Research On Automatic Test Data Generation For Dynamic Data Structure

Software testing is one of the most important methods to assure the quality and correctness of software. It takes up much more proportion of the whole software development along with more deeply people realized for the essentiality of it. In the process of software testing, the key step is test data generation, which influences the efficiency and effect of software testing directly. While manually generate test data is a task which very labor-intensive and expensive, the research for automatic test data generation has vastitude application foreground.Most existing approaches for test data generation only can deal with numerical data types, such as integer and real numbers. Although these data types are commonly useful in various application, they cannot be feasibly used where the data size is not fixed, and some related approaches have less efficiency in the aspects of dealing with dynamic data structure and pointer alias. These restrict the application of test data generation to a certain extent.This research presented a new test data generation approach for dynamic data structure, which includes four main phases: test path generation, path constraints collection, path constraints settlement and test data generation. Firstly, this research proposed a new test path generation approach, which can generate a test path automatically. Collected the control flow and data flow information of the tested program, and generated a path which can reach the target statement. Secondly, this research proposed a new test data generation approach which can deal with dynamic data structure. Simulated the real execution of the tested program, collected the constraints along the test path which generated previous, and at last generated test data in the process of constraints settlement or decided the target statement can not be reached. There are two difficults of the research. The one is how to generate a test path which can reach a given statement automatically, and the other is how to generate test data for dynamic data structures.The experiments showed that this approach can generate a test path which can reach the target statement quickly, and collected the constraints along the path efficiently, or decided the target statement can not be reached. The results of the research not only satisfy the need of testing, but also guide the bug detect at a certain extent, then improve the test efficiency.