Build System Analysis with Link Prediction

• Main Authors: XIA, Xin, LO, David, WANG, Xinyu, ZHOU, Bo
• Format: text
• Language: English
• Published: Institutional Knowledge at Singapore Management University 2014
• Subjects: Software Engineering
• Online Access: https://ink.library.smu.edu.sg/sis_research/2035; http://dx.doi.org/10.1145/2554850.2555134
• Institution: Singapore Management University

Compilation is an important step in building working software system. To compile large systems, typically build systems, such as make, are used. In this paper, we investigate a new research problem for build configuration file (e.g., Makefile) analysis: how to predict missed dependencies in a build configuration file. We refer to this problem as dependency mining. Based on a Makefile, we build a dependency graph capturing various relationships defined in the Makefile. By representing a Makefile as a dependency graph, we map the dependency mining problem to a link prediction problem, and leverage 9 state-of-the-art link prediction algorithms to solve it. We collected Makefiles from 7 open source projects to evaluate the effectiveness of the algorithms.