An automated Model-based Testing Approach in Software Product Lines Using a Variability Language.

Resumen

This paper presents the application of an automated testing approach for Software Product Lines (SPL) driven by its state-machine and variability models. Context: Model-based testing provides a technique for automatic generation of test cases using models. Introduction of a variability model in this technique can achieve testing automation in SPL. Method: We use UML and CVL (Common Variability Language) models as input, and JUnit test cases are derived from these models. This approach has been implemented using the UML2 Eclipse Modeling platform and the CVL-Tool. Validation: A model checking tool prototype has been developed and a case study has been performed. Conclusions: Preliminary experiments have proved that our approach can find structural errors in the SPL under test. In our future work we will introduce Object Constraint Language (OCL) constraints attached to the input UML model