Friday, March 17, 2006

I love anonymous delegates - I think they are extremely useful and allow me to solve some problems in very elegant ways.  However, once you really get into them, you start to see the dark side of anonymous delegates and that is unregistration.

Here's the basic problem: when binding an instance delegate to an event to handle some activity, the delegate will cache off the instance reference - thereby keeping the reference alive.  So for example, if I had a form which wanted to process some activity from an object:

class Publisher
{
   public event EventHandler OnEvent;
   ...
}

class MainForm : Form
{
   Publisher _pub = new Publisher();

   void ActivateChildForm() 
   {
      ChildForm f = new ChildForm(_pub);
      f.Show();
   }
}

class ChildForm : Form
{
   public ChildForm(Publisher pub) 
   {
      pub.OnEvent += ProcessEvent;
   }
   void ProcessEvent(object sender, EventArgs e)
   {
      listBox1.Items.Add("Event was fired!");      
   }
}

When the ChildForm instance is closed, the form won't be collected because the Publisher (_pub) is holding a reference to it.  This is easily fixed by adding some code into the FormClosing event:

class ChildForm : Form
{
   Publisher _pub;

   public ChildForm(Publisher pub)
   {
      _pub = pub;
      _pub.OnEvent += ProcessEvent;
   }

   void FormClosing(object sender, FormClosingEventArgs e)
   {
      _pub.OnEvent -= ProcessEvent; 
   }
}

Now, the form will be cleaned up when it's closed.  This is pretty standard stuff, and most people that have been using .NET for a while know all this.  Here's the rub: with .NET 2.0, we can simplify the code using anonymous delegates which are really helpful for these single-line processing event handling functions.  So, I could recode my handler as:

class ChildForm : Form
{
   public ChildForm(Publisher pub) 
   {
      pub.OnEvent += delegate { listBox1.Items.Add("Event was fired!"); }
      this.OnClosing += delegate { pub.OnEvent -= ????? }

   }
}

The issue is that I don't have a reference to the delegate as it's typed.  Under the covers, the C# compiler has generated a temporary function (or possibly even pulled it out to a separate inner class) and there's no way for me to get to the underlying function.  So, what can I do?  Well, the easiest thing to do is to save off the function:

class ChildForm : Form
{
   public ChildForm(Publisher pub) 
   {
      EventHandler eh = delegate { listBox1.Items.Add("Event was fired!"); }
       
      pub.OnEvent += eh;      
     
this.OnClosing += delegate { pub.OnEvent -= eh; }
   }
}

Now my code will function properly -- and is significantly reduced in size.  Of course, I've lost the benefit of being able to hook up events through VS.NET because it always generates seperate functions and I would need to cache off the Publisher instance as well as my delegate in that case. 

So, rule #1, always unregister the event when you are finished.  Rule #2, remember that anonymous delegates may be keep your instance alive so unregister them as well, unless the event is to be hooked up throughout the lifetime of the application.

posted on 3/17/2006 10:56:40 AM (Central Standard Time, UTC-06:00)  #   

I've posted a new version of ATAPI.NET which supports consultation transfers and a simple phone sample that shows off how to use the features.  In addtiion, this version of ATAPI.NET has a couple of bug fixes that were rolled in from a production project over the past couple of weeks to fix some weird startup/shutdown issues when a LINE_REINIT is reported by TAPI.  You can get the new code from our samples link -- http://www.julmar.com/samples.htm

 

posted on 3/17/2006 9:59:23 AM (Central Standard Time, UTC-06:00)  #   
 Wednesday, March 15, 2006

Forwarding lines with ATAPI.NET is simple and easy (assuming, of course, that the underlying TSP supports it).

