Design And Implementation Of A Directed Test Case Generation Tool

Testing is an valid way to reduce the software faults with the urgent demand for the computer software dependability. However, the cost of testing is high and both the mental and physical costs incite and stimulate the development of automated testing technique. Automated testing enables testers generate test cases automatically and troubleshoot the errors. Directed test case generation, which aims to reproduce the bugs, eliminates specific errors and collects software environment, is a sub-field technique of automated testing as well as an assistance for testers.Automated testing contains various technologies like random testing, concrete exe-cution dynamic symbolic execution (DSE), etc. DSE is an automated white box testing technique, which simplifies the restrains solving process by combining the idea of sym-bolic execution and concrete execution. Due to this improvement, DSE is qualified for unit testing and global testing under all kinds of criteria and brings its reputation both in academia and industrial companies. However, no particular algorithm is available on main stream DSE tools for directed test case generation problem.To tackle this problem, in this paper, we proposed an algorithm for fixed point test case generation. The algorithm, based on the critical edge, is able to prune the un-reachable states before the execution and guide the state selection through distance computation during the execution. Besides, the algorithm is realized in the test framework where LLVM is utilized as front processing and KLEE is applied as symbolic execution engine. Details of the algorithms and its implementation on the test framework are par-ticularly discussed. The test framework is tested against other directed testing algorithm and symbolic executions tools. The results on the test file of Coretuils demonstrate the effectiveness and stability of our algorithm as well as the feasibility of the test framework.