Dynamic analysis of shared execution in software product line testing

Abstract

Software product line (SPL), a family-based software development process, has proven to be a more effective technology than single software systems. Testing SPL products individually is redundant for product lines testing. Meanwhile, the complexity of systematically testing SPL programs is combinatorial, which limits the scalability of testing SPL. In this paper, I will introduce the idea of accelerating SPL testing by dynamic analysis. I intend to investigate a dynamic analysis of SPL testing, which analyzes all variants during the execution of a SPL, and forks the execution only a variant leads to a different state. In addition, to get the preliminary result, we can encode a family of variants into a single program to simulate a realization of SPL testing. The experimental result collected on the mutated C program, printtokens in Software-artifact Infrastructure Repository, shows our approach can accelerate the SPL testing with a speedup up to 4.83X than testing each variants at a time and pruning 87% of the variants.