The first step is knowing whether a given line device even supports forwarding.  This is trivial:

TapiManager mgr = new TapiManager("ForwardingTest");

foreach (TapiLine line in mgr.Lines)
{
   
if
(line.Capabilities.SupportsForwarding)
   {
      Console.WriteLine("Line {0} supports forwarding!"
, line.Name);
   }
}

Once we've identified a specific line, we can look at each address and get more information such as the types of forwarding supported.  For example, we might be able to forward to different numbers based on specific conditions such as whether the call goes unanswered vs. whether the address is in use and returning a busy signal.  We might also be able to forward specific inbound callers (very useful to get rid of your bosses calls).  We can get this information from the Capabilities of the TapiAddress object:

foreach (TapiAddress addr in line.Addresses)
{
   
Console.WriteLine("Forwarding modes supported on {0} are {1}"
, addr.Address, addr.Capabilities.SupportedForwardingModes);
}

We can also retrieve any existing forwarding information through the Status of the TapiAddress:

foreach (ForwardInfo fwd in addr.Status.ForwardingInformatioin)
   
Console.WriteLine("\t{0} to {1}:{2}"
, fwd.ForwardMode, fwd.DestinationAddressType, fwd.DestinationAddress);

This outputs: "Unconditional to PhoneNumber:1234" on a forwarded line I setup.

Finally, the big question is how to change the forwarding information, this is pretty easy as well.  You can set forwarding information on two levels, the entire line (which impacts all addresses), or a specific address.  This is done through two methods present on both TapiAddress and TapiLine which are Forward and CancelForward.  So, to cancel all forwarding in effect on every line we could do the following:

Console.WriteLine("Canceling all forwards:");
foreach (TapiLine line in
mgr.Lines)
{
   
if
(line.Capabilities.SupportsForwarding)
   {
      
try
      
{
         line.CancelForward();
      }
      
catch (TapiException
ex)
      {
         
Console.WriteLine("{0} - {1}"
, line, ex.Message);
      }
   }
}

Or, to setup the forwarding as above, I can issue a call to the Forward method:

ForwardInfo[] fwdInfo = new ForwardInfo[] {
      
new ForwardInfo(ForwardingMode.Unconditional, 0, "1234"
)
};

foreach (TapiLine line in mgr.Lines)
{
   
if
(line.Capabilities.SupportsForwarding)
   {
      
try
      
{
         line.Forward(fwdInfo, 5,
null
);
      }
      
catch (TapiException
ex)
      {
         
Console.WriteLine("{0} - {1}"
, line.Name, ex.Message);
      }
   }
}

The ForwardInfo class describes a single forwarding instruction and you pass an array of these info the Forward method to indicate how things are to be managed.  Exceptions need to be handled because the TAPI service provider might not allow the particular forwarding at this point in time, or the destination might not be allowed, etc.

Under the covers this will issue a lineForward request with a LINEFORWARDLIST setup for each of the ForwardInfo structures.

That about covers it!  Ping me with any questions if you want.

 

posted on 3/15/2006 2:43:58 PM (Central Standard Time, UTC-06:00)  #   
 Monday, March 13, 2006

I got hit with this problem from two separate clients last week - a .NET 1.1 application ported to .NET 2.0 is now terminating abruptly for no apparent reason.  Well, of course there's a reason - and it's that both applications had "hidden" exceptions being thrown in some background thread that weren't being caught.  Under .NET 1.1, the CLR would print any exceptions that occurred on threadpool threads to the console and then return the thread to the pool.  In addition, the CLR would silently eat any exception thrown on the finalizer thread (again printing the stack trace to the console).  Under .NET 2.0, the behavior has changed and it will cause the AppDomain to be unloaded (yikes!).  So, for example:

class BadClass
{
  ~BadClass()
  {
    throw new Exception("I'm Bad"); 
  }
}

class BadClass
{
  static void Main()
  {
     for (; ;)
       
new BadClass();
  }
}

The above code would run forever under CLR 1.1, but will terminate immediately when run under CLR 2.0 with an unhandled exception. 

Unhandled Exception: System.Exception: I'm bad
   at BadClass.Finalize() in W:\Projects\TestApp\TestApp\Program.cs:line 11

There are a couple of ways to deal with this - the best is to do all your testing and close the holes.  You really shouldn't have unhandled exceptions loose in your program.  But for a really large program, or one where you don't know where the exception is happening, Microsoft has given you a couple of options.  The first one is the <legacyUnhandledExceptionPolicy>  tag which you can put into your app.config file:

<configuration
  <
runtime

    <
legacyUnhandledExceptionPolicy enabled="1"
/> 
  <runtime
>
<configuration>

This will essentially revert the CLR to the 1.1 behavior.  Consider it a short-term fix because it will ultimately mask issues with your application that are really bugs.
 
