面向对象软件反编译中类的识别方法研究  

Research on Class Identification in Decompilation of Object-Oriented Software

作　　者：周国华[1] 陈凯明[1] 

机构地区：[1]中国科学技术大学,计算机科学与技术系,安徽合肥230027

出　　处：《计算机仿真》2008年第8期338-341,共4页Computer Simulation

摘　　要：在反编译面向对象软件的过程中,反编译工具必须能够揭示出该软件的体系结构,因此反编译器必须能够识别出软件中的类和类中所包含的方法和属性以及类之间的关系。通过对编译器产生的面向对象软件汇编码进行细致深入的分析,归纳出了类函数的特征码,由此给出了一种能够对软件所包含类中类成员函数进行识别和归类的方法,其中运用了程序切片,模拟执行算法等技术并据此开发了一个实验系统,此系统显示方法平均能识别出一个类中约80%的类函数。In decompiling OOP software, the decompilation must be able to reveal the framework of this software, so decompiler has to be able to identify the classes, the functions and attributes of classes, as well the relations between classes. Through the careful and deep analysis of the assembly code of object - oriented software generated by compiler, the paper concludes the special codes of class method and gives a method by which classes＇ member functions in software can be identified and classified. In this method, program slicing and simulation execution algorithm techniques etc are used. By using these techniques, an experiment system is developed, and the result of experiment shows that this method can identify about 80% of member functions in one class.

关 键 词：类 反编译 面向对象 类函数 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]