I’m working on a small application that hooks into the clipboard monitor chain and displays a list of options whenever text is saved to the clipbaord.  One of the options I wanted was to be able to send data to pastebin.com, a popular text/code sharing service.  This is a relatively mundane task, faking the postdata for a webpage, and sending the requests as a browser.  There were some hicups though, mostly with Finding the redirect URL using HttpWebRequest/HttpWebResponse.

A little background:

HttpWebRequest and HttpWebResponse live in the System.Net namespace, and provide the ability to quickly send HTTP requests to a webserver.  This is what a simple GET request would look like:

try
{
 
    HttpWebRequest request  = (HttpWebRequest) WebRequest.Create("http://davux.pastebin.com/pastebin.php");
 
    HttpWebResponse res =(HttpWebResponse) request.GetResponse();
    if (res.StatusCode == HttpStatusCode.OK)
    {
          StreamReader reader = new StreamReader(res.GetResponseStream());
          Console.WriteLine(reader.ReadToEnd());
    } else {
         // handle errors
    }
 
    res.Close();
 
}
catch (Exception ex)
{
    Console.WriteLine("Error: " + ex.Message);
}

Note: For most simple tasks, it is wise to use System.Net.WebClient, which is not suitable for this task, because it does not expose the ability to disable automatic redirection.

To send new data to pastebin, it needs to be sent via HTTP POST.  Using Firefox LiveHTTPHeaders, I determined what string needs to be sent in order to create a new entry.  This can also be determined by reading the source of the page, and looking at each of the fields contained in the <form> block.

string post = "&parent_pid=&format=text&code2=" + code_text_to_send + "&poster=" + poster_name + "&paste=Send&expiry=" + expiry + "&email=";

• code_text_to_send - The data which should be used to create a new page.
• poster_name - The author of the post
• expiry - How long the data should be retained for.  Options are ‘d’ for one day, ‘m’ for one month, and ‘f’ for indefinitely.

Note: Make sure to URLEncode each field that is sent!

After further examination of the headers and data sent back, I decided that the best way to extract the link was to catch the HTTP 302 Document Found response.

HTTP/1.x 302 Found
Date: Thu, 08 May 2008 01:06:08 GMT
Server: Apache/1.3.33 (Debian GNU/Linux) mod_python/2.7.10 Python/2.3.4 PHP/4.3.10-22 mod_perl/1.29
X-Powered-By: PHP/4.3.10-22
Location: http://davux.pastebin.com/m10a794d6
Keep-Alive: timeout=15, max=100
Connection: Keep-Alive
Transfer-Encoding: chunked
Content-Type: text/html; charset=iso-8859-1

HttpWebRequest, by default, has AllowAutoRedirect set to true, which means that, internally, it will resolve this message, and fetch the new URL.  This extra operation is needless in this circumstance, so it is best to disable it and extract the URL from the header.  The headers associated with each request or response are found in the Headers object for each class.  We wish to look at the Location field to find our URL.
C# WebClient HTTP POST Example:

try
{
 
    HttpWebRequest request = (HttpWebRequest)
        WebRequest.Create("http://davux.pastebin.com/pastebin.php");
 
    request.AllowAutoRedirect = false;
    request.Method = "POST";
 
    string post = "&parent_pid=&format=text&code2=" + HttpUtility.UrlEncode(text) + "&poster=Dave&paste=Send&expiry=m&email=";
    byte[] data = System.Text.Encoding.ASCII.GetBytes(post);
 
    request.ContentType = "application/x-www-form-urlencoded";
    request.ContentLength = data.Length;
 
    Stream response = request.GetRequestStream();
 
    response.Write(data,0,data.Length);
 
    response.Close();
 
    HttpWebResponse res =(HttpWebResponse) request.GetResponse();
    res.Close();
    // note that there is no need to hook up a StreamReader and
    // look at the response data, since it is of no need
 
    if (res.StatusCode == HttpStatusCode.Found)
    {
        Console.WriteLine(res.Headers["location"]);
    }
    else
    {
        Console.WriteLine("Error");
    }
 
}
catch (Exception ex)
{
    Console.WriteLine("Error: " + ex.Message);
}

(Some error checking and extra code have been removed for the sake of clarity)

The values for expiry and author have been hard-coded in order to save space.  This basic example will create a new Pastebin entry on my not so private pastebin, and print the URL of the entry to the console.  If there is an error, it will print the exception message.  If the HTTP code returned is not 302/Found, it will print a generic error.  (If the server is not returning a 302 for this request, there is nothing codewise that can be fixed, and thus the only option for error is generic.)

