基于程序依赖图的Aspect挖掘技术

为了有效地识别遗产代码中因横切关注点而导致的零散代码或交织代码,提出了一种基于程序依赖图的Aspect挖掘方法.根据横切关注点的本质特性,考察了横切关注点与核心关注点之间的依赖关系及其在依赖图上的表现形式,对于单个方法体,结合程序的语义信息,在程序依赖图中引入一个虚拟结点表征方法体的核心功能,所有影响核心功能的语句都与之建立依赖关系,然后对依赖图进行拓扑分析,先从虚拟结点开始分离出核心代码,再通过求解连通分量的方式挖掘方法体内横切代码.实例分析表明,该方法不仅能识别出方法体内的零散特性代码,而且还能有效地挖掘出交织特性的代码.

Apsect mining based on program dependence graph

In order to identify effectively scattering code and tangling code introduced by crosscutting concerns in legacy system,an approach based on program dependence graph(PDG) for aspect mining is proposed in this paper.According to the intrinsic attribute of crosscutting concern,the dependence relationship between crosscutting concerns and core concerns is firstly examined,followed by analyzing its structure presentation on PDG.For a single method,combining the semantic information of a program,a virtual node is firstly introduced into its PDG to represent its core functionality and all statements which impact the core functionality is virtually depended by it.Then,topology analysis on PDG is performed to figure out the core code starting from the virtual node and the crosscutting code determining by computing connected components in PDG.As a result,all potential crosscutting code can be separated from core functionality.A case study shows that our approach can discover effectively both scattering code and tangling code in legacy code.