Regression test selection and product line system testing

Abstract— Context: Software product lines (SPL) are used in industry to achieve more efficient software development. To test a SPL is complex and costly and often becomes a bottleneck in the product line organization. Objective: This research aims to develop and evaluate strategies for improving system test selection in a SPL. Method: Initially industrial practices and research in both SPL testing and traditional regression test selection have been surveyed. Two systematic literature reviews, two industrial exploratory surveys and one industrial evaluation of a pragmatic test selection approach have been conducted. Results: There is a lack of industrial evaluations as well as of useful solutions, both regarding regression test selection and SPL testing. Test selection is an activity of varying scope and preconditions, strongly dependent on the context in which it is applied. Conclusions: Continued research will be done in close cooperation with industry with the goal to define a tool for visualizing system test coverage in a product line and the delta between a product and the covered part of the product line.