P/Invoke, or Pinvoke stands for Platform Invocation Services.  PInvoke is a feature of the Microsoft .NET Frameowrk that allows a developer to make calls to native code inside Dynamic Link Libraries (DLL’s).  When Pinvoking, the .NET framework (or Common Language Routine) will load the DLL and handle the type conversions automatically.  The most common use of P/Invoke is to use a feature of Windows that is only contained in the Win32 API.  The API in Windows is extremely extensive and only some of the features are encapsulated in .NET libraries.  For example, Form.Show(); is really a wrapper for the ShowWindow() API found in shell32.dll.

ShowWindow Declaration API for C#:

[DllImport("user32.dll")]
static extern bool ShowWindow(IntPtr hWnd, int nCmdShow);

Note the use of the DllImport attribute.  This attribute, coupled with extern tells the .NET CLR where it can find the native declaration for ShowWindow.  shell32.dll must be registered on the system for this to work properly.

Note: Remember import the System.Runtime.InteropServices namespace!  This is where the [DLLImport] attribute is found.

using System.Runtime.InteropServices;

How are P/Invoked methods called?

Methods that are loaded in using DLLImport are Pinvoked by simply treating them as native .NET functions, and calling them as you would call any other.  Win32 API do tend to use data types that are not used often in .NET, for example - the IntPtr datatype, which is actually a pointer to an integer–in this case, a Window Handle.  nCmdShow is an integer that represents the show command.  But what values of nCmdShow mean what?  These can either be found on MSDN, or PInvoke.net.

C# Constants for ShowWindow in shell32.dll:

  const int        SW_HIDE            = 0;
  const int        SW_SHOWNORMAL      = 1;
  const int        SW_NORMAL          = 1;
  const int        SW_SHOWMINIMIZED   = 2;
  const int        SW_SHOWMAXIMIZED   = 3;
  const int        SW_MAXIMIZE        = 3;
  const int        SW_SHOWNOACTIVATE  = 4;
  const int        SW_SHOW            = 5;
  const int        SW_MINIMIZE        = 6;
  const int        SW_SHOWMINNOACTIVE = 7;
  const int        SW_SHOWNA          = 8;
  const int        SW_RESTORE         = 9;
  const int        SW_SHOWDEFAULT     = 10;
  const int        SW_FORCEMINIMIZE   = 11;
  const int        SW_MAX             = 11;

These constants can be used to call ShowWindow().  So in our origional example of Form.Show(), the P/Invoked way of doing this would be:

ShowWindow(Form.Handle, SW_SHOW);

This isn’t nearly as clean as the Object-Oriented way of using Form.Show(), but this is what is actually happening under the hood of .NET.

Why would I ever want something so overly complicated?

ShowWindow is an example of an API that has been wrapped nicely by the .NET CLR, it is unlikely that you will ever need to import and call it.  But something more useful might be: Deleting a file to the recycle bin.

What resources are out there for PInvoking?

The best resource I have found is PInvoke.net, which has declarations for both VB.NET and C# for commonly used Win32 API.

Often one needs to make a call to a webserver in .NET, and just as often there is Querystring data that needs to be sent along with this request. Thanksfully .NET has a great WebClient class in System.net.WebClient.  But you can’t just add the querystring data into your URL like such:

Wrong:
WebClient.DownloadFile("http://www.daveamenta.com/file.aspx?mydata=this is a string that I need to send to the server?");

The data needs to be encoded so that the server will not misinterpret the data. This is an example of using an encoded string:

Right:
WebClient.DownloadFile("http://www.daveamenta.com/file.aspx?mydata=this+is+a+string+that+I+need+to+send+to+the+server%3F");

The second string is encoded properly, this way the webserver will properly parse that last ?. This was done using the URLEncode function, .NET has built-in classes for this–but they are a bit hidden. They’re kept in the System.Web namespace. This means that if you are working with .NET on the desktop side, as opposed to ASP.NET, you’ll need to Add a Reference to System.Web.dll. Once you do, you’ll see new classes in System.Web that you can use for many things.

