API. I’m using Twitter a lot, to stay in touch with tech news, other developers and just for fun. But it’s getting more and more used for a lot of different things, and I needed it to integrate with an E-commerce platform I’m developing.

The code for post a picture to TwitPic looks like this:

    /// <summary>
    /// URL for the TwitPic API's upload method
    /// </summary>
    private const string TWITPIC_UPLADO_API_URL = "http://twitpic.com/api/upload";

    /// <summary>
    /// URL for the TwitPic API's upload and post method
    /// </summary>
    private const string TWITPIC_UPLOAD_AND_POST_API_URL = "http://twitpic.com/api/uploadAndPost";

    /// <summary>
    /// Uploads the photo and sends a new Tweet
    /// </summary>
    /// <param name="binaryImageData">The binary image data.</param>
    /// <param name="tweetMessage">The tweet message.</param>
    /// <param name="filename">The filename.</param>
    /// <returns>Return true, if the operation was succeded.</returns>
    public bool UploadPhoto(byte[] binaryImageData, string tweetMessage, string filename)
    {
      // Documentation: http://www.twitpic.com/api.do
      string boundary = Guid.NewGuid().ToString();
      string requestUrl = String.IsNullOrEmpty(tweetMessage) ? TWITPIC_UPLADO_API_URL : TWITPIC_UPLOAD_AND_POST_API_URL;
      HttpWebRequest request = (HttpWebRequest)WebRequest.Create(requestUrl);
      string encoding = "iso-8859-1";

      request.PreAuthenticate = true;
      request.AllowWriteStreamBuffering = true;
      request.ContentType = string.Format("multipart/form-data; boundary={0}", boundary);
      request.Method = "POST";

      string header = string.Format("--{0}", boundary);
      string footer = string.Format("--{0}--", boundary);

      StringBuilder contents = new StringBuilder();
      contents.AppendLine(header);

      string fileContentType = GetImageContentType(filename);
      string fileHeader = String.Format("Content-Disposition: file; name=\"{0}\"; filename=\"{1}\"", "media", filename);
      string fileData = Encoding.GetEncoding(encoding).GetString(binaryImageData);

      contents.AppendLine(fileHeader);
      contents.AppendLine(String.Format("Content-Type: {0}", fileContentType));
      contents.AppendLine();
      contents.AppendLine(fileData);

      contents.AppendLine(header);
      contents.AppendLine(String.Format("Content-Disposition: form-data; name=\"{0}\"", "username"));
      contents.AppendLine();
      contents.AppendLine(this.Username);

      contents.AppendLine(header);
      contents.AppendLine(String.Format("Content-Disposition: form-data; name=\"{0}\"", "password"));
      contents.AppendLine();
      contents.AppendLine(this.Password.ToInsecureString());

      if (!String.IsNullOrEmpty(tweetMessage))
      {
        contents.AppendLine(header);
        contents.AppendLine(String.Format("Content-Disposition: form-data; name=\"{0}\"", "message"));
        contents.AppendLine();
        contents.AppendLine(tweetMessage);
      }

      contents.AppendLine(footer);

      byte[] bytes = Encoding.GetEncoding(encoding).GetBytes(contents.ToString());
      request.ContentLength = bytes.Length;

      using (Stream requestStream = request.GetRequestStream())
      {
        requestStream.Write(bytes, 0, bytes.Length);

        using (HttpWebResponse response = (HttpWebResponse)request.GetResponse())
        {
          using (StreamReader reader = new StreamReader(response.GetResponseStream()))
          {
            string result = reader.ReadToEnd();

            XDocument doc = XDocument.Parse(result);

            XElement rsp = doc.Element("rsp");
            string status = rsp.Attribute(XName.Get("status")) != null ? rsp.Attribute(XName.Get("status")).Value : rsp.Attribute(XName.Get("stat")).Value;

            return status.ToUpperInvariant().Equals("OK");
          }
        }
      }
    }

ASP.NET MVC and Classic ASP

mn.nospam@nospam.dnknormark.net (martinhn) — Mon, 02 Feb 2009 00:09:12 +0200

As I pointed out in my blog post a few days ago, I see several reasons why you should make the switch to ASP.NET MVC. On the other hand, you have to remember that ASP.NET WebForms has been around for many years, and is a very mature technology. We all know that new technology has its initial flaws.

