Canonical Backs New Project to Translate Large C Codebases into Safe Rust

Canonical and the University of Bristol will launch a three-year research project in late 2026 to build a platform that automatically translates large, mature C codebases—potentially hundreds of thousands of lines—into safe, idiomatic Rust using a neurosymbolic approach that combines machine learning, program analysis, testing and formal methods. The workflow will split repositories into smaller units, translate them with language models trained on C-to-Rust examples, and verify functional equivalence, with Ubuntu’s AppArmor and snap-confine serving as real-world case studies.