11-30-2018 03:47 AM
Hi all,
up to now I used the Sequential Model and an wrapper who chooses the test cases and replaces the "dummy Test case" with the correct one to call it later on in a new execution; In the test case (test sequence) I used the Model Options und Report Options callbacks.
Later on
- a status with more states than passed / failed/terminated should be reported back to the caller to be able to react different
- the test cases should still use the callbacks independent if called by the wrapper or not
- the parallel Model should be used later on with the socket index
- a new user Interface needs to be written for that (4 independent test sockets).
As the Process Modell defines also the order of the callbacks, I better ask an expert :).
Questions:
how to pass a parameter into a sequence in a new execution, I found 2 possible approaches, one with the callbacks (use process modell) but I need to add a sequence name as additional parameter.....; Are there any other differences between the parallel model and the sequential modell that I need to have in mind?
So all these Decisions have an impact on the parameter list so that's why I need to be shure of the optimum approach before I change 180 Testcases (a lot of maintenance work when switching between sequential modell and parrallel model approach 🙂
Thanks a lot.
Best regards
M. Schreiner
11-30-2018 04:31 AM
One more Question about the parallel model: I (plan to) use 4 Testsockets. If a sequence is called from one of the 4 parallel runs and I use the same Process Model for the called sequence, does that mean that 4 more Testsockets (4 x 4 = 16) are created? So do I need to have the sequential model for the Testcase and the parallel model for the wrapper?
Thanks
Best regards
Matthias
01-17-2019 04:30 AM
Hi run321,
is this still actual topic? I see it was added some time ago.
Let us know, therefore we will try to help 🙂
Regards,
01-17-2019 06:28 AM
Hi Patrik,
thanks for your message. No the topic is not really an actual one any more.
After a lot of discussions we decided to use separate Testsand instances rather than the parallel model.
There was still a lot to do (to change) to enshure that
- with the wrapper the HW Adress is needed only once (for the whole sequence of tests)
- without the wrapper the HW Adress GUI starts in each single test
- with the wrapper if a test is interrupted there is a retry
- every single Test is a new Execution with it's own report
....
So thanks for your offer 🙂
Best regards
Matthias
01-17-2019 06:38 AM
Hi Matthias,
I believe you would be able to achieve your results with parallel process model as well, however as you wrote in the title - if it's about reducing maintenance effor, this could be a better way.
Thanks for reply and good luck with implementing every feature in your test sequences.
In case of any hesitation, feel free to reach this beautiful NI community 🙂
Regards,