I was fooling a little around with the technology, specifically some databinding where I had a DataTable and wanted to display a Table with the contents. My instant thought was: “This is like databinding back in the days of Classic ASP”…

Take a look at this:

<html>
<head>
<title>Get data from database</title>
</head>
<body>
<%
' Data connection
Set Conn = Server.CreateObject("ADODB.Connection")
DSN = "DRIVER={Microsoft Access Driver (*.mdb)}; "
DSN = DSN & "DBQ=" & Server.MapPath("/cgi-bin/database.mdb")
Conn.Open DSN

' SQL statement
strSQL = "Select * from Employees"
Set rs = Conn.Execute(strSQL)

Do

   Response.Write rs("Test") & "<br>"

   rs.MoveNext

Loop While Not rs.EOF

Conn.Close
Set Conn = Nothing
%>
</body>
</html>

Good old Classic ASP that is. You’ve got to admit that you had seamless control of the HTML that was being rendered!

Which is also the case with ASP.NET MVC. MVC offers you control and simplicity, as opposed to WebForms where you place a server control on the form, and sometimes you ended up with your HTML a complete mess…

Of course, MVC is a lot more maintainable and structured, than Classic ASP was. You instantly get a separation of concerns with the Model, View and Controller design pattern.

So we can fetch data in our Controller:

    public ActionResult Index()
    {
      DataTable dt = new DataTable();
      dt.Columns.Add("Column1");

      dt.Rows.Add("Test 1");
      dt.Rows.Add("Test 1");
      dt.Rows.Add("Test 1");
      dt.Rows.Add("Test 1");
      dt.Rows.Add("Test 1");
      dt.Rows.Add("Test 1");

      ViewData.Model = dt;
      
      return View();
    }

And display it, somewhat like we did back in the days of Classic ASP:

<% foreach (System.Data.DataRow item in ((System.Data.DataTable)ViewData.Model).Rows) 
   { %>
    <%= Html.Encode(item[0].ToString()) %><br />
<% } %>

To be honest, I actually likes going back to basics and getting full control of my HTML. Working in WebForms, you sometimes search for control settings to enable or disable the rendering of a certain piece of HTML. I think it is a lot easier to just give you full control.

But on the other hand, I could easily see a lot of the Classic ASP ‘bad habits’ return. Classic ASP can be *ugly* as hell, which means the maintainability is very poor. The same could happen with MVC if you aren’t disciplined enough. WebForms sort of put you in a box, but you could still make it look pretty ugly as well…

ASP.NET MVC Release Candidate links

ScottGu’s release candidate blog post

Stephen Walther’s guide to learning ASP.NET MVC

Is the use of AJAX on the web voting for a change to ASP.NET MVC?

mn.nospam@nospam.dnknormark.net (martinhn) — Fri, 30 Jan 2009 00:13:32 +0200

With ASP.NET MVC in Release Candidate now, the technology is getting a lot of attention these days. I personally haven’t given it any greater thought until just recently, but I have to admit that I’m tempted towards a change. I like ASP.NET WebForms. I think the WebForm approach provides us with great manageability and gives us a very rich set of controls.

But one thing that I don’t like about WebForms is, that it abstracts the anatomy of the web and how the web works. This was probably done to lure WinForms developers towards web development and making the change as easy as possible. But I think developers learning ASP.NET WebForms is missing out on the core web technology that is request, responses and headers etc. If you don’t know the core elements of what you’re working with, you can easily make some terrible mistakes.

As user experience is getting much more focus today, especially in the business of SaaS products, you have to deliver a great deal of responsiveness through the use of AJAX. In ASP.NET WebForms I see myself ending up rendering the page once, and doing AJAX callback through the core XmlHttpRequest in JavaScript when the user makes a change to some data. That’s to avoid postbacks. By doing this, I’m missing out on ViewState as well and what is left of ASP.NET WebForms now?

I think the high demand of quality (Testability) and user experience (Responsiveness = AJAX) is making ASP.NET WebForms obsolete.

I see several good reasons to dump WebForms for MVC:

- AJAX is used all over, so what good is postback, ViewState and the Page life cycle?
- WebForms is making your HTML look very ugly, especially because of the ViewState hidden field.
- ViewState can get very large, and slow down your pages.
- Performance is better on MVC –> You can handle more requests per second.
- The ability to test your code is better on MVC.

