Model-Based Test Cases Generation for Extended Data Flow Coverage Criteria

The paper presents methods and algorithms for automatic design of test cases from the formal models of transition systems. Certain specific extended data flow coverage criteria are used as a test goal targets. Particular attention is paid to the problem of infeasible coverage targets, which in fact are unreachable, nevertheless are required for static analysis. Such cases interfere with statistics and make it difficult to measure the completeness of the coverage results obtained. We describe a method for specifying data flow chains and propose algorithms that combine dynamic refinement of the coverage targets with the search for paths satisfying them. In many practical cases, our approach achieves a combinatorial reduction in the number of coverage targets and, if the algorithm is successfully terminated, the completeness of the desired coverage.