Network Programming in .NET

NuGet seems to be a great way to get windows developers to use your code, it makes it easy to package up an assembly to be used in another project. Seconds after uploading, I started to get developers downloading my code.

Here was a problem that I hit when I tried to package my assembly, there may be a better solution, but this worked for me.

I created a boiler-plate NuSpec file from my DLL by running

nuget spec MyAssembly.dll

I had to open the nuspec file in notepad, and edited the fairly obvious fields, like URL description, etc. Then, I went to pack the nuspec and dll together with this command:

NuGet Pack YourPackage.nuspec

However, I then got a warning:

WARNING: 1 issue(s) found with package ‘xxxx.dll’.

Issue: Assembly outside lib folder.
Description: The assembly ‘xxx.dll’ is not inside the ‘lib’ folder
and hence it won’t be added as reference when the package is installed into a project.
Solution: Move it into the ‘lib’ folder if it should be referenced.

The nupkg file was created, and I uploaded it to nuget.org, however, when I tried to download this again, using

PM> Install-Package xxxxx.dll
Installing ‘xxxxx.dll 1.0.0’.
Successfully installed ‘xxxx.dll 1.0.0’.

There was no reference added to my project, so, I couldn’t access the assembly.

So, after reading a bit, I edited my nuspec file, and added this node in the metadata

<references>
<reference file=”xxxx.dll” />
</references>

created a lib folder, and copied the dll into it. – and changed the version number

I repacked, and re-uploaded the nupkg, and when I downloaded it again, I got a reference automatically added to my test project.

And it’s all live on https://www.nuget.org/packages/OfflineAdvertsLib.dll

When porting some code from Silverlight 4 to Windows Phone 7.1, I  got an unusual error “Could not load type ‘System.Action’ from assembly ‘mscorlib, Version=3.7.0.0, Culture=neutral, PublicKeyToken=969DB8053D3322AC’.” – At this bit of code:

Deployment.Current.Dispatcher.BeginInvoke(someMethod);

The code is designed to run “someMethod” on the main UI Thread, so you don’t an illegal cross thread exception.

Invoking Dispatcher.BeginInvoke with a function name like this, is interpreted as an Action, rather than a delegate, so the solution is to use a delegate as follows:

private delegate void someMethodDelegate();

private static void SomeMethodCalledByBackgroundThread()
{
someMethodDelegate action = someMethod;
Deployment.Current.Dispatcher.BeginInvoke(action);
}

Although a little less easy to read, and it appears to have an extra level of indirection, this code works in both SL4 and WP7

The data uri is a string based representation of an image, which is base-64 encoded, so it’s easy to store in Javascript, and pass it in Post data. It bloats the image size, so it isn’t an efficient format, but it’s certainly easy to handle.

Here is an example of how to render a image from a data URI in Silverlight – which should also be suitable for Windows Phone 7, 8, and 8.1 (Silverlight).

Here’s my XAML:

<Grid x:Name=”LayoutRoot” Background=”White” Loaded=”LayoutRoot_Loaded”>
<Image x:Name=”imgDisplay”></Image>
</Grid>

And this is the underlying C# code:

public static BitmapSource BitmapSourceForDataUri(Uri dataUri)
{
var dataString = dataUri.OriginalString;
var b64Start = dataString.IndexOf(“,”, StringComparison.Ordinal);
if (b64Start == -1) return null;
var base64 = dataString.Substring(b64Start + 1);
Stream stream = new MemoryStream(Convert.FromBase64String(base64));
var image = new BitmapImage();
image.SetSource(stream);
return image;
}

private void LayoutRoot_Loaded(object sender, RoutedEventArgs e)
{
const string strDataUri =
“data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAUAAAAFCAYAAACNbyblAAAAHElEQVQI12P4//8/w38GIAXDIBKE0DHxgljNBAAO9TXL0Y4OHwAAAABJRU5ErkJggg==”;
var bmp = BitmapSourceForDataUri(new Uri(strDataUri));
imgDisplay.Source = bmp;
imgDisplay.Width = bmp.PixelWidth;
imgDisplay.Height = bmp.PixelHeight;
}

If you use Phonegap, and want to know how many people use your app regularly, and where they come from, then Google Analytics, with a small modification, is great for this purpose. – But you need a slight modifiction to the normal tracking code.

