CodeFrenzy

A couple of weekends ago, I was lucky enough to be able to go along to HackCamp at the Google offices in London together with the mflow team. It’s the first ‘hack’ day/camp I’ve been to – and although I didn’t stay for the whole weekend nor did I actively participate in the hacking, I got to see what a hack camp is all about. For those of you who have never heard of this term before, a hack day/camp is basically a congregation of hackers, developers, designers (geeks in other words) who get together and spend 24~48 hours building an application using whatever tools they have at their disposal, in small groups or alone.

There were about 10 or so different speakers presenting their APIs at HackCamp, and although you didn’t have to use those APIs in particular, it was certainly very interesting hearing about what others have to offer. One of the APIs that I thought was a great idea was this one – 0870.me. 0870.me is an API that allows you to plug in a standard 08* number (which mobile networks charge a fortune for here in the UK) and it will return to you a matching non-08* number if it can find one in it’s database. How neat is that??

There were speakers from Twitter, Google AppEngine and Buzz to name a few big ones, but also lots of smaller companies. Tum and Demis from mflow gave a talk about the mflow API.

Once all the hacked-up apps had been submitted, I had a quick look and found this little gem: tron. It is very cool to see that someone has used the mflow API in their app!

Anyway, having had a small taste of these ‘hack’ camps, I may just have to keep my eyes out for another that I can go to coz they sure do look like they could be lots of geeky-fun :)

It’s been ages since I last updated my blog and I’m feeling super guilty.

So I thought I’d get stuck into it again and start with a whinge – a whinge about something that had me stumped for way too long this morning.

I have to say that I hardly ever use the out-of-the-box ASP.NET server controls (like asp:DropDownList, asp:TextBox, asp:DataGrid, etc etc) because I prefer to use standard HTML controls, stick runat=”server” tags on them and handle them manually. Call me old-fashioned, tell me I’m doing it the “hard” way… I don’t care. It works and I know how to do it quickly. I always find that when I use the more complete and complex server controls, I get 95% of the work done quickly and then spend hours trying to get the last 5% to work. I don’t like that very much.

Anyway, sometimes I end up having to use a variety of these server controls and today was one such day. Everything was going swimmingly until I tried to insert / update a DateTime property on the type that I had bound my ListView control to, using an ObjectDataSource. I’m in the UK and here, we do not use US-formatted dates. For some reason, no matter where I put localization code to set the locale to “en-GB”, when I hit my “Save” button, my code would throw an exception because it could not convert the string representation of the date into a DateTime object. That’s because it was trying to convert it to an American date. Grr!!!

After a lot of Googling, I came across this article. And there was the answer I was looking for.

Turns out that this is a known bug since ASP.NET 2.0 that Microsoft have decided they “won’t fix”. WTF. They’ve decided not to fix it because ”... the potential fix is relatively dangerous and the workaround is relatively simple…”.

The bug occurs because behind the scenes, .NET is calling ConvertFromInvariantString on all the new values – which is the “en-US” culture. This is clearly a massive assumption on their part but alas, it seems it is “too risky” to fix this.

The “simple workaround” they suggest is to handle the OnItemInserting and OnItemUpdating or OnRowUpdating events and in those handlers, extract the date values from the EventArgs, parse them from strings into DateTime objects and save them back. The following is an example of how one might do this:

1
2
3
4
5
6
7
8
9
10
11

private static void MyList_ItemInserting(object sender, ListViewInsertEventArgs e)
{
	var myDate = e.Values["MyDate"].ToString();
	e.Values["MyDate"] = DateTime.Parse(myDate);
}
 
private static void MyList_ItemUpdating(object sender, ListViewUpdateEventArgs e)
{
	var myDate = e.Values["MyDate"].ToString();
	e.Values["MyDate"] = DateTime.Parse(myDate);
}

I realize this is not a hard fix – but the annoying part was figuring out why it was happening in the first place!!

Anyway, I hope this whinge helps someone else solve this problem… :)

In my chosen career, I solve problems everyday. Sometimes the problems I’m given are small and simple to solve, and sometimes they are most definitely not. That, I can deal with. What annoys me is when a problem seems like it should only take five or so minutes to solve and yet somehow, it takes you all day. It’s not that you’re bad at what you do, it’s just that the thing you’re trying to do wasn’t quite as simple as you or anyone else had thought it would be. Or perhaps the framework you’re working with has some nasty little quirks that you weren’t aware of. Or perhaps the thing you’re trying to do just hasn’t been done before, or at least hasn’t been well documented. Whatever the reason, spending all day on a “simple problem” isn’t good for your sanity and makes you feel like you ought to be doing COMP101 all over again.

