Skip to Main Content
Researchers have explored the application of combinatorial interaction testing (CIT) methods to construct samples to drive systematic testing of software system configurations. Applying CIT to highly-configurable software systems is complicated by the fact that, in many such systems, there are constraints between specific configuration parameters that render certain combinations invalid. In recent work, automated constraint solving methods have been combined with search-based CIT methods to address this problem with promising results. In this paper, we observe that the pattern of computation in greedy CIT algorithms leads to sequences of constraint solving problems that are closely related to one another. We propose two techniques for exploiting the history of constraint solving: (1) using incremental algorithms that are present within available constraint solvers and (2) mining constraint solver data structures to extract information that can be used to reduce the CIT search space. We evaluate the cost-effectiveness of these reductions on four real-world highly-configurable software systems and on a population of synthetic examples that share the characteristics of those four systems. In combination our techniques reduce the cost of CIT in the presence of constraints to that of traditional unconstrained CIT methods without sacrificing the quality of solutions.