ARC++: Effective typestate and lifetime dependency analysis

Abstract

The ever-increasing reliance of today's society on software requires scalable and precise techniques for checking the correctness, reliability, and robustness of software. Object-oriented languages have been used extensively to build largescale systems, including Java and C++. While many scalable static analysis approaches for C and Java have been proposed, there has been comparatively little work on the static analysis of C++ programs. In this paper, we provide an abstract representation to model C++ objects, containers, references, raw pointers, and smart pointers. Further, we present a new analysis called lifetime dependency analysis, which allows us to precisely track the complex lifetime semantics of temporary objects in C++. Finally, we propose an implementation of our techniques and present promising results on a large variety of open-source software.