Research And Implementation Of CSP-Based Data Flow Testing System For JavaME Programs

Software testing is time-consuming and hard sledding. In order to reduce the cost of manual software testing and to increase the reliability of the testing process, research in automatic software testing has always been very active. The focus of this thesis is the automatic testing of JavaME programs. Such programs have been widely used in embedded devices such as mobile phones, PDAs and the like, and JavaME (or J2ME) has almost become the standard development platform for embedded applications in recent years.Test data generator is one of the most important components in a automatic testing environment. Through the years several attempts in automatic test data generations have been made. At present, there are many techniques supporting generation of test input, among them are random-based algorithm, genetic algorithm, the predicate constraints and so on. However these techniques don't fit into the embedded computing environment very well, the research in automation of JavaME software testing is in its primary stage. The quality of test data they generated is not high. For example, random-based and genetic algorithms are with a certain blindness in choosing the test data. The constraint-based testing technique improved the test data quality but with a low solving efficiency for the constraint satisfaction problem and is not easy to find a suitable solution.Based on the above observations, we propose a constraint-based automated testing framework, where test data is generated through solving a constraint satisfaction problem (CSP) by using a data flow analysis and a dynamic domain reduction technique. We have verified the feasibility of our method by experiments with various programs, where the results show that both the quality and efficiency of test data generation are improved by the combination of dynamic domain reduction and data flow testing.