Along with the above, you can also be notified about unhandled exceptions prior to the AppDomain being unloaded.  This is done through several different methods based on the type of application.  The first basic method is through the AppDomain.UnhandledException event which is raised for Console applications:

static
void Main()
{
   
AppDomain.CurrentDomain.UnhandledException += delegate(object sender, UnhandledExceptionEventArgs
e)  
   { 
      
Console.WriteLine("{0} - IsTerminating = {1}"
, e.ExceptionObject.ToString(), e.IsTerminating); 
   
};

   for (; ;)
      
new BadClass
();
}

In this case, the application will still be terminated, but we will now get notified right before termination.  This allows our application to log the error (using the nifty System.Diagnostics.TraceSource support or through the EventLog) so we at least know what happened.

For Windows Forms applications, things are a little different -- instead of the AppDomain.UnhandledException event being raised, the Windows Forms infrastructure will raise the Application.ThreadException event for exceptions that occur on the primary thread.   The default behavior for this handler is to display the friendly System.Windows.Forms.ThreadException dialog:

If the exception occurs in the primary (message pumping) thread, the user will see the above dialog box and be given the choice to terminate the application or not.  The code behind it looks something like:

void Application_ThreadException(object sender, ThreadExceptionEventArgs e)
{
   
try
   
{
      
// Call user override
      
if (this.threadExceptionHandler != null
)
      {
         
this.threadExceptionHandler(Thread.CurrentThread, new ThreadExceptionEventArgs
(e.Exception));
      }
      
else
      
{
         
using (ThreadExceptionDialog excptDlg = new ThreadExceptionDialog
(e.Exception))
         {
            
DialogResult
result = excptDlg.ShowDialog();
            
if (result == DialogResult
.Abort)
               
Application
.Exit();
         }
      }
   }
   
catch
   
{
   }
}

That's great for the main thread, but if the exception occurs in a secondary thread, then the application will still be terminated after raising the AppDomain.UnhandledException event.  So, for non-UI threads, you must still register the unhandled exception handler on the AppDomain, log the failure and then watch your application die.  Having two handlers can be a pain and if you want to have the application terminate on any unhandled exception, you can direct Windows Forms to not catch primary thread exceptions automatically by using the Application.SetUnhandledExceptionBehavior method:

[STAThread]
static void
Main()
{
   Application.SetUnhandledExceptionMode(UnhandledExceptionMode.ThrowException);
   Application.Run(new Form1());
}

The above line will cause exceptions thrown in the main thread to be unhandled - thereby triggering the AppDomain.UnhandledException event.

For ASP.NET applications, things are also a bit different.  First, be aware that any exception thrown during a Page request (the normal Page rendering process) will be handled automatically by ASP.NET and rendered back to the client based on the error settings in web.config.  If I throw an exception while rendering the page, then I'll get an HTML response like:

Server Error in '/TestWebSite' Application.
Bye, Bye!

Description: An unhandled exception occurred during the execution of the current web request. Please review the stack trace for more information about the error and where it originated in the code.

The important thing here is that the AppDomain continues to run.  I can tell ASP.NET to redirect to a specific page through the <customErrors> section of the web.config.  Or, I can catch these exceptions and handle them myself through the global.asax Application_Error method (which is hooking the HttpApplication.Error event):

void Application_Error(object sender, EventArgs e)
{
   Exception
ex = Server.GetLastError();
   Context.ClearError();
  // Stop default error reporting
   
Response.Write(
"Application Error:");
   Response.Write(
"" + Server.HtmlEncode(ex.ToString()) + ""
);
}

A more popular way to display errors is to cache off the last error in a session variable and then execute Server.Transfer to some error.aspx page:

void Application_Error(object sender, EventArgs e)
{
   Exception
ex = Server.GetLastError();
   Session["LastException"] = ex;

   
Server.Transfer("myerrorpage.aspx");

}

Unfortunately, just like Windows Forms, if an unhandled exception occurs on a separate thread (such as a Threadpool thread), or outside the request processing framework, then the AppDomain is terminated.  This has significant consequences for ASP.NET applications - because the website itself is cycled.  This means you have lost session state and your users are going to notice the website outage.  Most of the time, there isn't much information to work off of -- you end up with a cryptic Event Log entry like:

EventType clr20r3, P1 w3wp.exe, P2 6.0.3790.1830, P3 4333d6f1, P4 app_web_5hu0gabf, P5 0.0.0.0, P6 4415a8be, P7 7, P8 a, P9 system.exception, P10 NIL.

