I'm really interested in somebody pursuing https://github.com/rust-lang/rust/issues/50753 -- anybody interested? I can try to leave some notes on the issue and advertise it a bit.
@mikeyhew, you would be the obvious candidate, if you have time.
@nikomatsakis can it be done? I tried using predicate_must_hold and predicate_may_hold, and neither worked
I think predicate_must_hold would be closer. The issue it ran into was that sometimes closures didn't coerce into FnOnce
I'm not entirely sure, but something like it must be done eventually if we are to transition to chalk :)
I didn't entirely expect it to be easy though, it may be that we have to modify chalk in some way to add a "custom mode" for this sort of evaluation
would be a shame though