关键词: 测试充分度准则, 偶然正确性, 动态数据流分析

Abstract: This paper proposes a new test adequacy criterion based on coincidental correctness probability. By considering the impact of coincidental correctness on the effectiveness of many test adequacy criteria, it designs an algorithm to calculate the probability that because of coincidental correctness, the output is still correct even when there is an error in the program. First, the use-definition dependence among statement instances is tracked in a program execution using control- and data- flow analysis, and the use-definition dependence graph is built based on the analysis result. Then the effect degrees of statement structure and transfer relationship on coincidental correctness probability are specified. Finally it gives an algorithm to get the test adequacy of the program in this execution. To evaluate the proposed approach, it conducts a mutation analysis on five simple C programs. Statement coverage cri-terion is compared with the new test adequacy criterion. The experimental results consistently approve the conjec-ture that using the coincidental correctness probability as testing adequacy criterion effectively quantizes the ability of a test suite to find errors.

Key words: test adequacy criterion, coincidental correctness, dynamic data-flow analysis