Initially, it looks like MVC is missing out on the rich set of controls that ASP.NET WebForms have. But I think this need will be driven by the awesome community that .NET have.

A buggy one: Introduction post

mn.nospam@nospam.dnknormark.net (martinhn) — Mon, 03 Nov 2008 22:08:35 +0200

Today, I was thinking about all those hours spent hunting and fixing bugs. I was hunting down a bug that was found after installing Flash Player 10 which resulted in Flash objects not being loaded. I will give the details shortly in the next post – this is just to clarify my intentions about upcoming posts, and the definition of them.

I thought it would be great, at least for myself, to have a place of some of the bugs that makes you laugh, cry, bang your head against the wall or whatever you do. I decided to write a blog post about bugs every once in a while when I find a “good” one. Bugs are never good, are they? I will post them in a category called ‘A buggy one’ and simply just tag them bugs, along with other relevant tags such as ASP.NET, HttpHandler, C# etc.

The bugs I share here, are not necessarily bugs in e.g. the .Net framework, Flash Player 10 or whatever – but bugs introduced by myself or another developer, causing our software not to work properly. No matter what you think about bugs, who you are and how good you think you are – you can just *NEVER* avoid bugs 100%. It is absolutely impossible.

I think the worst attitude you can have towards bugs, is to have so much pride about yourself and your work that you really think that you can write code 100% bug free.

Go hunt some bugs, first post is up for grabs shortly.

Technorati Tags: bugs

Early thoughts on Windows Azure

mn.nospam@nospam.dnknormark.net (martinhn) — Mon, 27 Oct 2008 17:53:11 +0200

So I just got home from work, and fired up my browser to check out the latest news from this year’s Microsoft PDC, by live streaming, and following the live blogging session. And as always, something big is underway. Windows Azure, is Microsoft’s answer to Amazon EC2. It is a cloud computing platform on Microsoft technology.

I’m truly thrilled by the thought. This gives me great possibilities. As an entrepreneur, I get loads of ideas. My biggest pain is always how to deploy this. Funding is a big issue, and this leads to compromises in server infrastructure and stuff. As a developer, I’m not that much focused on servers and thing alike, so to have a big platform is an advantage.

At PDC2008, they showcased a mobile app called Bluehoo. On stage, by editing an XML file online he scaled his platform from 1 to 20 nodes to be capable of the upcoming load due to the the announcement made at PDC.

My thoughts on pros:

Great scalability
Large content network = low latency around the world
Consolidation
Pay as you go premise – compared to buying hardware upfront
Easy application maturing – developing locally, deploying to a staging environment before deploying to production

Cons:

The inevitable: “What will they do with my data?”
Dependency on Microsoft
Unawareness – like deploying your app to ‘a black box’

On the Windows Azure site, the links to SDK’s and VS2008 tools are broken. They should be good shortly!

Update: For more information on Windows Azure, check out this blog post at Channel 10.

Or this: http://blogs.zdnet.com/microsoft/?p=1671

Technorati Tags: Software + Services,Windows Live,Windows Azure,cloud computing,PDC2008

Silverlight 2.0 provides better remote debugging experience

mn.nospam@nospam.dnknormark.net (martinhn) — Sun, 26 Oct 2008 01:25:04 +0200

Some months ago, I did some remote debugging of a Silverlight app on a Mac. To say the least, it wasn’t all that intuitive, and it required a lot of trying and guessing to achieve the desired goal. Personally I think remote debugging is cool, but I can’t really see the great need of it. Silverlight is cross-browser, cross-platform compatible – so when it works in a browser on Windows, shouldn’t it work on a Mac as well?

The release of Silverlight 2.0 has some great news in respect to remote debugging. It is no longer necessary to edit config files, and so on to authenticate the remote debugging session. This is all done from a nice user interface. JrzyShr Dev Guy has a excellent blog post describing the process of setting up remote debugging on a Mac on Silverlight 2.0. Furthermore, he suggests I mark my post as obsolete:

I ended up relying on a couple of random blog posts I came across on Live Search & Google to figure it out. If your search turns up any of these links, know that they are now obsolete:

Old 1.1 Alpha Way of Doing Things:
http://www.dnknormark.net/post/Windows-Vista-Mac-debugging-a-Silverlight-app-using-Visual-Studio-2008.aspx

So that I’ll do. Confusing information is all over the place, and I’d rather point someone in the right direction.

