Use of constraint solving for testing software product lines

摘要: A new software engineering methodology, product line (SPL) engineering, has been increasingly studied in academia and adopted industry the past decade. It allows delivery of similar, but customized, products to customers same domain within a short time period. Software produces an SPL by defining feature commonality variability, is supported well-managed asset base. can improve productivity from three ten times, however, we require more efficient testing methods, so that ensure correctness SPLs with resource allocation percentage as traditional engineering; methods applied longer lifecycle. In this dissertation, show how modern constraint solvers be used tackle challenge efficiently ensuring dependability two perspectives: sampling reuse. In sampling, key choose subset are representative whole configuration space. We focus on one technique, combinatorial interaction testing, samples combinations variability SPL. reuse, goal leverage inherent property SPLs: similarity, which stems fact all configurations generated core set common variable features. Our primary contributions improved sample generation techniques for incorporate constraints between features, reuse similarities during integration testing.