SBIR Phase I: OpenRefactory/C: An Infrastructure to Provide Automated Power Tools for C Programmers

小型企业创新研究(SBIR)第一阶段项目的更广泛影响/商业潜力是为C程序开发人员提供正确且复杂的程序转换工具，他们不仅可以使用这些工具来提高C程序的质量，还可以使程序更安全地抵御众所周知的漏洞。尽管C很受欢迎，但它已经将开发环境(IDE)与有限的此类电动工具组合集成在一起，可伸缩性有限，对实际程序的适用性也有限。成功地解决现有工具中存在的限制并提供健壮的程序转换基础结构应该会从根本上改变开发人员处理C代码的方式，因为重构是其他编程语言的软件开发实践的组成部分。这将影响C程序的编写方式以及它们的维护和发展方式。这项拟议的工作还探索了可以修复C程序中安全漏洞的行为增强型程序转换。行为增强转换的工作应该开启一种新的强大工具的范例，可以提供给一般程序员(不限于C程序员)。这个小型企业创新研究(SBIR)第一阶段项目将促进为C语言实现可扩展、健壮的自动化程序转换工具所需的新算法和数据结构的开发。它将专注于这个问题的以前工作中没有得到充分解决的部分-具体地说，C预处理器的合理处理和深入的程序分析。同时，这项工作将解决C平台的异构性，例如，不同的方言、基于底层硬件的不同配置、不同的IDE等。在过去三年中，原理调查员领导了一项由NSF支持的研究工作，以解决研究问题，并开发了一个大型研究原型(超过500Kloc)。所提出的工作将许多研究成果付诸实践。

The broader impact/commercial potential of this Small Business Innovation Research (SBIR) Phase I project is about empowering C program developers with correct and complex program transformation tools that they can use to not only improve the quality of C programs but also make the programs more secure against well-known vulnerabilities. C, in spite of its popularity, has integrated development environments (IDEs) with a limited portfolio of such power tools, with limited scalability and limited applicability to real-world programs. Successfully addressing the limitations present in existing tools and providing a robust program transformation infrastructure should fundamentally change how developers work on C code since refactoring is an integral part of software development practices for other programming languages. This will impact both how C programs are written and how they are maintained and evolved. The proposed work also explored behavior-enhancing program transformations that can fix security vulnerabilities in C programs. The work on behavior-enhancing transformations should open a new paradigm of power tools that can be made available to programmers in general (not limited to C programmers).This Small Business Innovation Research (SBIR) Phase I project will foster the development of new algorithms and data structures needed to implement scalable, robust automated program transformation tools for C. It will focus on the parts of this problem that have been inadequately addressed in prior work - specifically, sound handling of the C preprocessor and deep program analysis. At the same time, the work will address the heterogeneity of C platforms, e.g., different dialects, different configurations based on underlying hardware, different IDEs, etc. The Principle Investigator led an NSF-supported research effort to address the research problems in the last three years and has developed a large research prototype (over 500 KLOC). The proposed work brings many of the research results to practice.