Open AccessPSP‐Finder: A Defect Detection Method Based on Mining Correlations from Function Call Paths
Author(s) -
Cui Zhanqi,
Chen Xiang,
Mu Yongmin,
Pan Minxue,
Wang Rongcun
Publication year - 2018
Publication title -
chinese journal of electronics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.267
H-Index - 25
eISSN - 2075-5597
pISSN - 1022-4653
DOI - 10.1049/cje.2018.04.001
Subject(s) - computer science , function (biology) , data mining , pattern recognition (psychology) , artificial intelligence , biology , evolutionary biology
Large scale programs usually imply many programming rules, which are missing from the specification documents. However, if programmers violate these rules in the process of programming, it is possible to introduce software defects. Previous works on mining function call correlation patterns only use structural information of the program, while control flow, data flow or other semantic information of the program are not exploited in those approaches. As a result, the defect detecting ability is restricted and high false rate is caused. This paper proposes a defect detection method based on mining function call association rules from program paths, which can be provided by simple static analysis. Then, the programs are automatically checked against the function call association rules for detecting suspicious defects. Based on this approach, experiments are carried out on a group of open source projects. The experiment results show that this approach can improve the capability of detecting defects and find more bugs related to program execution path. In addition, the false positive function call patterns and the overhead for manually validating suspicious defects are reduced.