Wasabi
Wasabi is a toolkit designed to isolate and trigger retry bugs by combining static program analysis, large language models (LLMs), fault injection, and unit testing.
Install / Use
/learn @bastoica/WasabiREADME
This repository hosts WASABI, a toolkit for exposing and isolating bugs in retry logic (retry bugs) that surface in software systems. For design details, insights, and a comprehensive evaluation please refer to our paper [1].
Overview
WASABI implements two complementary bug-detection workflows: (1) a testing worfklow that triggers retry bugs using a combination of static analysis, large language models (LLMs), fault injection, and software testing; and (2) a static analysis workflow that identifies retry bugs using a combination of static control flow analysis and LLMs.
To get started, users can refer to:
- The step-by-step guide to build, install and run WASABI's testing workflow;
- The instructions to invoke WASABI's static analysis worfklow.
Users can also navigate to the sosp24-ae branch which contains guidelines and automation to replicate the key results from our paper [1].
References
[1] "If At First You Don't Succeed, Try, Try, Again...? Insights and LLM-informed Tooling for Detecting Retry Bugs in Software Systems". Bogdan Alexandru Stoica*, Utsav Sethi*, Yiming Su, Cyrus Zhou, Shan Lu, Jonathan Mace, Madan Musuvathi, Suman Nath (*equal contribution). The 30th Symposium on Operating Systems Principles (SOSP). Austin, TX, USA. November, 2024.
