Comments on: My Silverlight 2 blog roadmap for Summer ’08 http://www.jeff.wilcox.name/2008/07/my-summer-silverlight-blogging/ Silverlight, rich client apps and web development Fri, 30 Jul 2010 19:17:58 +0000 hourly 1 http://wordpress.org/?v=3.0.1 By: Jeff Wilcox http://www.jeff.wilcox.name/2008/07/my-summer-silverlight-blogging/comment-page-1/#comment-561 Jeff Wilcox Tue, 30 Sep 2008 07:36:42 +0000 http://www.jeff.wilcox.name/2008/07/13/my-summer-silverlight-blogging/#comment-561 @KierenH: You are correct. The CreateAsyncTest method is a little bit of syntax sugar on top of the same concept. It's really a psuedo-asynchronous concept. @Inquisitorjax: The web browser test harness doesn't use MSTest, since the Silverlight CLR and the execution environment are different enough that we went with a simpler, custom engine for hte time being - while keeping the same Visual Studio Team Test metadata. I should try and finish that post someday, sorry! @Mike, I don't think that there's an easy way to test WPF controls today in the same manner, but I know I'd love to see something happen there. My team's heavily involved in WPF as well, so if there isn't a way to do it today, rest assured we're going to work and find a solution someday. @KierenH: You are correct. The CreateAsyncTest method is a little bit of syntax sugar on top of the same concept. It’s really a psuedo-asynchronous concept.

@Inquisitorjax: The web browser test harness doesn’t use MSTest, since the Silverlight CLR and the execution environment are different enough that we went with a simpler, custom engine for hte time being – while keeping the same Visual Studio Team Test metadata. I should try and finish that post someday, sorry!

@Mike,
I don’t think that there’s an easy way to test WPF controls today in the same manner, but I know I’d love to see something happen there. My team’s heavily involved in WPF as well, so if there isn’t a way to do it today, rest assured we’re going to work and find a solution someday.

]]> By: Mike http://www.jeff.wilcox.name/2008/07/my-summer-silverlight-blogging/comment-page-1/#comment-506 Mike Wed, 17 Sep 2008 06:21:48 +0000 http://www.jeff.wilcox.name/2008/07/13/my-summer-silverlight-blogging/#comment-506 @Jeff, Please tell me is it the same way we can test wpf controls as well or we have to do it using Visual studio test framework and how can i test wpf controls asynchronously? @Jeff,

Please tell me is it the same way we can test wpf controls as well or we have to do it using Visual studio test framework and how can i test wpf controls asynchronously?

]]> By: http://inquisitorjax.blogspot.com/ http://www.jeff.wilcox.name/2008/07/my-summer-silverlight-blogging/comment-page-1/#comment-418 http://inquisitorjax.blogspot.com/ Wed, 27 Aug 2008 09:37:52 +0000 http://www.jeff.wilcox.name/2008/07/13/my-summer-silverlight-blogging/#comment-418 Jeff - you said: "One proof-of-concept project I did was get NUnit’s metadata to work within the Silverlight framework, just by creating a provider that maps NUnit metadata to generic unit test interfaces inside the Microsoft.Silverlight.Testing.dll assembly. I’ll be documenting this unit test provider functionality in a future blog post." can you expand a bit on this? Does this mean that you get NUnit defined tests to run against MS Test? Looking forward to mentioned post. Also - are there plans to enable a user to run selective tests from the test page? Jeff – you said:
“One proof-of-concept project I did was get NUnit’s metadata to work within the Silverlight framework, just by creating a provider that maps NUnit metadata to generic unit test interfaces inside the Microsoft.Silverlight.Testing.dll assembly. I’ll be documenting this unit test provider functionality in a future blog post.”
can you expand a bit on this? Does this mean that you get NUnit defined tests to run against MS Test?
Looking forward to mentioned post.
Also – are there plans to enable a user to run selective tests from the test page?

]]> By: KierenH http://www.jeff.wilcox.name/2008/07/my-summer-silverlight-blogging/comment-page-1/#comment-246 KierenH Tue, 05 Aug 2008 02:36:40 +0000 http://www.jeff.wilcox.name/2008/07/13/my-summer-silverlight-blogging/#comment-246 Thanks Jeff. Please correct me if I'm wrong, this is my understanding: It is a mechanism for ensuring tests that interact with the visual tree run on the UI thread. The CreateAsyncTest method in the SL2B1 controls source also makes sure the element has loaded first. Thanks Jeff. Please correct me if I’m wrong, this is my understanding:
It is a mechanism for ensuring tests that interact with the visual tree run on the UI thread.
The CreateAsyncTest method in the SL2B1 controls source also makes sure the element has loaded first.

