语句包裹模式的识别和分类

Identification and classification of statement encapsulation patterns

软件演化过程中会产生大量变更代码,对变更代码的识别有利于变更理解。其中普遍存在的把一个语句或语句序列移动到一个或多个不同的语法实体中的变更行为,对于这种语句包裹模式的识别和分类,提出了基于代码变更块和抽象语法树的语句包裹模式识别分类算法。首先从变更前后版本2个文件中筛选出代码变更块,根据语句包裹模式的特征找到候选代码变更块,再建立抽象语法树,通过语法分析找到代码变更块中存在的语句包裹模式并对其进行分类。该算法在4个开源项目中进行了实验验证,实验结果表明该算法对语句包裹模式的识别具有较高的准确率。

During the software evolution process,a large number of change codes will be generated,and the identification of the change code is conducive to change understanding.Among them is the universal behavior of changing a statement or a sequence of statement into one or more different grammatical entities.For the recognition and classification of such statement encapsulation patterns,the paper proposes a statement encapsulation pattern recognition classification algorithm based on code change blocks and abstract syntax trees.First,the code change block is filtered from the two files before and after the change,and the candidate code change block is found according to the characteristics of the statement encapsulation pattern,then an abstract syntax tree is established.Finally,syntactic analysis is used to find and classify the statement encapsulation patterns existing in the code change block.The algorithm has been experimentally verified in four open source projects,and the experimental results show that the algorithm has a high accuracy rate for statement encapsulation pattern recognition.