Closed proux01 closed 1 year ago
I guess the
ring
andfield
tactics have the same issue.
I don't know how it's handled there but all my tests with ring
did work. The issue doesn't seem to be there.
This would be a good occasion to think about supporting let-ins in general, using the PHOAS representation presented in Reification by Parametricity, although it would be a bit tricky in combination with homomorphism applications.
Sure, but given how annoying is the bug I'd first merge this as a quick fix.
A type cast
t : T
is represented as a let-inlet x : T := t in x
in Coq-Elpi. I guess thering
andfield
tactics have the same issue.This would be a good occasion to think about supporting let-ins in general, using the PHOAS representation presented in Reification by Parametricity, although it would be a bit tricky in combination with homomorphism applications, e.g., considering
let x := t in f x + g x
where bothf
andg
are homomorphisms.