Using URLEncode in C#:

WebClient.DownloadFile( "http://www.daveamenta.com/file.aspx?mydata=" +
System.Web.HttpUtility.UrlEncode( "this is a string that I need to send to the server?" ) );

It is unlikely you will have much use for URLDecode in a C# application, but it is applied the same way.

What exactly is URLEncode?

URLEncode returns a string in which all non-alphanumeric characters other than - (hyphen) _ (underscore), and . (period) have been replaced with a percent (%) sign followed by two hex digits that correspond to the charactor code and spaces encoded as plus (+) signs. Spaces can alternatively be encoded as %20.  20 is (16*2)+(1*0)=32 in decimal, which is the ASCII code for space.

URLDecode will reverse this operation.

I’ve seen the (perhaps VB-style) way of getting the MyDocuments folder, by using %HOMEDIR%, a batch-style variable.  Some of these environment variables can be modified by navigating to Control Panel -> System -> Advanced (tab).  In Vista, you’ll need to select Advanced System Settings at the System page in order to invoke the System Properties dialog.

These environment varibles work as pathes, but they are not really the accepted way of doing things.  What you should do, is use System.Environment.GetFolderPath().  This method takes a value from the System.Environment.SpecialFolder enum.  GetFolderPath will return a string for your location.  This is supported on all versionf of the .NET framework.

Below is an excerpt from MSDN of the possible values for SpecialFolder:

I had a friend ask me yesterday what the easiest/best way to delete a file to the recycle bin was, in C# that is.  I knew about SHFileOperation, but I didn’t know if .NET 2.0 had native support for sending files to the recycle bin.  As it turns out, there are two ways in which you can move files to the recycle bin.  Microsoft has included the support in the Microsoft.VisualBasic namespace, and referencing the Microsoft.VisualBasic assembly will allow you to use that function directly.

Deleting a file using Managed code in C# or VB.NET:

(Remember to add a reference to Microsoft.VisualBasic)

using Microsoft.VisualBasic;
 
string path = @"c:\myfile.txt";
FileIO.FileSystem.DeleteDirectory(path, FileIO.UIOption.OnlyErrorDialogs, .RecycleOption.SendToRecycleBin);

The other alternative is to use COM Interop and P/Invoke the SHFileOperation from the Windows shell32.dll.

Deleting a file using P/Invoke in C#:

I have written a class that provides methods to easily specify how the file should be sent to the recycle bin.  SHFileOperation can take flags, which I have put into an enum and commented.  These flags define how the operation should be executed.

• FOF_SILENT - Do not show the “Recycling…” dialog.
• FOF_NOCONFIRMATION - Do not prompt the use for confirmation of the operation.
• FOF_ALLOWUNDO - This is the option that specifies that the file should be sent to the recycle bin.  It is added by default in each Send method.
• FOF_SIMPLEPROGRESS - This option will surpress the display of the names of the files that are being deleted.
• FOF_NOERRORUI - This option will suppress any errors encountered during the operation.
• FOF_WANTNUKEWARNING - Adding this flag will allow a dialog to allow the user to cancel removing large files permanently, if they are too large for the recycle bin.

Any one of these flags can be added to change the operation.

Methods provided by the RecycleBin class:

• bool Send(string path)
• bool Send(string path, FileOperationFlags flags)
• static bool SendSilent(string path)

The first will simply send a path to the bin while showing the dialog, but surpressing the confirmation.

The second will allow you to customize the flags for the file operation, the only flag set for you is FOF_ALLOWUNDO. Flags can be combined with the | operator. e.g. FOF_NOERRORUI | FOF_SIMPLEPROGRESS.

The third, and probably most commonly used, will surpress all errors and warnings, and not show the user any operation is in progress.

