February 2005 - Mike Taulty's Blog

Following up on this previous post about SQL Server 2005 Service Broker , I've been writing a SOAP message transport for the Web Services Enhancements (WSE) 2.0 to move messages using ServiceBroker whilst still gaining advantage of the security, policy, addressing features of WSE2.0. It's important to say that this is purely for fun because it involves installing WSE2.0 on top of the beta .NET Framework 2.0 and there is no way in which the WSE team or the .NET framework team are going to support that kind of scenario. The solution files and source code are all in this zip file here which you can download and play with . I started with the December Community Tech Preview of SQL Server 2005 and the accompanying Visual Studio 2005 build - these are available from MSDN Subscriber Downloads for you to get hold of. Following up on my previous post I've already created constructs in SQL for SoapMessage, SoapContract and also that post details creating some services and queues: SendingService, SendingQueue and ReceivingService...

I saw a piece of code on a screen at DevWeek the other day and a description that went along side it that left me thinking about something that you don't see very often in .NET. The piece of code looked like something like this; public class ThrowingClass { public static void ThrowingMethod() { try { Console.WriteLine("Start statement"); <--- 1 MethodCallThatThrows(); <--- 2 } finally { Console.WriteLine("Finally statement"); <--- 4 Console.ReadLine(); } } private static void MethodCallThatThrows() { throw new ApplicationException("Exception"); <--- 3 } } Here we've got a piece of code which calls MethodCallThatThrows() which is a method which throws an exception and as part of the discussion it was mentioned that the execution is certain to be as indicated with the arrows above. So, when the MethodCallThatThrows method throws its exception we know that the next piece of code to execute is going to be our finally block. However, that's not always going to be the case ...

I've been trying to write some code and struggling to find continuous blocks of time to do it in. However, something that I keep coming across is a scenario where I want to wrap up some code that already has an asynchronous nature and keep the asynchronous nature of it. To try and explain what I'm on about...Imagine that you're writing some code that is largely based around a FileStream . Perhaps we're trying to write some class that offers up 64K of data from a FileStream at a time (yes, I know, it's pointless but my real example's more complicated). If we're trying to do that we might write some code something like; public class MyReader { public static IAsyncResult BeginRead( string path, byte [] array, AsyncCallback callback, object state) { return ( null ); } public static int EndRead(IAsyncResult result) { return (0); } private static void OnReadCompleted(IAsyncResult result) { } } In our BeginRead function we need to use the FileStream to start an asynchronous read and then in our EndRead we need to surface...

The list of "MSDN Nugget" videos that we have published on the UK MSDN website is growing nicely. We've gone from around 4 to around 10 separate items up there right now. It'll continue to grow as there's a pipeline sitting behind it of things waiting to be published. There's two sets of topics up there right now: there's the today stuff ".NET Nuggets" and the "tomorrow" stuff in the "SQL Server 2005 Nuggets". Take a look. At the very least you've a good chance of seeing a nice photo of me on my holidays in Cornwall last year :-) Whilst you're there, have a look at the MSDN Connection to get your own personalised news feed from MSDN.

And now for another "message from our sponsors" ;-) I'm speaking at DevWeek next week down in London Village somewhere. You can find the details from their website , I'm on the SQL Server track on Day 2 There's also a whole bunch of my colleagues speaking as well. We have the fantabulous Stephen Turner speaking just before me and we have the lovely Eric Nelson speaking on the first day . Over in the Visual Studio track we've got the dynamic Mr Mike Pelton talking about Office on Day 1 and we have the highly architectural chap Mike Platt speaking on Day 2 If anyone wants to meet up or have a chat I'll be around the conference on day 1 and 2 so drop me a line and we'll have a coffee or something.

Don Box (no less) has just posted an entry around ServiceContract's in Indigo with some detail as to how various bits fit together. Generally, I try and skim-read blog postings but, from time to time, I have to go and "Mark Post as Unread" with a view to going back and reading it again. With Don's post I had to do just that, go back and read it properly and, with a proper reading, it makes a lot of sense and (once again) Indigo looks to be designed in the way that you'd expect it to work.

I'm in the process of trying to put together a bit of code but I can't seem to find a spare 3-4 hours to sit down and sort something out. Sometimes I really miss those days where I could stick my head down for 16 hours of code and no-one complained or sent me email that needed answering! In the meantime, something that keeps interrupting my thoughts is this new HTC "magician" device which seems to be be a true mix between PocketPC and SmartPhone. Now...I've owned PocketPC's and I've owned the SmartPhones and my personal preference has long been for the SmartPhone. The primary reason I've got for this is that it fits into my pocket and portability is my number one requirement for a device like this. However, my one biggest complaint about the SmartPhone (and it's not really the fault of the device) is that it's a read-only device. I can read my mail on it (up to a point) and I read one or two websites on it if I'm stuck somewhere with nothing to do but I never, ever write anything on the device (although...