Application And Implementation Of Test Data Generation Based On Baseline Method

In the software life cycle, software testing plays a very important role, is animportant means to insure the quality of software. With the increasing of softwarecomplexity, the software test is more and more difficult, and the consumption of costis increasing. Test data generation is a key part of software testing, automation testdata generation process will improve the efficiency of software testing, and reduce thecost.First of all, in this paper, software testing’s principles, contents, process modeland technical classification is introduced, and the four types of software testingprocess model, static test and dynamic test is analyzed. This paper summarizes andanalyzes the existing automatic test framework model, and puts forward a test data tocreate frame model. The structural test data generation is deeply researched. Programcontrol flow graph automatic generation method, and test path automatic generationmethod is analyzed, program control flow graph generation algorithm, and testingpath automatic generation algorithm is proposed. At last, combination of Gupta’siterative relaxation method, automatic test data generation tool based on pathCoverage is developed.In which, the test paths of the automatically generated based on control-flowcoverage is the improvement of Baseline Method which broached by McCabe, is anew algorithm. The original baseline path is generated randomly and automatically. Inrespect of traversing the basic path group, the approach of depth-first search isintroduced to avoid storing more information about branch node; and reduce the timeof linear traversal. Using the column which S-node with no in-degree located at, andthe row which E-node with no out-degree located at, the case of nodes’ value andnumber of value case are stored for reducing the waste in storage. Giving priority tothe behind path-node when generating a path, the times of judgment for loops areeffectively reduced. Theoretical analysis and experimental results show that this newalgorithm has an advantage over the original Baseline Method in efficiency and spacecomplexity of the test path generation.