The Research And Implementation Of Test Suite Optimization For Unit Regression Test

Unit regression test is a regression test carried out in the unit level. Unit regression test contributes a lot to the whole process of software test due to its characteristics like fine granularity, high repeatability and heavy workload. Thus the efficiency of unit regression test plays an important role in the software test.This paper raises a new modification-based algorithm for test cases selection and construction, which takes both the selected subset To as well as the newly-added subset T1as relevant test cases. The algorithm includes:effective modification points detection based on control flow graphs comparison; software slicing based on the data and control dependency among nodes of expressions; the To subset selection from baseline test case library which covers the node set in software slices; runtime uncovered nodes analysis based on To; new test cases construction to gain a whole coverage of all the nodes; building final test case set by merging of subsets To and T1.Compared with other state-of-the-art unit regression test case selection algorithms, the research brings the following improvements: the modification points detection algorithm can compute the changed point effectively; the software slice based on the modification point can narrow the test range to a large extent; new test cases construction can ensure the adequacy of the regression test. Therefore, this approach works better both on effectiveness and efficiency thanks to its precise test case selection without compromising test coverage.