Hierarchical regression test case selection using slicing
Durga Prasad Mohapatra, Subhrakanta Panda · International Journal of Computational Science and Engineering · 2017
In this paper, we propose a novel regression test case selection approach by decomposing an object-oriented (OO) program into packages, classes, methods and statements that are affected by some modification made to the program. This decomposition is based on the proposed hierarchical slicing of an OO program. By mapping these decompositions to the existing test suite, we select a new reduced regression test suite and add some new test cases, if necessary, to retest the modified program. We apply hierarchical slicing on a suitable intermediate graph proposed for representing an OO program. This intermediate graph representation corresponds to all the possible dependences among the different parts of an OO program. We improve the scalability of the intermediate graph to a considerable extent by identifying and removing the redundant edges from the graph and thus detect the affected program parts in less time. The average reduction in time achieved for all the ten programs under experimentation is approximately 28.1%. The test cases that cover these affected parts of the program are then selected for regression testing. The average reduction in the number of test cases selected for regression testing of experimental programs is approximately 56.3%.