Dynamic Solution of Linear Constraints for Test Case Generation
Ernsting Marko, Majchrzak Tim A., Kuchen Herbert
The manual generation of test cases for unit testsis tedious. We have developed the tool Muggl, which generatestest cases based on symbolic execution and constraint solving.Solving constraints for this purpose is no trivial task and greatlyattributes to the total runtime. Hence, we developed a solver forlinear constraints adapted to the special needs of Muggl. It takesinto account the particularities of constraint retrieval throughsymbolic execution. Specifically, it is capable of incrementaladdition and backtracking of constraints. Moreover, we havedeveloped an approach to avoid rounding errors.
test-case generation; symbolic execution; Java; constraint solving; simplex algorithm