The test case can only be evaluated as a whole. So if you run the test case, and it fails because one test step failed, and then manually re-run that test step and the step passes, it will not update the test case status.
I would not call this a flaw, but rather a design method. The test case is evaluated as a whole, when the test case is run. The status will be set until the test case itself is ran again. If the test step is failing the first time, do you need to introduce a delay if it passes again when you manually run it later? Why is it failing the first time? That would be what I would point you to addressing.
I figured out why test step was failing.
parallel execution was overloading the web service and that failed some test steps,
If I force the test cases to run sequentially I do not have that issue.
Hey Sanj - how did you manage to fix this? Did you just change the order, or did you add a wait or something?