In order to get more information, we need to hook the AppDomain.UnhandledException event and then log the information, just like we have for Windows Forms and Console applications.  We could put a handler into our global.asax, but that would only handle that one web application.  It would be better to create a new HTTP Module with the event handler and hook that up through the web.config.  Luckily for me, Microsoft has already published a KB article which shows off this exact solution - http://support.microsoft.com/?id=911816.  In this article, Microsoft details creating a new HttpModule class which will log the exception information to the event log so you will know exactly where the exception occurred and can provide your own exception handling.  Unfortunately, without the above <legacyUnhandledExceptionPolicy> entry, the website will still be cycled.

Bottom line is: be aware of this new, breaking change -- look for places where you might not be handling exceptions properly and implement good audit trail mechanisms to log failures when/if they occur.  This will allow you to find and fix issues before they become production problems as you port your legacy code over to .NET 2.0!

posted on 3/13/2006 12:07:55 PM (Central Standard Time, UTC-06:00)  #   
 Tuesday, March 07, 2006

So, a question was asked "How do I determine what's happening in the TAPI3 wrapper"?  The answer is you turn on the internal trace source -- ITapi3 was built with a build in tracing facility to tell you when it had any underlying interface or COM failures and it's easy to activate.  First, add an Application Configuration File to your project.  Open that file and add the following lines:

<?xml version="1.0" encoding="utf-8" ?>
<
configuration>
   <
system.diagnostics>
      <
sources>
         <
source name="ITapiTrace" switchName="tapiSwitch" switchType="System.Diagnostics.SourceSwitch">
            <
listeners>
               
<add name="MyTraceLog" type="System.Diagnostics.TextWriterTraceListener" initializeData="MyTrace.txt" />
            </
listeners>
         </
source>
      </
sources>
      <
switches>
         <
add name="tapiSwitch" value="All" />
      </
switches>
   </
system.diagnostics>
</
configuration>

This will create a file called "MyTrace.txt" in your working directory.  The important line is the source tag which identifies the internal TraceSource object used by the ITapi3 library.  Inside this file will be the internal TAPI3 calls being made for your application.  As an example, the following trace shows me that several underlying COM errors occurred in the running of a simple TAPI3 application -- it was unable to retrieve the ITTerminal interface from the ITAddressEvent interface (which actually isn't really an error), failed to open the line (because Unimodem won't allow the media type VOICE to be passed for my modem), and failed to set the play list for this MSP -- [0x80040216] is actually a DirectShow error [VFW_E_NOT_FOUND].