Technorati Tags: silverlight,web development,debugging

Windows Live Development Part 1 - Authenticating users via Windows Live ID

mn.nospam@nospam.dnknormark.net (MartinHN) — Thu, 19 Jun 2008 23:01:34 +0200

With so many online services out there, you nearly claim a new online identity somewhere every day. You tend to use the same username and password couple for all of them, but sometimes your preferred username is taken, or your password does not meet the sites' password policy. I reckon I have well over 100 online identities out there - I rely on the same 3-4 combinations of username and password, and it's really messy. In this blog post I'm going to cover how to authenticate users of an ASP.NET application against Windows Live, using the SDK for Windows Live ID. I'm going use the built-in ASP.NET Membership mechanism and go from there.

I could have gone the OpenID way, but Scott Hanselman already did an example on OpenID, so I thought I'd try out Windows Live. Jeff Atwood asked the question, if we really need another username and password in the pursuit of the best authentication model for stackoverflow.com.

Conceptual idea of using Windows Live ID

Windows Live ID, is the technology you should be quite familiar with when authenticating on Microsoft web sites, and Messenger. If you want to know more about what Windows Live ID is, take a look at this video from Channel9.

When using Windows Live ID, you redirect the user to the Windows Live login-page, with a return URL specified. When the user authenticates, Windows Live will redirect the user back to your site, and delivering a user id and a token. Here's an overview of how it looks:

Getting your application ID to use Windows Live ID

To get started using Windows Live ID, first of all you need an application ID. Basically you need to sign up for one, and register your application. Go to https://msm.live.com/app/default.aspx and click 'Register an Application'. (You can find Microsoft's guide to this here: http://msdn.microsoft.com/en-us/library/cc287659.aspx)

Leave Domain Name blank. Save your self-chosen Secret Key for later use, and submit the form. When you see the confirmation page, your application ID is shown. Copy that, as we're going to use it later.

In the return URL field, type the URL of your application followed by a name of a page to handle the Windows Live ID communication, like http://localhost/demoapp/webauth-handler.aspx.

Download and install Windows Live ID Web Authentication SDK

Microsoft has made it quite easy for ASP.NET developers to get started, by providing an SDK. Download it here: http://www.microsoft.com/downloads/details.aspx?FamilyId=24195B4E-6335-4844-A71D-7D395D20E67B&displaylang=en and install the SDK. By default it is installed here: C:\Program Files\Windows Live ID\WebAuth

From the SDK we need a class called WindowsLiveLogin, located in the App_Code folder of the sample. Keep this file in mind for a few seconds.

Start coding - create a new website

Open Visual Studio 2005/2008 and create a new ASP.NET website the way you want it. First we will add the WindowsLiveLogin.cs file to our App_Code folder. (C:\Program Files\Windows Live ID\WebAuth\Sample\App_Code\WindowsLiveLogin.cs).

Leave the default.aspx page, and put the following HTML in the markup:

<html>
<head>
  <title>Windows Live&trade; ID</title>
</head>
<body>
  <iframe id="WebAuthControl" name="WebAuthControl" src="http://login.live.com/controls/WebAuth.htm?appid=<%=AppId%>&style=font-size%3A+10pt%3B+font-family%3A+verdana%3B+background%3A+white%3B"
    width="80px" height="20px" marginwidth="0" marginheight="0" align="middle" frameborder="0"
    scrolling="no"></iframe>
  <%   1:  if (UserId == null)
   2:      { 
%>
  <p>
    This application does not know who you are! Click the <b>Sign in</b> link above.</p>
  <%   1:  }
   2:      else
   3:      { 
%>
  <p>
    Now this application knows that you are the user with ID = "<b><%   1: =UserId
%></b>".</p>
  <%   1:  } 
%>
</body>
</html>

And this piece of code in its code-behind class.

using System;
using System.Web;
using System.IO;
using WindowsLive;

public partial class DefaultPage : System.Web.UI.Page
{
  const string LoginCookie = "webauthtoken";

