Automated Test Case Generation Considering Object States in Object-Oriented Programming

Testing for object-oriented programs is complicated and burdensome. One reason is the difficulty associated with generating method sequences that create instances and change object states to achieve high branch coverage. Automated test generation using mixed static and dynamic analysis is not only an effective approach to save time and reduce the burden of testing, but also an efficient way to find bugs. Seeker is an implementation for automated test generation that includes method sequences using static and dynamic analysis. However, when we need to change several values of variables to cover branches, the technique cannot generate method sequences that achieve all of the desired object states. In this paper, we extend the technique for automated test generation when multiple object states are required for the coverage of more complicated branches. Our approach identifies all variables that are involved in uncovered branches and evaluates method sequences according to a fitness function. Then we apply a retrieval strategy to suppress combinatorial explosion. Our results show that the proposed approach achieves higher branch coverage than that used in a previous study and also suggest that the effectiveness of the proposed approach tends to vary according to the specific characteristics of different projects.