Friday, 20 March 2009

Test framework - 'stop if too many errors feature'

Another feature I implemented a few weeks ago is 'stop if too many errors'.

Since an e-mail is sent for each test case failure, the e-mail inbox could be flooded if for example, an unhandled test environment issue occurs or the AUT build is completely broken. So to prevent that I have a parameter for how many test case failures that are tolerated before stopping a test run. Not that complicated either but good to have (especially if the e-mails are sent to a shared address...)

:-)

Friday, 13 March 2009

Test framework - Stop feature

I implemented a new feature in my automation framework today...

My test suites are usually very long scenarios and sometimes it takes hours to get to a test case which is by the end of a scenario. Imagine creating new test cases at the end, running the suite and a script failure occurs which in worst case require total rerun (that is also a major reason why I have a 'Resume' test feature)...

...so I thought a stop feature would be nice to have so I implemented a new state/status for a test case run status; 'Stop'

It work like this: Once a test run is created (more clues about my data model will be published here within a month or so, I hope) you set test case run status to Stop for a test case (via Report GUI) and the test execution will stop when the test case is up next in the test case execution queue (and sending an e-mail about it). Now, you can resume the test in debug mode and fix errors on the fly (remember to change status to 'No run' first). Or perhaps backup the AUT database before you continue?

That's it, not that complicated (I wonder why I haven't thought of it before)...