One such problem that I’ve come across more than once in the last few weeks is how to make a text string fit inside a label or other similar control. If the string is too long to fit, the end should be truncated and replaced by an ellipsis. Sounds pretty straightforward eh? Nope, think again.

After coming up with a few complicated solutions including but not limited to iterating through every character in the string and measuring the length of the string (yes, highly inefficient, I know!), I finally found a solution which seems to do the trick. I feel this amazing solution needs to be shared with everyone so here it is:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23

private string textToDraw = "Hello, how are you? Why does this string not fit in the label that I have drawn? Maybe it is because it is too long?";
 
public Form1()
{
    InitializeComponent();
    label1.Text = ShortenString(textToDraw, this.Width - 10, label1.Font);
}
 
public string ShortenString(string myString, int width, Font font)
{
    string result = string.Copy(myString);
    TextRenderer.MeasureText(result, font, new Size(width, 0), TextFormatFlags.EndEllipsis | TextFormatFlags.ModifyString);
 
    return result;
}
 
private void Form1_ResizeEnd(object sender, EventArgs e)
{
    // Set the width to something - I've just set it to the width of the form - 10 to allow for the ellipsis but
    // I guess this would need to depend on which control the label is to be displayed inside of
    int width = this.Width - 10;
    label1.Text = ShortenString(textToDraw, width, label1.Font);
}

The trick to solving this problem was to use the TextRenderer class and calling its MeasureText function. MeasureText doesn’t seem like the most obvious function to use but it seems to work.

Note that this example is for a C# winforms application. The same could be done for a web application although you’d have to use JavaScript to re-calculate the string when the browser or element is resized.

Has anyone else ever had to solve this particular problem? If so, I’m interested to hear how you did it!

I know this is not rocket science but it’s actually something I’d never had to do until recently. Normally I populate my drop down lists (and repeaters, etc etc) using data from databases, not XML.

However, fear not. Using XML as a datasource in .NET is super easy. Here is one way to do it:

1
2
3
4
5
6
7
8

XmlDataSource source = new XmlDataSource();
source.DataFile = Server.MapPath("Values.xml");
source.XPath = "ParentElement/ValueList/ValueItem";
 
myDropDownList.DataSource = source;
myDropDownList.DataTextField = "Text";
myDropDownList.DataValueField = "Value";
myDropDownList.DataBind();

And your xml file, Values.xml, may look something like:

<?xml version="1.0" encoding="utf-8"  standalone="yes" ?>
<ParentElement>
  <ValueList>
    <ValueItem Text="Item 1" Value="1" />
    <ValueItem Text="Item 2" Value="2" />
  </ValueList>
</ParentElement>

Happy coding!

The following article is part 4 and therefore the last in a 4 part series of articles about developing .NET applications using SSL. Part 1 in this series can be found here, part 2 can be found here and part 3 can be found here.

How to set up a web application to automatically redirect to HTTPS if accessed via HTTP when SSL is required

SSL can be “required” at the site level, folder level or even file level.

If any file that has SSL set to “required” is requested via HTTP, HTTP Error 403;4 “Forbidden” will be returned automatically by IIS.  As far as I can tell, there is no simply way to tell IIS to re-direct to the same page via HTTPS.  This effect can be achieved in two different ways and if “defence in depth” is the preferred approach then both solutions can be implemented together.  Note that both solutions require custom code to be written however.

Redirect with IIS and code:

1. Create a new Web Form (aspx) in your Web Application called RedirectToSSL (or similar).
2. In the Page_Load event handler, insert the following code:
   
   
   ^?View Code CSHARP

   1 2 3 4 5 6 7 8

   string redirectUrl = HttpContext.Current.Request.Url.Query; redirectUrl = redirectUrl.Substring((redirectUrl.LastIndexOf(';')+1) , redirectUrl.Length - (redirectUrl.LastIndexOf(';')+1));   System.UriBuilder secureUrlBuilder = new System.UriBuilder(redirectUrl); secureUrlBuilder.Scheme = System.Uri.UriSchemeHttps; secureUrlBuilder.Port = -1; // Use default port HttpContext.Current.Response.Redirect(secureUrlBuilder.Uri.ToString(), true);

