A Feature Model (FM) is a compact representation ofall the products of a software product line. The automatedextraction of information from FMs is a thriving researchtopic involving a number of analysis operations, algorithms,paradigms and tools. Implementing these operations is farfrom trivial and easily leads to errors and defects in analysissolutions. Current testing methods in this context mainly relyon the ability of the tester to decide whether the output ofan analysis is correct. However, this is acknowledged to betime-consuming, error-prone and in most cases infeasibledue to the combinatorial complexity of the analyses.In this paper, we present a set of relations (so-calledmetamorphic relations) between input FMs and their setof products and a test data generator relying on them.Given an FM and its known set of products, a set ofneighbour FMs together with their corresponding set ofproducts are automatically generated and used for testingdifferent analyses. Complex FMs representing millions ofproducts can be efficiently created applying this processiteratively. The evaluation of our approach using mutationtesting as well as real faults and tools reveals that mostfaults can be automatically detected within a few seconds
展开▼
