Open shonfeder opened 10 months ago
Maybe check out this paper which was published at OOPSLA '23:
Getting into the Flow: Towards Better Type Error Messages for Constraint-Based Type Inference https://dl.acm.org/doi/10.1145/3622812 (PDF)
I don't know if the technique described there is directly applicable to the Quint type checker but perhaps it'll yield some inspiration towards a solution.
Thanks for the pointer, @romac !
Currently, the constraint system etc. doesn't let us produce very friendly or informative error messages: it's all in terms of unification failures of rows. We need a way to guide or decorate the constraint checker with error messages on constraint failures.