3. The aspx needs to be at least 512 bytes, otherwise IIS won’t redirect to it. So make sure you don’t delete everything out of it, it should never get displayed anyway.
4. In IIS, right-click your Web Site and go to Properties > Custom Errors.
5. Select the “403;4” HTTP error in the list HTTP errors and click the Edit Properties button.
6. Select “URL” from the Message type drop-down list and enter “/RedirectToSSL.aspx” into the URL textbox, as shown below.
7. Press OK twice to close the dialogs down.
8. Restart the website and/or reset IIS to ensure all settings have been applied.
9. Open Internet Explorer and go to: https://<servername>/<securedpage> and verify that you are automatically redirected to the https version of the same page.

1. Insert the following code at the beginning of the Page_Load event handler of every page that needs to be secured:
   
   
   ^?View Code CSHARP

   1 2 3 4 5 6 7 8 9

   // Make sure the request has come through SSL (ie that HTTPS has been used) if (!HttpContext.Current.Request.IsSecureConnection) { System.Uri currentUrl = HttpContext.Current.Request.Url; System.UriBuilder secureUrlBuilder = new System.UriBuilder(currentUrl);   secureUrlBuilder.Scheme = System.Uri.UriSchemeHttps; secureUrlBuilder.Port = -1; HttpContext.Current.Response.Redirect(secureUrlBuilder.Uri.ToString(), true); }




• 

The following article is part 3 in a 4 part series of articles about developing .NET applications using SSL. Part 1 in this series can be found here and part 2 can be found here.

Installing Certificates into the correct Certificate Stores

Applications depend heavily on server SSL and client certificates being installed in the correct certificate stores. Certificates may need to be copied and pasted or exported and imported into stores in order to get the systems correctly working together in a secure manner.

Client application (with client certificate):

1. Local Computer > Trusted Root Authority store should contain the Root CA certificate of the Server SSL Certificate Issuer of the Web application.
2. Current User > Personal store should contain the Client Certificate for the Client application.
3. Current User > Trusted Root Authority store should contain the Root CA certificate of the Client Certificate Issuer of the Client application

1. Local Computer > Personal store should contain the Server SSL Certificate for the Web application.
2. Local Computer > Trusted Root Authority store should contain the Root CA certificate of the Server SSL Certificate Issuer of the Web application.
3. Local Computer > Trusted Root Authority store should contain the Root CA certificate of the Client Certificate Issuer of the Client application.
4. Current User > Trusted Root Authority store should contain the Root CA certificate of the Client Certificate Issuer of the Client application.

1. From the Start menu, select Run, type “mmc” and click OK. This will open the Microsoft Management Console.
2. Click on the File menu option then on “Add/Remove Snap-in”.
3. In the Add/Remove Snap-in dialog, click the Add button.
4. From the Add Standalone Snap-in dialog, select “Certificates” from the list of snap-ins, and click Add.
5. Select the “Computer account” option and click Add. Select “Local computer” and click Finish.
6. If you are logged in as the user that will be used to run the client application, add the “My user account” as well.
7. Now you will be able to look through the Certificates that are installed in all the different stores.

The following article is part 2 in a 4 part series of articles about developing .NET applications using SSL. Part 1 in this series can be found here.

Setting Up Client Certificates

Client certificates can be used to authenticate the client connecting to a web application. The client can either be a browser or a custom written client application.

Requesting a client certificate:

1. Open IE (Firefox didn’t work so well for me…) and type in: http://<certauthservername>/certsrv/
2. Click on the “Request a certificate” link.
3. Click on the “Web Browser Certificate” link.
4. Enter all required information into the form shown above click “Submit”

1. Log onto the Windows Server 2003 machine and open the Certification Authority dialog from Administrative Tools.
2. Expand the root certification authority “Pending Requests” folder.
3. The certificate you just requested should be in the “Pending Requests” folder. Right-click on this request and select All Tasks > Issue. The certificate request will be removed from the “Pending Requests” folder and appear in the “Issued Certificates”.

1. Go back to IE on your web server and type in http://<certauthservername>/certsrv/ again.
2. Click on the “View the status of a pending certificate request” link.
3. Click on the certificate request you want to view.
4. Click on the “Install this certificate” link.
5. Click “Yes” on the dialog that opens to install the client certificate.
6. Click “Yes” on the next dialog as well.

