Forum Discussion

andreeailiuta's avatar
andreeailiuta
New Contributor
8 years ago

Execution inconsistency when running chained batches vs running batches one by one

Hello everyone,   

 

We have several test cases split in batches. When running all batches one by one all test cases are passing, but when running chained batches, there are some tests that are constantly failing due to "window invisible" error.

We tried to investigate if there are some left overs remained between test cases, but as we can see, the application is cleaned up before staring a new test case.

Environment details:

The test cases are developed in TestComplete V12.42 and are run on a test machine, using TestExecute V12.42 where is configured Team Foundation Build Service.

 

Could anybody give us an idea of what could lead to this strange behavior?

 

Thanks in advance!

  • tristaanogre's avatar
    tristaanogre
    Esteemed Contributor
    8 years ago

    Timing....  there's probably some test step execution being attempted before the object is fully resolved/rendered on screen or the page hasn't completely loaded.  Issues like this I've seen occur more frequently when tests are running for a while against an application, especially as the application starts caching more memory.

     

    Take a look at the steps where the error is happening and make sure that you have all the necessary "Wait" calls and such added to have the automation be "smart".

     

    Another thing... is this a web application?  And is there a difference in environment between running them one-by-one versus running them in batch?  The reason being is that, while TestComplete is pretty good at making sure it scrolls components onto screen before it attempts to execute an action, it's not always successful.  We've occasionally needed to add a call on components of 'scrollIntoView(true)' to make sure that we scroll the component into view on screen before we attempt to execute a test step.  That's something else to investigate