Regression test is a very important but expensive process during software test. Therefore research on how to reduce the cost of regression test is especially important, and regression test selection technique has allured many eyes of researchers.This paper first briefly introduces regression test selection (RTS). Then we discuss two methods of RTS, one method used for RTS of procedure-oriented software, and the other for RTS of object-oriented software. Regarding the problem that algorithms available cannot handle sequential-call, this paper presents a new method called 'link', which can thus improve preciseness of RTS.Moreover, to improve confidence of software modification in regression test on condition that not all tests can be executed, we introduce test cases prioritization. After discussing several algorithms of test cases prioritization, this paper presents a new algorithm, which is called JIG-based brach-coverage prioritization algorithm. By combining the RTS technique metioned above and JIG-based brach-coverage prioritization algorithm, this paper presents a new method, which is called JIG-based regression test selection with prioritization.This paper also discusses how such factors as test frequency and demand of fault detection affect effectiveness of RTS.Finally, we implement a regression test selection system based on JIG-based regression test selection with prioritization.
|