I too have commonly used various 'retry-on-failure' methods in my testing strategies. Some of these have been available in testing frameworks, others have been home grown.
Having Test Studio offer something out of the box to make this work nicely with UI tests without requiring custom code or custom deployments/schedule hacks would be very useful.
Anything to help cut down on the false positives which sometimes occur in UI testing!
Maybe Telerik can think of a way to differentiate they way they do this from everyone else. Perhaps rather than failing, catching the failure on the outside of the test and re-executing, what if the flow of the steps within an executing test could be somewhat dynamic. Sort of like the 'continue on failure' option for any step - what if there was a 'move to step 1 on failure'. It would need to be controlled to not loop infinitely. That would effectively give the effect of restarting the test on failure but be contained within the execution of a test rather than being handled externally after a hard failure.
ADMIN
Daniel Djambov
Posted on:11 Nov 2014 10:42
Telerik: here is an implementation for such scenario using Test Studio Execution extension: http://www.telerik.com/forums/automatic-re-run-of-failed-tests-from-a-test-list-execution
ADMIN
Konstantin Petkov
Posted on:26 Jul 2013 10:17
Hello,
A related issue for the execution in general:
http://feedback.telerik.com/Project/117/Feedback/Details/54452-after-a-test-list-is-executed-allow-failed-tests-to-be-executed-as-a-test-list
Thanks!