Rewriting SQLite in Rust: The Limbo Project

00:00 Introduction and Sponsorship

A discussion about the SQLite Rewrite in Rust begins with the speaker clarifying their sponsorship while expressing genuine interest in the topic due to its popularity on social media. The speaker shares past efforts to encourage the project leaders to explore Zig before committing to Rust.

01:30 Forking SQLite

The team forked SQLite two years ago, seeking an open development model. This decision stemmed from dissatisfaction with SQLite's closed testing and contribution processes. The new project, Lib Squeal, has gained significant traction, illustrating a successful community-driven initiative.

03:00 The Limbo Project

The Limbo project aims to completely rewrite SQLite in Rust, striving for a memory-safe alternative while maintaining compatibility. The speaker discusses the challenges of forking and the benefits of building a new codebase from scratch, particularly in achieving innovative features like asynchronous I/O.

05:30 Testing and Reliability

Efforts to ensure reliability through deterministic simulation testing are highlighted, along with partnerships with Antithesis for enhanced testing methodologies. The use of this innovative testing approach aims to surpass SQLite's reliability.

08:00 Performance Comparisons

Initial performance tests suggest Limbo is already outperforming SQLite in certain queries, though the speaker emphasizes the need for comprehensive stress testing to validate these results. The focus is on ensuring that Limbo handles real-world usage scenarios effectively.

10:00 Conclusion and Invitation

The project remains in its early stages, but the team invites others interested in building an enhanced embedded database to contribute. The speaker expresses excitement about the possibilities that Limbo offers beyond just a simple rewrite.