Testing the Evolution of Feature Models with Specific Combinatorial Tests
Andrea Bombarda, Silvia Bonfanti, Angelo Gargantini · 2024
Software Product Lines (SPLs) are commonly used for representing highly configurable systems, by using Feature Models (FMs). Like any other systems, SPLs (and consequently FMs) undergo changes over time, and developers may inadvertently introduce errors during this process. Consequently, it is of utmost importance to rigorously test SPLs during their evolution, in particular, by analyzing the changes introduced from one version of FM to its evolved counterpart.This paper presents the concept of “specificity”, enabling testers to discern tests specifically tailored for evaluating a model evolution. In addition, we present SPECGEN, a BDD-based algorithm that aims at producing more specific combinatorial test suites. We conduct a comparative analysis between SPECGEN and established approaches. Our experiments demonstrate that employing SPECGEN results in test suites with higher specificity, maintaining the same size, and accomplishing this in a shorter time compared to conventional methods.