A Systematic Literature Review of Test Breakage Prevention and Repair Techniques

TL;DR

This systematic literature review consolidates current research on test breakage prevention and repair, highlighting the causes, existing tools, evaluation methods, and gaps, especially the lack of industrial application and publicly available tools.

Contribution

It provides a comprehensive taxonomy of test breakage causes and evaluates the state of test repair tools, identifying gaps and directions for future research.

Findings

Only four test repair tools are publicly available.

Most studies evaluate approaches on open-source case studies.

Few tools are adopted in industrial contexts.

Abstract

Context: When an application evolves, some of the developed test cases break. Discarding broken test cases causes a significant waste of effort and leads to test suites that are less effective and have lower coverage. Test repair approaches evolve test suites along with applications by repairing the broken test cases. Objective: Numerous studies are published on test repair approaches every year. It is important to summarise and consolidate the existing knowledge in the area to provide directions to researchers and practitioners. This research work provides a systematic literature review in the area of test case repair and breakage prevention, aiming to guide researchers and practitioners in the field of software testing. Method: We followed the standard protocol for conducting a systematic literature review. First, research goals were defined using the Goal Question Metric (GQM). Then we formulate research questions corresponding to each goal. Finally, metrics are extracted from the included papers. Based on the defined selection criteria a final set of 41 primary studies are included for analysis. Results: The selection process resulted in 5 journal papers, and 36 conference papers. We present a taxonomy that lists the causes of test case breakages extracted from the literature. We found that only four proposed test repair tools are publicly available. Most studies evaluated their approaches on open-source case studies. Conclusion: There is significant room for future research on test repair techniques. Despite the positive trend of evaluating approaches on large scale open-source studies, there is a clear lack of results from studies done in a real industrial context. Few tools are publicly available which lowers the potential of adaption by industry practitioners.