Generating Refactoring Proposals to Remove Clones from Automated System Tests

Automated system tests often have many clones, which make them complex to understand and costly to maintain. Unfortunately, removing clones is challenging as there are numerous possibilities of how to refactor them to reuse components such as subroutines. Additionally, clones often overlap partly which makes it particularly difficult to decide which parts to extract. If done wrongly, reuse potential is not leveraged optimally and structures between tests and reuse components will become unnecessarily complex. We present a method to support test engineers in extracting overlapping clones. Using grammar inference algorithms, we generate a refactoring proposal that demonstrates test engineers how overlapping clones can be extracted. Furthermore, we visualize the generated refactoring proposal to make it easily understandable for test engineers. An industrial case study demonstrates that our approach helps test engineers to gain information of the reuse potential of test suites and guides them to perform refactorings.