]]> By: Jeff Wilcox http://www.jeff.wilcox.name/2008/07/my-summer-silverlight-blogging/comment-page-1/#comment-245 Jeff Wilcox Mon, 04 Aug 2008 20:22:50 +0000 http://www.jeff.wilcox.name/2008/07/13/my-summer-silverlight-blogging/#comment-245 KierenH, I have a complete draft of the asynchronous post and am just working on a few graphics to help illustrate the system. To answer your question, most of the "unit tests" that shipped with the controls are borderline "end-to-end" or integration tests: they obviously do interact with the underlying control framework and Silverlight platform. When working with Silverlight controls and visuals, a lot of the time it's important for those things to actually be placed within the visual tree for the events and other properties to fire and to be ready to test. So by using the Asynchronous test system, the CreateAsyncTest can help by giving the runtime the time to place the control in the visual tree. The asynchronous test isn't happening in another thread, but rather in the same user interface thread. So, about 5ms after that Create call was run, the callback action to perform the assertion is called. That is a long time in the CLR world, enough time for any pending visual tree updates and native interop to happen. Let me know if this makes sense or not.. -Jeff KierenH,
I have a complete draft of the asynchronous post and am just working on a few graphics to help illustrate the system.

To answer your question, most of the “unit tests” that shipped with the controls are borderline “end-to-end” or integration tests: they obviously do interact with the underlying control framework and Silverlight platform.

When working with Silverlight controls and visuals, a lot of the time it’s important for those things to actually be placed within the visual tree for the events and other properties to fire and to be ready to test.

So by using the Asynchronous test system, the CreateAsyncTest can help by giving the runtime the time to place the control in the visual tree.

The asynchronous test isn’t happening in another thread, but rather in the same user interface thread. So, about 5ms after that Create call was run, the callback action to perform the assertion is called. That is a long time in the CLR world, enough time for any pending visual tree updates and native interop to happen.

Let me know if this makes sense or not..
-Jeff

]]> By: KierenH http://www.jeff.wilcox.name/2008/07/my-summer-silverlight-blogging/comment-page-1/#comment-244 KierenH Mon, 04 Aug 2008 09:59:34 +0000 http://www.jeff.wilcox.name/2008/07/13/my-summer-silverlight-blogging/#comment-244 Jeff, hanging out for a deep-dive into [Asynchronous] testing. I can see the scenario for testing async services, but would like more info on how it is used in the SL2B1 source (that ships with tests). For example, why is this testing asynchronously: [Asynchronous] public override void FontSizeDefaultValue() { Button control = new Button(); CreateAsyncTest(control, () => Assert.AreEqual(11, control.FontSize, Common.HighPrecisionDelta)); } Taken from ButtonTest. Cheers Jeff, hanging out for a deep-dive into [Asynchronous] testing.

I can see the scenario for testing async services, but would like more info on how it is used in the SL2B1 source (that ships with tests). For example, why is this testing asynchronously:
[Asynchronous]
public override void FontSizeDefaultValue()
{
Button control = new Button();
CreateAsyncTest(control,
() => Assert.AreEqual(11, control.FontSize, Common.HighPrecisionDelta));
}

Taken from ButtonTest.

Cheers

]]> By: Miguel http://www.jeff.wilcox.name/2008/07/my-summer-silverlight-blogging/comment-page-1/#comment-233 Miguel Tue, 22 Jul 2008 00:01:48 +0000 http://www.jeff.wilcox.name/2008/07/13/my-summer-silverlight-blogging/#comment-233 Wow, would be great to be able to get it from codeplex. Thanks Wow, would be great to be able to get it from codeplex.

Thanks

]]> By: Jeff Wilcox http://www.jeff.wilcox.name/2008/07/my-summer-silverlight-blogging/comment-page-1/#comment-226 Jeff Wilcox Sat, 19 Jul 2008 22:17:07 +0000 http://www.jeff.wilcox.name/2008/07/13/my-summer-silverlight-blogging/#comment-226 Miguel, I've removed the debugger break statement from the code and will try and release an interim binary release soon. I'm working to get the framework out there on CodePlex with a MS-PL license, and hope to have some details when it's available. These things always take a little longer than it feels like they should... sorry I don't have a known date yet. Sincerely, -Jeff Miguel,
I’ve removed the debugger break statement from the code and will try and release an interim binary release soon.

I’m working to get the framework out there on CodePlex with a MS-PL license, and hope to have some details when it’s available. These things always take a little longer than it feels like they should… sorry I don’t have a known date yet.