  // Initialize the WindowsLiveLogin class.
  private static WindowsLiveLogin wll = new WindowsLiveLogin(true);
  
  protected static string AppId = wll.AppId;
  protected string UserId;

  protected void Page_Load(object sender, EventArgs e)
  {
    HttpRequest req = HttpContext.Current.Request;
    HttpApplicationState app = HttpContext.Current.Application;

    HttpCookie loginCookie = req.Cookies[LoginCookie];

    if (loginCookie != null)
    {
      string token = loginCookie.Value;

      WindowsLiveLogin.User user = wll.ProcessToken(token);

      if (user != null)
      {
        UserId = user.Id;
      }
    }
  }
}

Default.aspx will work as a login page. Users click the Sign-in button, and is redirected to Windows Live. When they authenticate, they're redirected back to our handler page, that we will now create...

Create a new aspx page, and name it webauth-handler.aspx. Recall that this was the page we provided as the return URL when we registered the application at Windows Live.

This page will serve as the handler page for the user, when Windows Live redirects back to your site. Windows Live will post authentication specific values for you to use. Because this page is never seen by a user (we will send the user back where they began), we can simply remove all the markup.

Put the following code in your code-behind class for webauth-handler.aspx.

using System;
using System.Web;
using System.IO;
using WindowsLive;

public partial class HandlerPage : System.Web.UI.Page
{
  // Relative path to the login- and logoutpage.
  const string LoginPage = "default.aspx";
  const string LogoutPage = "default.aspx";
  const string LoginCookie = "webauthtoken";
  static DateTime ExpireCookie = DateTime.Now.AddYears(-10);
  static DateTime PersistCookie = DateTime.Now.AddYears(10);

  // Initialize the WindowsLiveLogin class.
  static WindowsLiveLogin wll = new WindowsLiveLogin(true);

  protected void Page_Load(object sender, EventArgs e)
  {
    HttpRequest request = HttpContext.Current.Request;
    HttpResponse response = HttpContext.Current.Response;
    HttpApplicationState application = HttpContext.Current.Application;

    string action = request["action"];

    if (action == "logout")
    {
      HttpCookie loginCookie = new HttpCookie(LoginCookie);
      loginCookie.Expires = ExpireCookie;
      response.Cookies.Add(loginCookie);
      response.Redirect(LogoutPage);
      response.End();
    }
    else if (action == "clearcookie")
    {
      HttpCookie loginCookie = new HttpCookie(LoginCookie);
      loginCookie.Expires = ExpireCookie;
      response.Cookies.Add(loginCookie);

      string type;
      byte[] content;
      wll.GetClearCookieResponse(out type, out content);
      response.ContentType = type;
      response.OutputStream.Write(content, 0, content.Length);

      response.End();
    }
    else if (action == "login")
    {
      HttpCookie loginCookie = new HttpCookie(LoginCookie);

      WindowsLiveLogin.User user = wll.ProcessLogin(request.Form);

      if (user != null)
      {
        loginCookie.Value = user.Token;

        if (user.UsePersistentCookie)
        {
          loginCookie.Expires = PersistCookie;
        }
      }
      else
      {
        loginCookie.Expires = ExpireCookie;
      }

      response.Cookies.Add(loginCookie);
      response.Redirect(LoginPage);
      response.End();
    }
    else
    {
      response.Redirect(LoginPage);
      response.End();
    }
  }
}

The first login test

Now we've got everything setup as we should, and it should be possible to login using Windows Live. View the default.aspx page in your favorite browser, and let's try.

First you will see out login page:

Click the Sign in link, which will take you to Windows Live:

Sign in with your Windows Live ID, and you will return to our login page - which will now welcome you and show your user ID:

Now your user has authenticated against Windows Live, using the Windows Live ID. Now it's up to you to handle the user's Windows Live ID, which is what I'm going to blog about in part 2 of the Windows Live Development series here...