ITapiTrace Verbose: 0 : Creating ITTAPI instance
ITapiTrace Verbose: 0 : Hooking up connection sink to ITTAPI interface
ITapiTrace Information: 0 : ITTapi::put_EventFilter(0x8001F) hr=0x0
ITapiTrace Error: 0 : COM Hresult 0x80040004 "The MEDIATYPE passed in to this method was invalid." generated 
   at JulMar.Tapi3.TapiException.ThrowExceptionForHR(Int32 hr)
   at JulMar.Tapi3.TTapi.RegisterCallNotifications(ITAddress* pitf, Int16 vbMonitor, Int16 vbOwner, Int32 supportedMediaTypes)
   at JulMar.Tapi3.TAddress.Open(TAPIMEDIATYPES supportedMediaTypes)
ITapiTrace Error: 0 : ITAddressEvent::get_Terminal failed hr=0x80040055
ITapiTrace Verbose: 0 : Processing TapiAddressChangedEventArgs: Evt=AE_RINGING, Address=DSSP Line #1 - Address 5, Terminal=
ITapiTrace Error: 0 : ITAddressEvent::get_Terminal failed hr=0x80040055
ITapiTrace Verbose: 0 : Processing TapiAddressChangedEventArgs: Evt=AE_RINGING, Address=DSSP Line #1 - Address 4, Terminal=
ITapiTrace Error: 0 : ITAddressEvent::get_Terminal failed hr=0x80040055
ITapiTrace Verbose: 0 : Processing TapiAddressChangedEventArgs: Evt=AE_RINGING, Address=DSSP Line #1 - Address 3, Terminal=
ITapiTrace Error: 0 : ITAddressEvent::get_Terminal failed hr=0x80040055
ITapiTrace Verbose: 0 : Processing TapiAddressChangedEventArgs: Evt=AE_RINGING, Address=DSSP Line #1 - Address 2, Terminal=
ITapiTrace Error: 0 : ITAddressEvent::get_Terminal failed hr=0x80040055
ITapiTrace Verbose: 0 : Processing TapiAddressChangedEventArgs: Evt=AE_RINGING, Address=DSSP Line #1 - Address 1, Terminal=
ITapiTrace Error: 0 : ITAddressEvent::get_Terminal failed hr=0x80040055
ITapiTrace Verbose: 0 : Processing TapiAddressChangedEventArgs: Evt=AE_RINGING, Address=DSSP Line #1 - Address 0, Terminal=
ITapiTrace Verbose: 0 : Processing TapiCallNotificationEventArgs: Event=CNE_OWNER, Call=TCall: 171360625 CS_OFFERING
ITapiTrace Verbose: 0 : Processing TapiCallStateEventArgs: Call=TCall: 171360625 CS_OFFERING, State=CS_OFFERING, Cause=CEC_NONE
ITapiTrace Error: 0 : COM Hresult 0x80040216 "" generated 
   at JulMar.Tapi3.TapiException.ThrowExceptionForHR(Int32 hr)
   at JulMar.Tapi3.TTerminal.set_MediaPlayList(String[] fileList)
   at AnsMachine.AutoAttendantForm.AnswerCall()
   at AnsMachine.AutoAttendantForm.OnCallState(Object sender, TapiCallStateEventArgs e)
ITapiTrace Verbose: 0 : Processing TapiCallStateEventArgs: Call=TCall: 171360625 CS_DISCONNECTED, State=CS_DISCONNECTED, Cause=CEC_DISCONNECT_NORMAL
ITapiTrace Error: 0 : ITAddressEvent::get_Terminal failed hr=0x80040055
ITapiTrace Verbose: 0 : Processing TapiAddressChangedEventArgs: Evt=AE_RINGING, Address=DSSP Line #1 - Address 5, Terminal=
ITapiTrace Error: 0 : ITAddressEvent::get_Terminal failed hr=0x80040055
ITapiTrace Verbose: 0 : Processing TapiAddressChangedEventArgs: Evt=AE_RINGING, Address=DSSP Line #1 - Address 4, Terminal=
ITapiTrace Error: 0 : ITAddressEvent::get_Terminal failed hr=0x80040055
ITapiTrace Verbose: 0 : Processing TapiAddressChangedEventArgs: Evt=AE_RINGING, Address=DSSP Line #1 - Address 3, Terminal=
ITapiTrace Error: 0 : ITAddressEvent::get_Terminal failed hr=0x80040055
ITapiTrace Verbose: 0 : Processing TapiAddressChangedEventArgs: Evt=AE_RINGING, Address=DSSP Line #1 - Address 2, Terminal=
ITapiTrace Error: 0 : ITAddressEvent::get_Terminal failed hr=0x80040055
ITapiTrace Verbose: 0 : Processing TapiAddressChangedEventArgs: Evt=AE_RINGING, Address=DSSP Line #1 - Address 1, Terminal=
ITapiTrace Error: 0 : ITAddressEvent::get_Terminal failed hr=0x80040055
ITapiTrace Verbose: 0 : Processing TapiAddressChangedEventArgs: Evt=AE_RINGING, Address=DSSP Line #1 - Address 0, Terminal=
ITapiTrace Error: 0 : ITTapi::Shutdown hr=0x0

 

posted on 3/7/2006 5:18:50 PM (Central Standard Time, UTC-06:00)  #   

Matthias Moetje (Tapi MVP) has graciously taken one of the VB6 Platform SDK samples and ported it to the ITapi3 library and allowed me to distribute it in the Tapi3 Wrapper sample set.  It allows you to select an address, place a call on it and then monitor or generate digits.  Great sample - thanks Matthias!

posted on 3/7/2006 5:04:47 PM (Central Standard Time, UTC-06:00)  #   
 Monday, March 06, 2006

I got an email question today asking about how to write a .WAV file out to an active call using our TAPI3 wrapper.  It's actually pretty easy to do and it follows along with the normal SDK method used in C++.  Here's the relevant code (for the full example, download the http://www.julmar.com/samples/tapi3wrapper.zip and look at the AutoAttendant sample).

