Research On Automatic Test Case Generation For Complex Data Types

As an important technology to assure software quality, software testing is the indispensable part of software development. Test case generation, which is the core of software testing, has the profound significance to improve testing efficiency. Along with the rapid development of software technology, complex data types such as tree, graph, XML documents, are widely used in practice. Yet a complex data type generally consists of several members including intricate constraint relations, which makes it harder to generate effective test cases using traditional methods. Test case generation for complex data structure is a worth research theme.Valid imperative predicate, equivalent imperative predicate and finitization function are used to express constraints of complex data types. The results which satisfy the constraints are test data of complex data types. Automatic test case generation for complex data types based on equivalence, which is fully consulted for software robustness and cost, improves bounded exhaustive testing. Test case generation for program with complex data types works as follows, first generate test data for each parameter, and then combine the data according to certain rules. Solving process is as follows: get finitization function and imperative predicate, establish the state space according to the finitization function, select structures from the candidate structures, generate additional invalid structures and eliminate redundancy.Experiments demonstrate that comparing with bounded-exhaustive testing, this technique is more practical and achieves better testing results using less test cases.