We present a three objective formulation of regression test prioritisation. Our formulation involves the well-known, and widely-used objectives of Average Percentage of Statement Coverage (APSC) and Effective Execution Time (EET). However, we additionally include the Average Percentage of Change Coverage (APCC), which has not previously been used in search-based regression test optimisation. We apply our approach to prioritise the base and the collection package of the Guava project, which contains over 26,815 test cases. Our results demonstrate the value of search-based test case prioritisation: the sequences we find require only 0.2% of the 26,815 test cases and only 0.45% of their effective execution time. However, we find solutions that achieve more than 99.9% of both regression testing objectives; covering both changed code and existing code. We also investigate the tension between these two objectives for Guava.
CITATION STYLE
Bian, Y., Kirbas, S., Harman, M., Jia, Y., & Li, Z. (2015). Regression test case prioritisation for Guava. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 9275, pp. 221–227). Springer Verlag. https://doi.org/10.1007/978-3-319-22183-0_15
Mendeley helps you to discover research relevant for your work.