First, when a new offering call shows up, get hold of the file playback terminal for it and set it up to play your .WAV files.  We cannot play the files until the call is answered, but this will essentially "queue" it up.

// Method called when TE_CALLSTATE == OFFERING raised
private void AnswerCall()
{
   // Get the playback terminal from the call
   try
   {
      playbackTerminal = activeCall.RequestTerminal(TTerminal.FilePlaybackTerminal, 
                  TAPIMEDIATYPES.AUDIO, TERMINAL_DIRECTION.TD_CAPTURE);
      if (playbackTerminal != null)
      {
         playbackTerminal.MediaPlayList = new string[] { "Hello.wav" };
         activeCall.SelectTerminalOnCall(playbackTerminal);
         activeCall.Answer();
      }
      else
      {
         MessageBox.Show("Failed to retrieve playback terminal.");
         activeCall.Disconnect(DISCONNECT_CODE.DC_REJECTED);
      }
   }
   catch (TapiException ex)
   {
   }
}

Next, watch for the call media to change indicating we have an active stream for our terminal.  When that happens, start the playback stream:

// Method called when TE_CALLMEDIA is raised
private void OnCallMedia(object sender, TapiCallMediaEventArgs e)
{
   try 
   {
      if (activeCall != null && e.Event == CALL_MEDIA_EVENT.CME_STREAM_ACTIVE &&
            e.Terminal.Direction == TERMINAL_DIRECTION.TD_CAPTURE &&
            playbackTerminal != null)
      {
         playbackTerminal.Start();
         SetStatusMessage("File Playback Terminal started ");
      }
   }
   catch (TapiException ex)
   {
   }
}

Finally, when the file terminal is finished, close and dispose the stream.  This is done just to cleanup the resources properly:

// Method called when TE_FILETERMINAL is raised
private
void OnFileTerminal(object sender, TapiFileTerminalEventArgs e)
{
   // We are interested in TMS_IDLE because we will un-select playback and 
   // select recording
   if (e.State == TERMINAL_MEDIA_STATE.TMS_IDLE)
   {
      if (e.Terminal.Direction == TERMINAL_DIRECTION.TD_CAPTURE && playbackTerminal != null)
      {
         try
         {
            // Remove the playback terminal
            activeCall.UnselectTerminalOnCall(playbackTerminal);
            playbackTerminal.Dispose();
            playbackTerminal = null;
         }
         catch (TapiException ex)
         {
         }
      }
   }
}

That's pretty much it -- there's a full sample of this that should work with voice modems or any streaming-capable TSP. 

Updated: 3/16/05 -- It doesn't appear to completely work with the H.323 provider; the file terminal gets connected but apparently never reports an IDLE state and so never gets disconnected in the above sample.  This appears to be true of the Platform SDK samples as well.

posted on 3/6/2006 5:02:34 PM (Central Standard Time, UTC-06:00)  #   
 Friday, March 03, 2006

New announcement - TAPI3 wrapper for .NET available on our samples page -- http://www.julmar.com/samples/tapi3wrapper.zip

Recently, I released the ATAPI.NET project into the wild for people to be able to easily code up .NET applications that utilize TAPI.  That assembly is based on the C API exported from TAPI 2.1 and so uses the older form of TAPI programming.  With Windows 2000, Microsoft released a COM version of TAPI - dubbed TAPI 3.0 that was designed to allow VB developers to access the telephony API.  However, as I noted in my previous post, Microsoft claims that the object model is too complex and that there are significant issues with accessing it from managed code.

