Can a Zephyr test that is marked as Failed be re-executed

[Jeanne Howe]

Currently, if we want to re-execute a test that has Failed, we must first move it to the WIP status. Can this be changed so that Failed test cases can be rerun without needing to be reset to WIP?

[Swapna Kumar Vemula [Zephyr] · Accepted solution]

Hello JT,

As per the functionality of Zephyr for JIRA it is not mandatory to change the execution status to WIP before you execute a failed test.

An execution status can be changed to any status as per the execution results and there is no restriction on this at any state.

So, You can directly change the status of a failed Testcase to any without resetting to WIP.

Thank You.

Swapna Kumar Vemula.

[Saradha Narayanan]

Is there any suggestion or updates from zephyr for handling this

[Paul Henckel]

But there is no versioning when changing the state of a testcase is there? So if a testcase failed 4 times before it finally passed, all the information about the failed executions will be lost right? That development progress is too valuable to loose, testcases that fail frequently indicate aspects of the program that are too shaky.

The only workaround I have been able to find is to selectively add the failed tests to a new test suite for re-execution.

But in my opinion that procedure is rather cumbersome so I want to point you to a concept which could help impove it. The add-on Structure.Testy provides a grid interface where you have all your testcases as individual rows and then each progressive Test Run can be added as a new column where you can specify whether the test failed again or passed.

So my feature request is this: Make it possible for the user to add new Test Runs as new columns on the existing list of issues in the test suites the way that Structure.Testy does.

[Nishant Shah]

Hi Team,

This is a really important feature and will really appreciate your some sort of reply on this.

Yes, currently I can change a failed test in a test cycle to Pass. But, there is no versioning happening in this case.
It will be really helpful if some versioning happens in this case as well.

Even for a single failure, currently, I have to create a new Test Cycle and add the failing test case to it and then execute it. This is quite painful.