Reproduce “Parser Knows Best: Testing DBMS with Coverage-Guided Grammar-Rule Traversal” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Yu Liang, Hong Hu

Penn State University researchers developed ParserFuzz, a novel fuzzer that automatically extracts SQL grammar rules from DBMS parser definition files and integrates coverage-guided mutation. This approach enabled the discovery of 81 previously unknown bugs, including 29 segmentation faults, across five widely used DBMSs, while also achieving higher grammar and code coverage compared to existing testing tools.

Key insights

  • Directly leveraging a DBMS's internal parser grammar files ("ground truth SQL grammar rules") ensures comprehensive exploration of supported syntax features, including obscure ones.
  • Combining grammar-rule-based query generation with coverage-guided mutation is essential for intelligently exploring complex feature combinations and discovering deep memory errors.
  • Mitigating path explosion in recursive grammar rules through rule classification (simple, normal, complex) and prioritization significantly improves the success rate of generating valid SQL queries.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.