I think that's interesting

If that works as a drop-in replacement for a significant amount of normal C applications, that would be a low-effort, instant fix for tons of security vulnerabilities. There are a lot of legacy project that work perfectly fine, but where we have no real-world resources to recreate them in a safe language – and even if applications need some patches to compile with, that is much more feasible to sell to the maintainers than "rewrite it in rust".

What would be really goat is if there'd be an ABI translation layer for Fil-C-ABI to C-ABI (not sure if that's feasible though). That way I could compile not-performance-critical legacy library in Fil-C and tie it to Rust/Golang/Swift or other legacy-C. Of course that would invalidate safety guarantees at external ABI boundaries; but those are easier to audit/prove than library internals...

Overall, that sounds like a really tasty project; it'll be very interesting to see how it cooks, and how it works out IRL...