1. On the client machine, open a Microsoft Management Console with the Certificates snap-in. Export the client certificate without the private key. This will generate a .cer file. Copy this file over to the server hosting the web application.
2. On the server hosting the web application, right-click the website in IIS and select “Properties”.
3. Under Properties, go to the “Directory Security” tab.
4. Under the Secure Communications section, click the “Edit” button.
5. In the Secure Communications dialog, change the “Client Certificates” option to “Require client certificates” and check the “Enable client certificate mapping” checkbox.
6. Click on the “Edit” button in the “Enable client certificate mapping” section.
7. This will open up the Account Mappings dialog.
8. Click the “Add” button.
9. Select the client certificate .cer file recently exported from the server that hosts the client application or client browser.
10. Give the mapping a relevant name and choose the windows user that the web application will be run under.
11. Click OK.
12. Click OK to close the Account Mappings dialog and Secure Communications dialogs.
13. Remove all other authentication options in the IIS manager Directory Security setting by clicking on the “Edit” button under Anonymous access and authentication control in the Directory Security tab.
14. Uncheck all authentication options and press OK.
15. Open up a browser and verify that you cannot reach the web site unless you navigate to it from the machine that has the client certificate installed on it.

The following article is part 1 in a 4 part series of articles about developing .NET applications using SSL. These articles are based on my experience and I hope that they will help others develop secure applications as well.

Setting up a Development or Test SSL Server Certificate on a Website in IIS (5.0/6.0)

When developing an application that requires SSL you may want a test SSL certificate to develop against.  You can request test SSL certificates from companies such as Verisign however these only last 14 days (and you will be called up by Versign representatives asking you when you would like to buy the real deal).  A much simpler and more flexible solution is to issue your own test SSL and client certificates by setting up a Windows Server 2003 machine that has the Certification Services windows component installed.

First, create an SSL server certificate request:

1. In IIS, right-click the Upload web site and select “Properties”.
2. Under Properties, go to the “Directory Security” tab.
3. Click the “Server Certificate” button and then click Next.
4. Select the “Create a new certificate” option and click Next.
5. Select the “Prepare the request now, but send it later” option and click Next.
6. Go through the next few forms and fill in all required information.
7. At the end of the wizard you will be asked to save the certificate request as a text file. Do this so you can send this request to a CA later.
8. You will be shown a summary of your SSL certificate request. Check this thoroughly to make sure you have all the correct values.
9. Click Next to finish the IIS Certificate Request Wizard.

1. Open IE (Firefox didn’t work so well for me…) and type in: http://<certauthservername>/certsrv/
2. Click on the “Request a certificate” link.
3. Click on the “advanced certificate request” link.
4. Click on the “Submit a certificate request by using a base-64-encoded CMC or PKCS #10 file, or submit a renewal request by using a base-64-encoded PKCS #7 file” link.
5. Copy and paste the contents of the certificate request generated by IIS (certreq.txt) into the “Saved Request” text box.
6. Click the “Submit” button to request the certificate

1. Log onto the Windows Server 2003 machine and open the Certification Authority dialog from Administrative Tools.
2. Expand the root certification authority “Pending Requests” folder.
3. The certificate you just requested should be in the “Pending Requests” folder. Right-click on this request and select All Tasks > Issue. The certificate request will be removed from the “Pending Requests” folder and appear in the “Issued Certificates”.

1. Go back to IE on your web server and type in http://<certauthservername>/certsrv/ again.
2. Click on the “View the status of a pending certificate request” link.
3. Click on the certificate request you want to view.
4. Click on the “Download certificate” link. A “File Download” dialog will open. Click Save to save the new SSL certificate as a .cer file.
5. Go back to the Web Site properties in IIS and from within the Directory Security tab, click on the Server Certificate button.
6. Select “Process the pending request and install the certificate” and click Next.
7. Browse to the certificate file saved to disk in step 4 above that contains the SSL certificate issued to this server by your test Certification Authority.
8. Follow the wizard until the SSL certificate has been successfully installed.

1. Once an SSL server certificate has been installed on the website, go back to IIS > > Properties > Directory Security tab > Secure Communications section and click on the “Edit” button.
2. Check the “Require secure channel (SSL)” checkbox but leave the “Require 128-bit encryption” checkbox empty.
3. Leave the “Client Certificates” section set to “Ignore client certificates” and the other two checkboxes unchecked for now.
4. Close this window down and restart the website and/or reset IIS.
5. Open up a browser and verify that you cannot reach the web site by using http only, instead you must use https.

