Handling Failures in Automated Acceptance Tests

One of the aims of automated functional testing is to run many tests and discover multiple errors in one execution of the test suite. However, when an automated test runs into unexpected behavior-system errors, wrong paths taken, incorrect data stored, and more-the test fails. When a test fails, additional errors, called inherited errors, can result or the entire test can stop unintentionally. Either way, some portion of the system remains untested, and either the error must be corrected or the automation changed before proceeding. Alexandra Imrie describes proven approaches to ensure that the most tests will continue running despite errors encountered. She begins by sharing a specific way of designing tests to minimize the disturbance from an error. Using this test design as a foundation, Alex describes the strategies she exploits for handling and recovering from error events that occur during automated functional tests. In conclusion, Alex explains some process considerations for dealing with errors within the project team.