.NET has a built-in facility to access COM objects, it creates .NET wrappers (called RCW's) around the COM interfaces and then allows you to call the COM object as if it were a .NET object.  That's great until you need to mesh the COM style of cleanup (AddRef/Release) with the .NET style (GC).  Essentially, the COM object doesn't get released until the managed wrapper gets collected.  In addition, if you have multiple wrappers for a single interface (TAPI returns the same interface through different methods), the interface won't be released until all the wrappers are collected.

Sometimes, this isn't really a problem - so we keep a COM interface alive a little longer than normal.  In TAPI's case however it can be a huge problem because some TAPI service providers will not let you create new calls unless the existing call interface has been released - even if it isn't connected to anything.

So, two options exist today:

1) Don't use TAPI 3.x - use TAPI 2.x where you have more control over the underlying TAPI call handle and can call lineDeallocate yourself.  In .NET this means using some wrapper like ATAPI.NET.  If you don't need terminal and stream support, this is fine.  This is what the newsgroups typically will recommend to people.

2) Call Marshal.ReleaseCOMObject on every outstanding interface.  Just plain ugly and extremely error-prone.

Now, a new option exists.  I've been in conversations with Matthias Moetje - one of the TAPI MVPs and went back to my original TAPI3 wrapper which was written in MC++ (yuck) and ported it to C++/CLI (which, BTW, is very cool).  It has almost everything supported except agents and call center support.  It solves the above problem in a couple of steps:

1) It guarantees that the same interface will match 1:1 with a managed object as long as the object hasn't been collected yet.  This means the object will only have one holding managed object.
2) It implements IDisposable on most of the wrappers allowing the client to get rid of interfaces immediately.
3) It "auto-disposes" calls when they hit the disconnected state.  This can be turned on or off depending on need through the TTapi.AutoDestroyCalls flag

Instead of a Register function, each address has an Open and Monitor method and a Close method. Since you have to pass the address in anyway, I moved the function to that level and made it distinct rather than having you pass in two booleans to indicate intent.

There are some shorthand functions: for example, the TCall object (which represents ITBasicCallControl, ITCallInfo, et.al.) has a SelectDefaultTerminals method which enumerates the streams and hooks up the default static/video terminals for each stream - basically the code that was in every TAPI sample provided by MS. There's also a FindTerminal method on the TStream class which will locate a terminal of the proper media type/direction if it exists.

The library itself follows the TAPI3 interfaces pretty closely - except it combines the various control interfaces together.  So for example:

Class

Interfaces

TTapi

ITTapi, ITTapi2

TAddress

ITAddress, ITAddress2, ITAddressCapabilities, ITAddressTranslation, ITLegacyAddressMediaControl, ITLegacyAddressMediaControl2, ITMediaSupport, ITTerminalSupport

TCall

ITCallInfo, ITBasicCallControl, ITBasicCallControl2, ITStreamControl, ITLegacyCallMediaControl, ITLegacyCallMediaControl2

In each case, the same properties and methods are exposed by the reference class and it will do the underlying cast necessary to get to the functionality.  The downside of this, is that if the interface isn't supported, the library must throw an exception - in this case a TapiException which holds a string message indicating the failure and an error code (the HRESULT).  The TTapi class exposes the TE_xxx events you generally registered a sink for.

I didn't bother to encapsulate the DirectShow library - instead each of the TAPI wrappers provides a QueryInterface method which you can use to get to an interface if necessary.  So, for example, if you want to get to the venerable IVideoWindow interface from DirectShow, you can include an interop reference to Quartz.DLL and then do the following on the TTerminal object:

IVideoWindow videoWindow = thisTerminal.QueryInterface(typeof(IVideoWindow)) as IVideoWindow;
if (videoWindow != null) { ... }

So, here's an example of using it - I think it's much cleaner that the comparable RCW code, and it's thread-aware and works around the general problem of TAPI3 under .NET through a Dispose mechanism and the TTapi.AutoDestroyCalls flag.

using System;
using System.Collections.Generic;
using System.Text;
using JulMar.Tapi3;

namespace TestTapi
{
   class Program
   {
      static void Main(string[] args)
      {
         TTapi tapi = new TTapi();
         
TCall call = null; TAddress modemAddr = null;

         tapi.Initialize();
         tapi.TE_CALLNOTIFICATION +=
delegate(object sender, TapiCallNotificationEventArgs e)  
         {
            
