Research On Regression Test Case Suite Reduction Based On Invariants

Regression test is costly,it is absolutely necessary in test process.With the deepening of the software development process,it needs frequent regression testing,the size of test suite grows rapidly, and there is large number of redundant test case in it.The key and difficult point in regression testing is how to reduce the regression test case set and perform effective testing,and reduce the test cost of regression testing,improve the efficiency of testing.The autonomy of nuclear power software is the key factor of China's nuclear power autonomy.There is high demand on the safety and reliability of nuclear power software because of its particularity.Based on NESTOR's software testing,we propose a method of reducing the regression test case in order to solve the problem of test cases redundancy.Nuclear power software has complex calculations and a large number of functional program invariants,however,Daikon's built-in invariant library does not involve the complex function invariants.According to this,this paper proposes a method based on GEP-RNC exponential logarithmic invariant discovery method,which extends the types of invariants in Daikon.Then,a solution is proposed for the problem that the regression test suite is too large.Using Daikon to extract program invariants,to determine whether changed the current program invariants.We can classify the test cases according to the information.Combined with the invariant information and change information of thetested program,to achieve the regression test case set reduction.And then apply this method to the regression test case set of nuclear power software.Through the test,this method can reduce the workload of regression testing and improve the efficiency of the test.