The approach that we have taken is to use the batch script to start the Telerik test scripts execution (ArtOfTest.Runner.exe) and then check the out through a custom C# console program that checks the output for execution status (Pass/Fail).
When we are executing the Telerik batch file without using Hudson the test script shows successful execution and simultaneously the C# code picks the same status. But when we are trying to execute the same through Hudson the test script executes but shows Fail status.
One thing that we did notice was that in the non-Hudson execution the Runner simulates the test script by opening the browser but in the Hudson environment the browser simulation is not executed by the Hudson environment as it is executing it as a background process. This might be a potential cause of the issue.
Thanks
10 Answers, 1 is accepted
do you have an active Remote Dekstop Connection opened to the Hudson Slaves executing your tests? Without a RDC Windows will simply render some GUI elements since there's no one to see them anyway. Since we're performing the GUI of an application, not keeping a RDC open to your slave might results in failed tests (I'm not sure whether the open connection that Hudson keep to its slave has any effect on this).
In any case I've used Hudson and I can confirm it can be integrated with WebAii tests. I'm not 100% sure what you mean when you say "Hudson environment" though - would you please clarify?
Greetings,
Rank 1
can you please provide as more detailed Information on how we can integrate Telerik WebAii tests with Hudson CI.
there's really no special integration required in order to do it. I simply create a FreeStyle project and add windows batch commands. The batch commands start ArtOfTest.Runner.exe and pass it the appropriate test. This is probably what you're doing as well.
Obviously, there's a lot more configuration/setting up required but that will depend on the particular situation.
Unfortunately UI testing offer some unique challenges. The Windows OS chooses not to render some UI elements when it deems on one is looking and that gets to be a problem. In order to avoid this you'll probably need an active desktop connection to any Hudson slaves you intend to use for testing. The Desktop connection will have to be open while you're performing the tests, don't even minimize it.
Of course you'll also need to be logged in with an administrator account to your slave machines (sleep and locking should be turned off).
Could you please give me the error message your tests generates on failure - it might hold some clue as to what you're doing wrong.
Hope to hear from you soon!
Regards,
Rank 1
If so then how will this impact on GUI tests
yes, if I remember correctly you could set the "Master" machine to also server as an Executor. It was somewhere in the configuration - I'm sure you'll be able to find how to do it on Hudson's site.
Please keep in mind that this is not the right place for purely Hudson-related questions - I just happen to have some previous experience with it.
Don't hesitate to contact us if you have any other questions related the WebAii Framework.
Regards,
Rank 1
Thanks for your support
Rank 1
So does this mean that Telerik cannot fully test UI components/websites in a batch or background mode?
Thanks,
Nick
That sort of depends on how you're defining "batch" and "background". There are multiple possible interpretations in an automated testing scenario.
Yes we can run a whole set of tests on an automated schedule. This really depends on your build & test environment for how this is accomplished. There are almost as many different environments as there are organizations.
No we cannot run more than one test at a time on one machine. Each test must be run serially.
Our tests that require moving the mouse or simulating keyboard input must be run in the foreground i.e. have the browser open, have focus on an open unlocked machine.
Cody
the Telerik team
Rank 1
Hi,
when I ran test scrips as batch execution some of the test scripts are not executed and also in log file the reason also not captured. Due to this I am unable to identify actual reason why some of the scripts are not executed.
Thank you for contacting us.
I would like to ask you to describe your exact scenario in more details as the thread you have updated is since more than 5 years. Do you use Hudson as well in your setup? Please elaborate more on your query.
Thanks in advance and hope to hear from you soon.
Regards,
Elena Tsvetkova
Telerik
Test Studio Trainings