Enhancing Incremental Dataflow Analysis in an IDE
Koko Harianto, Feng-Jian Wang, Mahmoud M. Abouzeid · 2024
Incremental dataflow analysis is a conventional technique adopted in syntax-directed editors, popularly used in Integrated Development Environments (IDEs). However, dataflow anomaly detection during program editing in IDEs remains a challenge due to the interaction with newer functionalities, such as AI-powered plugins like Copilot, which might insert and modify code lines directly. These interactions impact both the efficiency and effectiveness of dataflow anomaly detection processes. This paper introduces an enhanced incremental dataflow analysis approach in IDEs, focused on the analysis of dataflow within program structures. Our approach introduces a modified version of SP-tree, named SP-graph. This modification, involving the addition of supporting edges, facilitates data storage and enhances dataflow analysis incrementally. We also present a mechanism that streamlines dataflow anomaly detection within IDEs by identifying modifications that may impact dataflow, updating relevant data, and conducting immediate analysis. In addition, incremental concurrent anomaly detection in SP-graph shall be studied further. We demonstrate the effectiveness of our approach through a case study.