Designing for Internationalization

mn.nospam@nospam.dnknormark.net (MartinHN) — Tue, 10 Jun 2008 20:17:05 +0200

Web sites and Web Applications today, are very often exposed beyond the borders of your home country, and therefore users speak different languages, and has a different currency, date- and time formats etc. ASP.NET provides you with an entire namespace for handling things like this. That is the System.Globalization namespace, where you will find a lot of classes for handling your every day globalization tasks. I'm not going to cover anything in this namespace now, if you want to get your hands dirty take a look at this video: http://asp.net/learn/videos/video-40.aspx where you will see how to use local and global resources for your application.

Using a global resource file for a place to store display text on buttons, labels, validation controls etc. is fine. But if you have an e-commerce site, selling products in multiple regions with different languages, you need an extra level. That level is a way to globalize e.g. the name and description of your product. When a user changes language, the name and description of your "display product details page" should change accordingly.

For me, the ideal solution should not result in extra database columns like name_us, name_da, name_es. This would be a very static solution, as you would have to change your database whenever a new language is added to your application. Nor should it require extra tables, so you need to join like hell, when you need to select a product.

I've decided to store e.g. the Name values as XML in the database, and parse that XML into a Dictionary<string, string> property on my Product object, with the key of the Dictionary being the language code (en-US for US English). To me this seems to work just fine. My database design is not getting more complex, and I can get and set values quite easily.

The XML string that goes into my ProductName column in the database table, looks like this:

<cultures><culture code="en-us">Logitech SmartCam 124</culture></cultures>

On my Product object, the Name property is a Dictionary:

private Dictionary<string, string> _Name;

public Dictionary<string, string> Name
{
  get { return _Name; }
  set
  {
    if (_Name != value) MarkDirty("Name");
    _Name = value;
  }
}

When I need to get the US English value of the Product.Name property, I call:

p.Name["en-us"]

To get a Dictionary from my XML, I use this helper method:

public static Dictionary<string, string> GetDictionary(string xml)
{
  if (String.IsNullOrEmpty(xml))
    return new Dictionary<string, string>(owner);
    
  XmlDocument doc = new XmlDocument();
  doc.LoadXml(xml);
    
  Dictionary<string, string> dic = new Dictionary<string, string>();

  foreach (XmlNode node in doc.DocumentElement)
  {
    dic.Add(node.Attributes["code"].Value, node.InnerText);
  }

  return dic;
}

And when I want to update my database, after I change the Name of the Product, I convert to Dictionary to an XML string using this helper method:

public static string GetXmlDocument(Dictionary<string, string> dic)
{
  XmlDocument doc = new XmlDocument();
  XmlNode docElement = doc.CreateNode(XmlNodeType.Element, "cultures", "");

  foreach (string key in dic.Keys)
  {
    XmlNode node = doc.CreateNode(XmlNodeType.Element, "culture", "");
    XmlAttribute att = doc.CreateAttribute("code");
    att.Value = key;
    node.Attributes.Append(att);

    node.InnerText = dic[key];

    docElement.AppendChild(node);
  }

  doc.AppendChild(docElement);

  return doc.OuterXml;
}

If you need more information on this huge topic, take a look at the ASP.NET Wiki: http://wiki.asp.net/page.aspx/55/internationalization/

ASP.NET Server Control: FeedList - List most recent items from an RSS feed as links

mn.nospam@nospam.dnknormark.net (MartinHN) — Mon, 21 Apr 2008 21:30:38 +0200

RSS feeds is everywhere, and rightly so. Just about any modern browser, e-mail client, search engine or mobile phone is compliant with the RSS technology, and it gives web developers an easy and very convenient way to share and consume content across the web. From time to time I need to show a list of recent items form an RSS feed. This is a very easy thing to do, and I always end up copying the same usercontrol around, which becomes nothing but a mess! Eventually I need to change a few things, which means I get different versions of the same control spread throughout my projects.

This mess has got to stop, and from now on I'll put all my common controls in a class library project, which I can reference from all the other projects.

Back to my RSS feed reader - this is the first control I've made into a server control, from one of my many usercontrols around the projects folder. The control downloads the XML at the Url of the RSS feed, and renders a list of links. The controls contains a few properties:

FeedUrl: The Url of the RSS feed to download
NumberOfItems: The number of items which should be rendered. Defaults to 5 if none is specified.
CssClass (Part of the WebControl class, from which FeedList derives): Use this so you have style the control from CSS.

