Efficient Testing of Different Loop Paths

Loops can represent an infinite number of possible execution paths and therefore purse a major challenge for current static analysis frameworks and test input generators. In this paper, we introduce a new loop exploration algorithm to examine different iteration orders (i.e. loop paths) in order to test distinct loop behaviour. To reduce the complexity of testing all possible combinations of iterations, we introduce a criterion to group different paths into equivalence classes and show how to specifically generate test cases that cover the different equivalence classes. We demonstrate how this approach helps to achieve higher coverage rates and helps to find software failures that are not discovered by current test case generation frameworks.