Code Sample:

SendToRecycleBin.zip [ZIP] [Visual Studio 2008] (43KB)

I’ve spent a bit of time tweaking this XML Serialization class I wrote to make use of the System.Xml.Serialization objects Microsoft provides in .NET 2.0.

XML Serialization is a great way to store complex data types, and also a great alternative to binary serialization. Unfortunately, using these classes is not as straightforward as I would like, there isn’t any simple Save or Load methods, I’ve bridged that gap with a a small class that provides this functionality.

Basic Methods:

These are the two basic methods you would need to save and load an object from a file.

• public static T Load<T>(string filename)
• public static void Save<T>(string filename, T cls)

Additional Methods:

These methods are included for those that would like to bypass the files, and send objects directly over a different medium.

• public static String SerializeObject<T>(T pObject)
• public static T DeserializeObject<T>(String pXmlizedString)

Important Note:

Remember to wrap all of these methods in try-catch blocks and catch any exceptions thrown. If the file cannot be accessed, or there is an error during serialization, it will be passed on.

Code Example:

Load an object from file:

List MyList = null;
try
{
MyList = XSerial.Load<list>("MyList.xml");
}
catch (Exception e)
{
Console.WriteLine("Error: " + e.Message);
// make sure the object won't cause errors.
MyList = new List();
}
 
Save an object to file:
 
try
{
XSerial.Save</list><list>("MyList.xml", MyList);
}
catch (Exception e)
{
Console.WriteLine("Error Saving: " + e.Message);
// nothing can be done about recovery.
}</list>

Download Demo Project:

XMLSerialTest [ZIP] [Visual Studio 2008]

Twitter is great, and it’s nice that there is more than one way to send a tweet (message) out, you can use Web, SMS or Google Talk (or even Digsby!). But what if you’re a heavy mIRC user, or if you just like to have more ways to tweet? Clint, on DALnet/#system was working on this, and I’ve added a few things that I think are worth sharing.

How can I use twitter through mIRC?

This would be entered into the Remote script section in mIRC:

alias tweet {
; use /tweet
set %tw.username twitter_username_here
set %tw.password twitter_password_here
; exit if the message is too long for twitter to take.
if ($len($1-) > 140) {
echo -a Sorry, that was $calc($len($1-)-140) characters too long!
halt
}
; connect to twitter on HTTP port 80
set %authentication $encode($+(%tw.username,:,%tw.password),m)
sockclose twitter
sockopen twitter twitter.com 80
; set a 2-second timeout in case there is a problem
; this could be changed to 5 or 10
.timertwitter 1 2 twitter_timeout
set %tweet $$1-
; store the text
}
 
; encoding is important for sending data to a web server
; the decode function is not used, but is here for the sake of
; completeness
alias urlencode return $regsubex($1-,/\G(.)/g,$iif(($prop && \1 !isalnum) || !$prop,$chr(37) $+ $base($asc(\1),10,16),\1))
alias urldecode return $replace($regsubex($1-,/%(\w\w)/g,$chr($iif($base(\t,16,10) != 32,$v1,1))),$chr(1),$chr(32))
 
; when the socket accepts our connection
on *:sockopen:twitter:{
sockwrite -n twitter POST /statuses/update.json HTTP/1.1
sockwrite -n twitter Host: twitter.com
sockwrite -n twitter User-Agent: Mozilla/5.0 (Windows; U; Windows NT 6.0; en-US; rv:1.9b5) Gecko/2008032620 Firefox/3.0b5
sockwrite -n twitter Content-Length: $calc($len($urlencode(%tweet)) + 9)
sockwrite -n twitter Authorization: Basic %authentication
sockwrite -n twitter $crlf
sockwrite twitter status=
sockwrite -n twitter $urlencode(%tweet)
sockwrite twitter $crlf
sockwrite twitter $crlf
; send the postdata
}
 
; the data wasn't returned, kill the socket
; and tell the user.
alias twitter_timeout {
echo -a Message Failed to Send - Socket Timeout
sockclose twitter
}
 
; we have data back
on *:sockread:twitter: {
.timertwitter off
sockread -f %string
echo -a Message Returned ( $+ %string $+ )
sockclose twitter
}

Hopefully this is useful to someone else, I often find myself updating twitter this way, because I always have mIRC open, albeit in the background.

