The Research And Design Of Test Case Generation Tool Based On Constraint Programming

Test case generation is an important part of software testing. Using man-power to calculate test cases is a traditional test case generation method. But it can only reach a low test case amount, as well as not efficient. Some automatic test case generation software use random test case generation method to generate test cases. Random test case generate method can be very efficient to generate test cases, but it also have its disadvantages, because the test cases generated by random method will have a low-level cover ability.Test case generation based on constraint programming is researched in this paper. First, some related techniques are discussed, such as the static analysis of source code, control flow path selection and some other test case generation tools. Then the techniques used in constraint programming are discussed in detail. The variable interval initialization is one of these techniques. It can calculate an interval for every variable properly, using interval analysis results of control flow graph. The test case generation in such proper interval is efficient, and the cover ability of test cases generated is improved. At last the process of test case generation based on constraint programming is introduced. It uses constraint programming to analyze all constraint relations in the source code, and uses reverse symbolic symbol analysis to process constraint expressions. After that, it solves constraint expressions, and generates the test cases. The collaboration of test case generation based on constraint programming with the other modules is also introduced.Some open-source C projects were selected for evaluate the constraint programming test case generate method. The result of the test showed that using constraint programming test case generate properly, could improve the coverage of test cases and reach a better test effect.