Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Watch this 15-minute conference presentation from OOPSLA 2025 that introduces Moon, a novel selective object-sensitive pointer analysis framework designed to balance precision and efficiency in static program analysis. Learn about the theoretical foundation that establishes a sound over-approximation for identifying objects that truly benefit from context sensitivity, reformulating this identification as graph reachability problems over Pointer Flow Graphs (PFG). Discover how the researchers from Nanjing University developed a Variable Flow Graph (VFG) optimization that enables efficient backward and forward traversal to systematically identify precision-improving objects. Explore the experimental results demonstrating Moon's impressive performance gains of 37.2X speedup for 2-object-sensitive analysis and 382.0X speedup for 3-object-sensitive analysis across 30 Java programs, while maintaining negligible precision losses of only 0.1% and 0.2% respectively. Understand how this approach addresses the longstanding challenge in pointer analysis where context sensitivity improves precision but at significant computational cost, and see how Moon's theoretically-backed methodology outperforms previous selective context-sensitive analysis approaches that relied on specific code patterns rather than comprehensive theoretical foundations.