1. If the browser tells you that your certificate was issued by an untrusted CA when you navigate to your website, you will have to download the CA certificate from your test Certification Authority website and install it into the Trusted Root Authority store of the Local Computer.
2. You may also find that the browser may tell you that the certificate is not valid because it was issued to a different name – that will happen if you are accessing your site using http://localhost/ rather than the server name. The certificate was most likely issued with the actual server name and the browser will think that your site is being spoofed if the domain name in the browser and the certificate don’t match. Use the server name instead.

The other day while on a bus in Brisbane, I witnessed a classic error message FAIL:

I have just come across a strange problem when trying to disable some radio buttons on a web page using .NET and JavaScript. It seems to work well in Firefox but not IE, what a surprise!

The situation:

I have two sets of radio buttons – two RadioButtonList controls. The first set needs to control the second set – that is, when one of the radio buttons in the first set is selected, the other RadioButtonList control should be enabled, and when the other of the radio buttons in the first RadioButtonList control is selected, the second RadioButtonList control should be disabled.

In my code-behind, I had an event handler for the first RadioButtonList control to check if the selected index had changed. If so, check the value of the first RadioButtonList control and if it’s say equal to 1, then set the Enabled property of the second RadioButtonList control to false. Otherwise, set the Enabled property to true.

Obviously this alone doesn’t fix the problem unless the first RadioButtonList control is set to AutoPostBack = True. I needed some code on the client side to do this when while the user wasn’t posting the page back.

The code to disable some radio buttons is simple:

1
2
3
4

var secondRadioButtons = document.form1.secondRadioButtonList;
for(var i=0; i<secondRadioButtons.length;i++ ){
    document.form1.secondRadioButtons[i].disabled = false;
}

The problem:

When you wire all this up, it works fine in Firefox… But does it work in IE? No, of course not. So why not? I found that when I first loaded the page, the radio buttons were enabled / disabled as they should have been. Clicking on the first set of radio buttons was disabling / enabling the second RadioButtonList control as it should have. But as soon as I posted the page back (the page had a button that submitted a form and displayed some results), the second RadioButtonList control was completely disabled and no matter which of the first radio buttons I selected, I couldn’t re-enable it. I stuck some debug code in my JavaScript only to find that it was running correctly and firing when it should. As far as I could tell, the JavaScript was setting disabled = false (as above) on all of the radio buttons in the second RadioButtonList control, but they were still disabled. Why!?

I took a look at the source code for the page and noticed something that I thought might have been causing the problem I was experiencing. Basically, browsers renders an asp:RadioButtonList control like this:

1
2
3
4
5
6

<span id="secondRadioButtonList">
<input id="secondRadioButtonList_0" name="secondRadioButtonList" value="1" type="radio" />
    <label for="secondRadioButtonList_0">First radio button</label>
<input id="secondRadioButtonList_1" name="secondRadioButtonList" value="2" type="radio" />
    <label for="secondRadioButtonList_1">Second radio button</label>
</span>

If you disable a RadioButtonList in your code-behind, the same asp:RadioButtonList control will be rendered like this:

1
2
3
4
5
6
7
8

<span id="secondRadioButtonList" disabled="disabled">
    <span disabled="disabled">
<input id="secondRadioButtonList_0" name="secondRadioButtonList" value="1" disabled="disabled" type="radio" />
        <label for="secondRadioButtonList_0">First radio button</label>
<input id="secondRadioButtonList_1" name="secondRadioButtonList" value="2" disabled="disabled" type="radio" />
        <label for="secondRadioButtonList_1">Second radio button</label>
    </span>
</span>

This is precisely why I don’t like asp controls – they look nice in the html that you write yet when they render out to the browser, they’re filled with extra stuff like these two spans.

Anyway, the problem were these spans – obviously even though I was setting ‘disabled = true’ on all the radio buttons inside my asp:RadioButtonList control, there were a couple of spans around the radio buttons that were also disabled causing my radio buttons to appear disabled regardless of their individual status.

The solution:

My solution to this was to simply remove the code-behind that I had written to enable or disable the RadioButtonList control on the server side. My JavaScript is all I really need.

However, if you desperately want or need to have some server-side code also disabling your radio buttons, you should do so like this:

1

secondRadioButtonList.InputAttributes.Add(”disabled”, “disabled”);

rather than

1

secondRadioButtonList.Enabled = false;