The code

It really is straight forward. Below is the method that renders the links.

private void RenderFeedItems(HtmlTextWriter writer)
{
  if (!String.IsNullOrEmpty(FeedUrl))
  {
    XmlDocument doc = new XmlDocument();
    doc.Load(FeedUrl);

    XmlNodeList items = doc.DocumentElement.SelectNodes("channel/item");

    if (items != null)
    {
      int bounce = NumberOfItems;
      if (items.Count < NumberOfItems)
        bounce = items.Count;

      for (int i = 0; i < bounce; i++)
      {
        XmlNode titleNode = items[i].SelectSingleNode("title");
        XmlNode urlNode = items[i].SelectSingleNode("link");
        XmlNode pubDateNode = items[i].SelectSingleNode("pubDate");

        string title = String.Empty;
        string url = String.Empty;
        DateTime pubDate = new DateTime();

        if (titleNode != null)
          title = titleNode.InnerText;

        if (urlNode != null)
          url = urlNode.InnerText;

        if (pubDateNode != null)
          DateTime.TryParse(pubDateNode.InnerText, out pubDate);

        writer.RenderBeginTag(HtmlTextWriterTag.Li);

        writer.AddAttribute(HtmlTextWriterAttribute.Href, url);
        writer.AddAttribute(HtmlTextWriterAttribute.Title, title);
        writer.RenderBeginTag(HtmlTextWriterTag.A);
        writer.Write(title);
        writer.RenderEndTag();

        writer.RenderEndTag();
      }
    }
    else
    {
      RenderErrorMessage(writer, String.Format("No RSS-feed items found at: {0}", FeedUrl));
    }
  }
  else
  {
    throw new Exception("FeedUrl cannot be empty!");
  }
}

To use the control, put it in your website's App_Code folder, and add the following to the <controls> collection of your <pages> section in web.config:

<add tagPrefix="dnknormark" namespace="dnknormark" />

This should make your <pages> section look like this one:

<pages>
  <controls>
    <add tagPrefix="dnknormark" namespace="dnknormark" />
  </controls>
</pages>

And to put it on an aspx page, and show the latest 5 items from this blog, add this line to the page:

<dnknormark:FeedList ID="lstFeeds" runat="server" CssClass="feedList" 
NumberOfItems="5" FeedUrl="http://www.dnknormark.net/syndication.axd" />

Use it as you'd like. Feel free to change it, the way you need.

Adding a Website to IIS7 programmatically

mn.nospam@nospam.dnknormark.net (MartinHN) — Wed, 16 Apr 2008 19:32:06 +0200

Some time ago, I blogged about Adding an Application Pool to IIS7 programmatically. The result was a new Application Pool, that uses the Integrated Pipeline in IIS7.

In this post I will show you how to add a new Website, that uses the Application Pool from the other blog post.

I've loaded the Console Application I used to add the Application Pool, and moved the logic from Main to a method called AddApplicationPool, to split it up nicely. It is actually even easier to add a website programmatically. Below is the code for doing just that:

private static void AddWebSite()
{
  ServerManager mgr = new ServerManager();

  if (!Directory.Exists(@"c:\inetpub\wwwroot\iis7test"))
  {
    Directory.CreateDirectory(@"c:\inetpub\wwwroot\iis7test");
  }

  // Add a new Site to the server, configured to use our the iis7test home directory.
  Site site = mgr.Sites.Add("MyWebSite", @"c:\inetpub\wwwroot\iis7test", 80);
  
  // Set the application pool name of the site, to use the MyAppPool application pool.
  site.ApplicationDefaults.ApplicationPoolName = "MyAppPool";

  // Clear all bindings.
  site.Bindings.Clear();

  // Make the site listen to incoming HTTP requests using host header iis7test, on port 80.
  site.Bindings.Add("*:80:iis7test", "http");
  
  // Set auto start to true.
  site.ServerAutoStart = true;

  // Commit the changes
  mgr.CommitChanges();
}

Notice how we add Bindings to the website. Bindings is the information that tells IIS7 when to serve our website. We use this string to configure bindings: *:80:iis7test. The first * tells IIS to listen on all IP addresses on your system. 80 is the port number, and iis7test is the host header value for this site.

To browse our website, we need to add iis7test to the computers hosts file (located in %WINDIR%\System32\Drivers\etc), and point it to 127.0.0.1.