Forum Discussion

touch725's avatar
touch725
New Contributor
14 years ago

GUI Problems Whent Testing with Network Suite

I've set up a task in a Network Suite that runs a test on another workstation.  Now, the test residing on the other workstation runs fine when the tester is logged into the workstation and the test is displayed, but when the test is run as a remote test through a network suite, the tested application's GUI doesn't seem to respond to test commands.  After the failed test run, the test logs indicate a lot of "The object does not exist." errors.  Are the tested application's GUI objects being repainted?  It would seem like it would have to be a requirement for this type of testing to work.  Am I missing something?

3 Replies

  • Can you provide some more details about your physical setup ? VM Vs physical ? Fireall status ?  subnets ?





    Can you verify the task by itself ?

    Can you run the task by itself ?



    The more details the better.  no small detail is too small. everything counts.
  • touch725's avatar
    touch725
    New Contributor
    I'm working with two workstations on a company intranet.  No firewalls.  No VM.  I suppose I'm not clear on how the firewall / subnet has anything to do with this, since the automated test on the remote workstation is starting fine. 



    Please follow me here.  In the scenario where user does not use distributed testing but is simply running an automated test on the actual workstation the test resides, the test runs fine.  Now, if, while that test is running, the user were to lock his screen, wait say a half an hour, unlock his screen, stop the test and view the log, there would be "The object does not exist" errors.  This indicates to me that the automated test, when behind the locked screen, is unable to hook into the GUI of the application. I've tested this out and that seems to be a fact.



    I then change scenarios and run the test via distributed testing.  When I do so, I see the same results as in the first scenario.  The task on the immediate workstation starts the test on the remote workstation just fine, as indicated in the previous email, but since the running test is not displayed, again it doesn't seem like the automated test is able to hook into the GUI.  Since the test actually started, it would seem that this would eliminate the firewall issue, since, if it were the firewall, the automated test on the remote workstation wouldn't start in the first place.  Rather, it is DURING the test run that I'm getting the "The Object does not exist" errors, as indicated in the test log after the test is stopped.



    Attached is the .mht log from the test.
  • Hi Jon,



    This is expected behavior for TestComplete 7 because the GUI of your application is not drawn by the system indeed. Please see the Running Tests on Locked Computers help topic for more information.



    This restriction is overcome in TestComplete 8, as this version of the tool automatically creates an RDP connection to the slave machine.