Jack Ek
Jack Ek
>Thanks for finding these bugs and giving easy instructions. I've had to track these down because my tax data contained discrepancies, so might as well provide you with all the...
>We're still missing a mechanism to prove that something doesn't change after deployment. Would it be completely crazy to allow all types of expressions as `invariants`, not only booleans? So...
How about a function which takes - a refined AST - an "address into the refined AST" (e.g. `[Int]`, where each element tells us which constructor argument we should recurse...
I'm not sure I understand. I'm thinking like this: 1. When working with a refined AST, it should be relatively easy to keep track of your current position in the...
Something like this? But with waaaaay more cases. Kind of unsatisfactory in that way tbh... ```haskell import Syntax import RefinedAst type ExpPos = [Int] getRefinedPos :: ExpPos -> Exp a...
But note that we don't need explicit cases for when `refined` and `simple` don't match. All of those can just be an error. So at least we don't need `n^2`...
Yeah agreed tbh
I guess one thing we could do is to create a `Map ExpPos Pn` at the same time as we refine the AST. So like ```haskell inferExpr :: Env ->...
Hey sorry, didn't see this. I haven't been maintaining Act for quite a few years now and so am not qualified to assess whether this is still desirable or not...
This looks very good! A few comments and thoughts: >“Regularity” and “homoiconicity” and not well known terms; update them to something more understandable We need to talk about the value...