Firstly, download the ga.js file from http://www.google-analytics.com/ga.js – open it in a text editor and look for the word file:, and replace it with _file then stick it into your www folder. make a reference to it from every page you want to track, then use the script:

<script type="text/javascript" src="ga.js"></script>
    <script>
 var _gaq = _gaq || [];
    _gaq.push(['_setAccount', 'UA-YOUR_ID_HERE']);
    _gaq.push(['_setDomainName', 'none']);
    _gaq.push(['_trackPageview', 'NAME_OF_PAGE']);
    </script>

To get your UA-ID then you have to go to Google Analytics, create a new property – selecting “WEBSITE” not “APP”, and pick it out of the code.

To test it, then you run the app, and look at Google Analytics “Real time”, if you see activity appearing, then it’s working!

If your server has multiple IP addresses, you can specify which IP address to make a HTTP request from in C#. This is useful if you want to need to ensure that requests do or do not originate from a particular IP address. This little console application demonstrates this, by listing available IP addresses, then making a request to an external website to reveal the outbound IP address as visible to the outside world.

private static IPAddress selectedIP;
static void Main(string[] args)
{
var ips = LocalIPAddress();
Console.WriteLine(“Select an IP Address”);
for (var i = 0; i < ips.Count(); i++)
{
Console.WriteLine(i + “. ” + ips[i]);
}
var choice = Console.Read() – 48 ;
selectedIP = ips[choice];
const string strUrl = “http://myexternalip.com/raw&#8221;;
var http = GetWebRequest(new Uri(strUrl));
var stream = http.GetResponse().GetResponseStream();
if (stream != null)
{
var sr = new StreamReader(stream);
Console.WriteLine(sr.ReadToEnd());
}
Console.ReadLine();
}

protected static WebRequest GetWebRequest(Uri uri)
{
var request = (HttpWebRequest)WebRequest.Create(uri);
request.ServicePoint.BindIPEndPointDelegate = BindIPEndPointCallback;
request.Proxy = null;
return (request);
}

public static IPEndPoint BindIPEndPointCallback(ServicePoint servicePoint, IPEndPoint remoteEndPoint, int retryCount)
{
return new IPEndPoint(selectedIP, 0);
}

private static List<IPAddress> LocalIPAddress()
{
if (!System.Net.NetworkInformation.NetworkInterface.GetIsNetworkAvailable())
{
return null;
}
var host = Dns.GetHostEntry(Dns.GetHostName());
return host
.AddressList
.Where(ip => ip.AddressFamily == AddressFamily.InterNetwork).ToList();
}

To see it, you’ll need to run this on a multi-homed machine, type the number beside the required IP address, and see the external IP appear momentarily afterwards.

The FileReader object is a new feature of HTML5, and it allows handling of file-data on the client side of a web-application, so for instance, you could provide extra validation on an uploaded file before sending it to the server. 

Here’s a simple example of it’s use:

<html>
<head>
<script src=”http://code.jquery.com/jquery-1.11.0.min.js”></script&gt;
<script language=”javascript”>
$(init);
function init()
{
console.log(“Good to go!”);
$(“#myFile”).change(myFile_change);
}
function myFile_change(evt)
{
console.log(“myFile_change”);
console.log(evt.target.files[0]);
var f = evt.target.files[0];
reader = new FileReader();
reader.readAsDataURL(f);
reader.onload = function(e) {
console.log(reader.result);
$(“#imgUpload”).attr(“src”,reader.result);
}
}
</script>
</head>
<body>
<input type=”file” id=”myFile”>
<img id=”imgUpload”>
</body>
</html>