Explore a 17-minute video presentation from the SOAP 2024 conference that introduces a more precise approximation for interleaved Dyck reachability. Delve into the challenges of context- and field-sensitive static analysis, and discover how the presenters extend the mutual-refinement algorithm to achieve higher precision. Learn about the development of refined CFLs for expressing each type of sensitivity and the application of on-demand analysis to mask out irrelevant graph parts. Examine the experimental results showing significant improvements over existing approaches, with a focus on a challenging benchmark where the new method achieves 51% reduction in reachable pairs compared to recent alternatives.
Overview
Syllabus
[SOAP24] A Better Approximation for Interleaved Dyck Reachability
Taught by
ACM SIGPLAN