GlobalCommand is a utility for sending commands to the computer via hotkeys (new), or a sequence of keystrokes.  It’s been more than 3 years since GlobalCommand 1.0, and I’ve nearly finished up version 3.

For example: Typing $w can send ‘Now Playing: Paula Deanda - Walk Away’ to the program you typed it into. $p could also just tell iTunes to pause.  You could also replace commonly misspelled words. “teh” can be mapped to “the.”

GlobalCommand is, as the name implies, program independent.  GlobalCommand works in any application that uses keyboard input on Windows*.  GlobalCommand can also map hot keys, Windows Flag Key + A could tell iTunes to start playing, or could start Microsoft Word.  The possibilities are endless, and GlobalCommand is meant to all a power user the ability to easily take control of their computer by providing a means for easy hot key registration and management.

What’s new in GlobalCommand 3?

• GlobalCommand 3 is a complete rewrite, from the ground up.  GC3 is written in C#, and leverages the power of the Microsoft .NET Framework v3.5 in order to provide an easy to use application which took minimal time to develop and test.
• GlobalCommand 3 implements an interface for plug-in applications to take advantage of the easy to use GC framework.  Each output or control module is a plug-in.  This provides maximum flexibility and allows for easy change or modification of components.
• GlobalCommand 3 will store all data in the Application Data folder for windows.  Commands will be kept separate on a per-user basis.  Logging data is also stored in the Application Data folder.
• Supported Operating Systems: Windows XP, Windows Vista.  Support for Windows 9X/ME and 2000 has been removed.  Microsoft .NET v3.5 is not supported by those operating systems, and thus GC is not compatible with computers that lack the proper framework.

What components come with GlobalCommand?

GC Ships with 4 types of components:

1. The GC Plugin Framework: GCPluginFramework.dll.  This library is fully documented, and provides the interface for building plug-ins.
2. The GC binary.  GlobalCommand.exe  This file should not be renamed (for purposes of updating components)
3. Packaged Plug-ins for GlobalCommand.
4. The PluginDebugger.  Both Compiled binary (all you actually need), and source solution to the plugin debugger are included in the download.  This project will alllow you to easily test plug-ins, as well as verify information and security keys.

What plug-ins come with GlobalCommand?

• Updater - This plugin allows you to select a repository and update GlobalCommand, as well as its’ plug-ins and dependencies.  You can also download new items in the repository through this facility.
• iTunes - get song information and set commands to the player.
• Language Translation - translate a string of text into several other languages.
• mIRC - simple output plugin (supports perview) for sending color codes to mIRC.
• Shell - Start applications and run commands.
• Winamp - get the current song from Winamp.
• Keys - common keystrokes that can’t be represented by printable text.
• Sample - this plugin (source attached) is an example that can be used to develop more powerful plug-ins.

Can I have the source code?

GlobalCommand is not open source at this time.

What does it look like?

Screenshots:

Where can I get it?

Download: GlobalCommand 3 Beta 1 [ZIP] (1.11MB)

* GlobalCommand may not work well with all Video Games.

Today was the first annual Imagine RIT Festival, hosted at the RIT campus in Rochester, New York.  RIT was expecting a maximum of 30,000 people to attend, and although the weather wasn’t favorable for the event, drizzling without much sun, more than 10,000 people came to the campus.  There were more than 400 exhibits around the campus–not only in the academic buildings–but also in the quads and Field House.  The Xerox ‘WOW’ center was setup much the same way our career fairs are setup.  Some of the memorable exhibits included RIT EyeTracking, and Calling Earth to Second Life.  I spent a few hours walking around the campus checking everything out, there was some seriously cool stuff.  Digsby was also present, and I scored a t-shirt from Steve.

The purpose of Imagine RIT, aside from kicking off the Rochester Festival season is to promote the university as not only a steller teaching university, but also as one that is is an innovation university, with an emphasis on research and creativity.  RIT is ranked highly for undergraduate education, and having one of the oldest co-op programs has earned it a reputation which attracts many of the best employers in the USA.