We are trying to evaluate our Silverlight Application with Test Studio and having some problems while recording the tests.
Currently we are using version 2012.2.920 of Test Studio since we had the problem "Unable to connect to Silverlight Application" with the most recent version 2012.2.1420.  Here is what we tryed:
After reviewing the documented solution we registred the MIME type application/x-silverlight-app to our servers. Afterwards Fiddler showed the correct response headers. Unfortunately the recording didn't work. We checked all other solutions from the documentation but the windowless mode was the only one which was helpful. Adding this parameter to the Silverlight object made the recording possible. But due to performance and compatibility issues we can not use the windowless mode for our application!

So my first question is: are there any other solutions to try?
Recording on other Silverlight Applications (like the Telerik Demos) works just fine.

---

With version 2012.2.920 we were able to successfully record and run multiple tests (without setting application/x-silverlight-app or enabling windowless mode!) so we decided to keep this version for now. A couple of days ago we wanted to create some new tests but encountered another problem. When navigating to our Silverlight apps the recorder freezes and its process has to be terminated. The strange thing is: in 1 out of 10 recording attempts it works, but >90% of the time it fails! The app is basically the same as before. There were no changes in the visual part of the screen. There is no difference when trying to record a local deployment/execution of the XAP (Visual Studio) and the deployment on the server. Both fail.

The Test Studio log file gives no information on the cause of the error because it logs just to the point where the recorder freezes. I attached the log to this post. Any suggestions how to find a solution to this problem?

I really hate the fact that IF I do a "run to here" ( see http://bit.ly/13TCXdK )
recording starts  upon finishing steps.  I am learning to turn it off with <printscreen>
right away as I keep getting burned by it recording steps I dont want,  BUT
I think an ounce of prevention would be very nice and keep me bit more sane.
 so, is there any way to set whether it is enabled or disabled?
If not, I think it would be a great enhancement 

(as a side note  - wouldn't it be cool if newly recorded steps added to an existing saved steps be color coded  or have state for each step with a flag for  new / edited  / saved)

Hi,

I am using  drag and drop functionality  as  shown in the link below between radscheduler and a radgrid to create appointments.
On dropping the record, It  gathers information related to dropped record and send the request to  webservice.
 Response received from webservice is processed  and displayed in a Confirmation RadWindow.
This RadWindow  has a label control display message received from webservice  and two buttons("Ok"/"Cancel")
User clicks  on "Cancel" or "OK" to create/cancel an appointment.

http://demos.telerik.com/aspnet-ajax/scheduler/examples/draganddropintegration/defaultcs.aspx?product=scheduler

 My webtest runs fine by creating new appointment on clicking "Ok" button, but when I use the same webtest for load test, it does
not create any appointments. Can someone  help me in solving this issue?

Thanks,
Prava



 

I've setup a db in Microsoft SQL Server Management Studio 2012 and connected it to the Reporter just fine. When running a test the Reporter signal goes green to red continuously (unless I stay on the Environment window) throughout the run. Tables and data are showing up in my db but not on my Analyze screen in Test Studio. I'm getting absolutely no test results. Any idea why this occurs?

Hello,

 

In our team testers are not developers whatsoever, so I’m looking to for a way to automate built testing without QA coding. I visualize it similar to this:

1.     1. Developers create basic test blocks which can accept parameters, for instance “Create new user with given name, email, etc.” Hopefully Devs can provide “Lego” for QA to build their sand castles;
1.     2. QA hopefully should be able to assemble their test suites from those “Lego” blocks using something like primitive scripting: a. Call Loging test, b. Call Create User test, c. Call Some other test, d. call verification “test” with particular parameters.

 

Does this sound feasible and easily doable with Test Studio?

 

Thank you!

Konstantin

 

Has anyone successfully scripted a step that Adds a URL to the Favorites in Internet Explorer? Most examples I find are out of date or do not work on TestStudio.

I have single trivial test which tries to navigate somewhere and executes 4 other steps.
I made it data bound so it does it for 3 different URLs. One of the navigations fails (bad URL).

When this test is part of the test list and it is run with the list, then its result looks quite clear: 14 out of 15 passed. Test has 5 steps, was run three times, so 15 total. In one case one step failed, so 14/15 looks right.

Now when I run the same test using Quick Execution it again executes 15 steps for exactly the same reason, but after that the status on top shows 5/5 passed! It only shows the last iteration, not all of them. This is VERY bad because unless sumebody is careful enough to open test run log and search for the word "Fail" there is no way to tell that one of the iterations actually failed.

Please let me know if I missed something.

Thank you!
Konstantin

While this topic seems to be discussed many times, I still cannot quite grasp it. I care about the scenario when test A calls test B and both of them are data bound. When those tests use separate data sources it works as expected. But if I tick "Inherit parent Data Source" for test B, then.. I don't understand what is going on.

Actual scenario I'm trying to achieve is - test B normally has its own data source which supposed to be used when test B is run standalone, BUT when test B is invoked as part of a bigger test I want it to inherit parent's data source and use that one instead. HOW test A datasource should look like so that it serves both A and B in this case?

Is there decent downloadable example for this scenario?

Thank you!
Konstantin

Is it planned for future release to implement "bunch of steps" or something similar?
I want to group several steps into logical groups.
As workaround it could be used LOOP steps with 1 count and text description of the steps meaning but it could be more convenient to be able to group several steps into group with some logical operation.

Hello

My test step 'Navigate to' fails because a redirection occurs once the URL is opened, to a URL which includes a query string.

How can I prevent this issue occuring? it happens quite frequently on the website i'm working with.

Regards


Hamish