| With the development of the technology of software testing, in order to discover and fix the defects in the software as early as possible, many enterprises and research institutions have used the code analysis tool based on static analysis technology to detect the defects in the code. However,the static analysis of the code can't have both reliability and completeness,the initial results are large and there may be a lot of false positives, so we may need to spend a lot of manpower and time to confirm. The Defect Testing System (DTS), developed by the author's team, can be used for static analysis of programs developed in C, C++, and Java languages. Since the technique used by DTS is static analysis, it is also necessary to manually confirm the initial results of the analysis. However, the manual verification of the defect to the software tester to bring a great burden, and it is difficult in a relatively short period of time to complete the confirmation of all defects. Based on this problem, this paper proposes a defect priority partitioning method based on multi-dimensional analysis,and prioritizes the easily recognizable defects in the code.The priority classification of the defect alarms mainly considers the following factors: the computational accuracy of the defect, the complexity of the data source of the defect and the complexity of the solution. These three aspects are the progressive relationship between them, each of them will give a part of the defect of the final grade value. In view of these three aspects, this paper respectively designed the hierarchical strategy and its key algorithm, and wrote the code to realize it on the DTS system platform.Finally, according to the result of experiments with some open source projects, we verify the correctness and effectiveness of the proposed method. The experimental results show that the method can accurately classify the initial results of static analysis, and make the tester more specific in manual verification period. |
PDF Full Text Request