Here the verus github page with some code samples in color:
Very interesting. I wonder what its limits are. Will have to give it a try.
This project is still new, but the theory is well-examined, and there are widely accepted (but more difficult to use) systems like TLA+.
There are a number of comments on lobste.rs on this. Especially David Chisnall, which participates in the C++ standard meetings, and is an researcher and expert on verification, (link to Google scholar) has a high opinion on it.
That said, comprehensive correctness proofs for software are hard. They are usually a lot of work. But this one could be very useful for unsafe Rust.
In a past life I worked on a state visualiser for TLA+, but I barely understood it then, and I didnt think it went anywhere.
I more meant what its limits are in terms how much code complexity/recursion/etc. I’ll have a play with it and see if I can hit any.