Sincerely,
-Jeff

]]> By: Miguel http://www.jeff.wilcox.name/2008/07/my-summer-silverlight-blogging/comment-page-1/#comment-225 Miguel Sat, 19 Jul 2008 16:59:48 +0000 http://www.jeff.wilcox.name/2008/07/13/my-summer-silverlight-blogging/#comment-225 I misundestood the use of the property. I tought, than it wouldn't break if InterceptUnhandledException was true. I looked at the code in reflector and notice the first thing it does is a Debugger.Break call inside the GlobalUnhandledExceptionListener method. So what happens is that on any exception, even if handled by this Listener, the debugger breaks, this includes exception thrown by asserts. One option is to run it without debugging, but that means, we cant debug other methods either. For failed tests and excepted exception we dont need the debugger to break. That's the behavior of other testing frameworks. By the way, I noticed BrowerUtility also has an InterceptUnhandledException property, but it always returns true. I guess is confusing, specially because it has a set that has no effect. In your post 'Unit Testing with Silverlight 2' you mentioned you were goint to consider releasing the UTFx Source Code. Have this gone any further? I misundestood the use of the property. I tought, than it wouldn’t break if InterceptUnhandledException was true. I looked at the code in reflector and notice the first thing it does is a Debugger.Break call inside the GlobalUnhandledExceptionListener method.

So what happens is that on any exception, even if handled by this Listener, the debugger breaks, this includes exception thrown by asserts.

One option is to run it without debugging, but that means, we cant debug other methods either. For failed tests and excepted exception we dont need the debugger to break. That’s the behavior of other testing frameworks.

By the way, I noticed BrowerUtility also has an InterceptUnhandledException property, but it always returns true. I guess is confusing, specially because it has a set that has no effect.

In your post ‘Unit Testing with Silverlight 2′ you mentioned you were goint to consider releasing the UTFx Source Code. Have this gone any further?

]]> By: Jeff Wilcox http://www.jeff.wilcox.name/2008/07/my-summer-silverlight-blogging/comment-page-1/#comment-223 Jeff Wilcox Thu, 17 Jul 2008 19:34:01 +0000 http://www.jeff.wilcox.name/2008/07/13/my-summer-silverlight-blogging/#comment-223 Miguel, Definitely! Although, I actually think it is true by default right now (if you attach and look at the value, is that the case?). There are actually a lot of situations in today's Silverlight bits where the UnhandledException handler for the Application is -not- called. And that's where I hook up the InterceptUnhandledExceptions property... The reason for including this property was actually for a set of tests that I have that actually test the unhandled exception property - and so I disable the property before the test, and re-enable it after that. But my understanding is that it was True by default :-/ -Jeff Miguel,
Definitely! Although, I actually think it is true by default right now (if you attach and look at the value, is that the case?).

There are actually a lot of situations in today’s Silverlight bits where the UnhandledException handler for the Application is -not- called. And that’s where I hook up the InterceptUnhandledExceptions property…

The reason for including this property was actually for a set of tests that I have that actually test the unhandled exception property – and so I disable the property before the test, and re-enable it after that.

But my understanding is that it was True by default :-/

-Jeff

]]> By: Miguel http://www.jeff.wilcox.name/2008/07/my-summer-silverlight-blogging/comment-page-1/#comment-222 Miguel Thu, 17 Jul 2008 19:19:50 +0000 http://www.jeff.wilcox.name/2008/07/13/my-summer-silverlight-blogging/#comment-222 Jeff, Could you consider making InterceptUnhandledExceptions true as default as it is in VSTS and other testing frameworks. Jeff,

Could you consider making InterceptUnhandledExceptions true as default as it is in VSTS and other testing frameworks.

]]> By: Jonas Follesø http://www.jeff.wilcox.name/2008/07/my-summer-silverlight-blogging/comment-page-1/#comment-212 Jonas Follesø Mon, 14 Jul 2008 22:32:10 +0000 http://www.jeff.wilcox.name/2008/07/13/my-summer-silverlight-blogging/#comment-212 Hi Jeff, Thanks for the link back. Sounds like great stuff. Really looking forward to the next revision of the testing framework. Hopefully you'll have it out before September. Doing a "Silverlight 2 for developers" talk at TechEd New Zealand and Sydney first week of September. Unit testing and UI patterns will be an important part of the talk, so would be cool to have some fresh bits to show :). Keep up the good work and enjoy the Seattle summer ! Cheers, Jonas Hi Jeff,

Thanks for the link back. Sounds like great stuff. Really looking forward to the next revision of the testing framework. Hopefully you’ll have it out before September. Doing a “Silverlight 2 for developers” talk at TechEd New Zealand and Sydney first week of September. Unit testing and UI patterns will be an important part of the talk, so would be cool to have some fresh bits to show :) .

Keep up the good work and enjoy the Seattle summer !

Cheers,
Jonas

]]>