多粒度数据变更影响域分析方法

Analysis of multi-granularity data change impact domain

完整准确提取程序变更影响范围是分析由数据变化引发程序错误的关键,能够有效缩短回归测试的周期。为此,提出一种基于多粒度的数据变更影响域分析方法。通过静态分析,结合程序局部控制信息和数据信息流规则,提取关键函数,得到影响函数调用路径,以程序影响函数调用路径为分析集,提取受变更影响的基本路径集,即从函数粒度和语句粒度对数据变更影响域进行分析,分别得到影响函数调用路径和影响基本路径集。大量实验结果表明,该方法能够完整准确提取程序影响函数调用路径和影响基本路径集,定位数据变更影响域,提高回归测试的效率。

Extracting the impact range completely and accurately is the key to the analysis of the program errors caused by the change of data,and it can effectively shorten the cycle of regression testing.A multi-granularity analysis method of data change impact domain was proposed.The key functions were extracted by static analysis,combined with local control information and data information flow rules,the affected function call path was then got.The basic path set affected by the data change was extracted based on the program affected function call path.The influence domain of data change was analyzed from the perspectives of function granularity and sentence granularity,and the affected function call path and the affected basic path set were obtained respectively.Results of a large number of experiments show that the proposed method can extract the program affected function call path and affected basic path set completely and accurately,locate data change impact domain,and improve the efficiency of regression test.