Console.WriteLine("New call {0} detected from {1}", e.Call.ToString(), e.Event);
         };
         
tapi.TE_CALLSTATE += delegate(object sender, TapiCallStateEventArgs e)
         {
            
Console.WriteLine("{0}:{4} has changed state to {1} due to {2} - current={3}:{5}"
                  
e.Call, e.State, e.Cause, e.Call == call, e.Call.GetHashCode(), (call != null) ? call.GetHashCode() : 0);
            
if (e.State == CALL_STATE.CS_INPROGRESS && e.Call == call)
            {
               
Console.WriteLine("Dropping call");
               e.Call.Disconnect(
DISCONNECT_CODE.DC_NORMAL);
            }
         };

         foreach (TAddress addr in tapi.Addresses)
         {
            
if (String.Compare(addr.ServiceProviderName, "unimdm.tsp", true) == 0 && 
               
addr.QueryMediaType(TAPIMEDIATYPES.AUDIO))
               modemAddr = addr;
         }

         if (modemAddr != null)
         {
            
Console.WriteLine("{0} = {1} ({3}) [{2}]", modemAddr.AddressName, modemAddr.State, modemAddr.ServiceProviderName, modemAddr.DialableAddress);
            modemAddr.Monitor(
TAPIMEDIATYPES.AUDIO);
            
ConsoleKey ki = ConsoleKey.A;

            while (ki != ConsoleKey.Q)
            {
               
// Flip the auto-destroy flag
               
if (ki == ConsoleKey.D) 
               {
                  tapi.AutoDestroyCalls = !tapi.AutoDestroyCalls;
                  
Console.WriteLine("Set AutoDestroy to {0}", tapi.AutoDestroyCalls);
               }
               
// List existing calls
               else
if (ki == ConsoleKey.L)
               {
                  
foreach (TCall _call in modemAddr.EnumerateCalls())
                  {
                     
Console.WriteLine("Existing call found: {0}:{1}", _call, _call.GetHashCode());
                     _call.Dispose();
// Go ahead and dump it
                  
}
               }
               
// Create a new call
               
else
               
{
                  call = modemAddr.CreateCall(
"5551213", LINEADDRESSTYPES.PhoneNumber, TAPIMEDIATYPES.DATAMODEM);
                  
Console.WriteLine("Created new call {0}:{1}", call, call.GetHashCode());
                  
try
                  
{
                     
// This will fail if existing call interface is still around (i.e. not disposed)
                     
call.Connect(false);
                  }
                  
catch (TapiException ex)
                  {
                     
Console.WriteLine(ex.Message);
                  }
               }
               
Console.WriteLine("Press a key to try another call.. Q to quit");
               ki =
Console.ReadKey().Key;
            }
         }

         // This will destroy any outstanding interfaces
         tapi.Shutdown();

         // Call should be disposed here.. state will be CS_UNKNOWN 
         
if (call != null)
            
Console.WriteLine("{0} {1}", call, call.CallState);
      }
   }
}

 

posted on 3/3/2006 12:05:50 PM (Central Standard Time, UTC-06:00)  #   
 Thursday, February 16, 2006

I've really gotten into using code snippets and the compiler expansions in VSNET 2005 -- it cuts my typing way down and I love it.  I've always thought it odd that VB.NET had a ton more snippets pre-installed than C# (my language of choice).  But Microsoft has rectified that with the release of C# snippets on MSDN (http://msdn.microsoft.com/vstudio/downloads/codesnippets/default.aspx) which includes most, if not all the VB.NET ones but recoded for C#.

Now, the hard part will be knowing what's there!

posted on 2/16/2006 7:17:20 PM (Central Standard Time, UTC-06:00)  #   

So, at the GNET2 last week, I found out that you can now call Delegate.Invoke instead of using the C# functor syntax.  This makes code much more readable IMO --

delegate void MyCallback(string msg);
...
MyCallback callback;
...
callback += delegate(string msg) { Console.WriteLine(msg); };
...
callback.Invoke("Hello");   // instead of callback("Hello");

Very nice.

posted on 2/16/2006 8:58:44 AM (Central Standard Time, UTC-06:00)  #