Testing Method of Code Redundancy Simplification Based on Program Dependency Graph

Abstract

This paper presents a testing method of code redundancy simplification from the aspect of software static analysis. We study on the dependency relationships between the variables, branches and functions in source or intermediate code of the tested object by PDG (Program Dependence Graph). This method establishes an effective testing to discover and locate the redundant functional modules and the unreachable paths based on dependency relationship. Compared with the conventional code optimization which emphasizes the improvement of time efficiency, we compress the size of source code and object code, independent of the programming language the source program uses. Experimental results show that the source or object code size decreases by 1 to 3 percent approximately after our simplification, not only reducing the occupied space but also assuring the functional consistency.