Customizing Static Analysis using Codesearch

TL;DR

This paper introduces StarLang, a simplified and efficient variant of Datalog, to facilitate easier development of customizable static analysis tools for security and verification tasks.

Contribution

The paper presents StarLang, a new language that simplifies and accelerates the creation of custom static analysis tools by reducing complexity and runtime overhead.

Findings

StarLang enables faster static analysis rule implementation.

It maintains a balance between expressiveness and performance.

Developers can build custom tools more easily and efficiently.

Abstract

Static analysis is a growing application of software engineering, leading to a range of essential security tools, bug-finding tools, as well as software verification. Recent years show an increase of universal static analysis tools that validate a range of properties and allow customizing parts of the scanner to validate additional properties or "static analysis rules". A commonly used language to describe a range of static analysis applications is Datalog. Unfortunately, the language is still non-trivial to use, leading to analysis that is difficult to implement in a precise but performant way. In this work, we aim to make building custom static analysis tools much easier for developers, while at the same time, providing a familiar framework for application security and static analysis experts. Our approach introduces a language called StarLang, a variant of Datalog which only includes programs with a fast runtime by the means of having low time complexity of its decision procedure.