Structural Statistical Software Testing with Active Learning in a Graph

Structural Statistical Software Testing (SSST) exploits the control flow graph of the program being tested to construct test cases. Specifically, SSST exploits the feasible paths in the control flow graph, that is, paths which are actually exerted for some values of the program input; the limitation is that feasible paths are massively outnumbered by infeasible ones. Addressing this limitation, this paper presents an active learning algorithm aimed at sampling the feasible paths in the control flow graph. The difficulty comes from both the few feasible paths initially available and the nature of the feasible path concept, reflecting the long-range dependencies among the nodes of the control flow graph. The proposed approach is based on a frugal representation inspired from Parikh maps, and on the identification of the conjunctive subconcepts in the feasible path concept within a Disjunctive Version Space framework. Experimental validation on real-world and artificial problems demonstrates significant improvements compared to the state of the art.