Retargeting an Abstract Interpreter for a New Language by Partial Evaluation

TL;DR

This paper introduces a novel method using partial evaluation to automatically adapt an existing abstract interpreter to new programming languages, reducing manual effort and increasing efficiency.

Contribution

The paper presents a new technique that leverages partial evaluation to automatically retarget abstract interpreters for different languages, streamlining static analyzer development.

Findings

Successfully retargeted an abstract interpreter to a new language

Eliminated manual development of analyzers for target languages

Ensured correctness of the retargeted analyzers

Abstract

It is well-known that abstract interpreters can be systematically derived from their concrete counterparts using a "recipe," but developing sound static analyzers remains a time-consuming task. Reducing the effort required and mechanizing the process of developing analyzers continues to be a significant challenge. Is it possible to automatically retarget an existing abstract interpreter for a new language? We propose a novel technique to automatically derive abstract interpreters for various languages from an existing abstract interpreter. By leveraging partial evaluation, we specialize an abstract interpreter for a source language. The specialization is performed using the semantics of target languages written in the source language. Our approach eliminates the need to develop analyzers for new targets from scratch. We show that our method can effectively retarget an abstract interpreter for one language into a correct analyzer for another language.