static-rc

Rust's reference topology is too restrictive. You can't have back references. This is what drives many C++ programmers nuts. It's common in C++ to have A point to B, and for B to have a pointer back to A. This happens implicitly with class inheritance, too. As a result, common C++ idioms don't translate to Rust at all. This is fixable. Because you can have back references. You just have to use Rc, Rc::Weak, .upgrade(), RefCell, .borrow, and .borrow_mut(). This works, but only if the upgrades and borrows never fail. A failed .borrow() is a panic. The implication is that if you use .borrow() or .borrow_mut(), there's some good reason to think it will never fail. For Rc::Weak, the key constraint is that all weak pointers must drop before all strong pointers have dropped. If you can prove that, .upgrade() doesn't need a run-time check. For Re

Good to see some more uses of linear types! Very few languages can use linear types, and their potential is huge IMO. OP shows the benefits of a linear RC that tracks references dynamically, and one can even imagine a version where the compiler tracks the references instead. For example (for those who are more familiar with Rust) imagine that it had linear types and that we used them for matthieum's static-RC [0] (like suggested by frank-king at [1] to prevent leaks) and you'd have a pretty good idea of linear static RC. I also talked about a possible alternative linear static RC at [2] and [3] (I didn't explain it well at all, happy to elaborate), though lately I suspect that one would need to add a GhostToken-ish [4] substance to allow safely reading/writing the shared value, or perhaps a RefCell. I suspect that something like GhostToke