Seems like the original solution was automaticly ported to a new GHC version, and that broke it. The code compiles fine with NoPolyKinds language extension set, so i propose enabling it

My solution passes sample tests, but fails to compile on submission

test/ImperativeSpec.hs:216:1: error:
    • Uninferrable type variable k10 in
      the type synonym right-hand side:
      forall v st (s :: k10).
      HasValue st v b =>
      Var st a -> v -> Imperative @{k10} @{k10} st s s ()
    • In the type declaration for ‘Op’
    |
216 | type Op a b = forall v st s. HasValue st v b => Var st a -> v -> Imperative st s s ()
    | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

Idk, it looks like an internal problem

Haskell GHC 9.2.5

tan related tests should accept answers expressed with cos

simple expressions

f(x) = (tan x)
f'(x) = (/ 1 (^ (cos x) 2))

 actual:   `"(/ 1 (^ (cos x) 2))"`,
 expected: `"(+ 1 (^ (tan x) 2))"`

Nested expressions

f(x) = (tan (* 2 x))
f'(x) = (* 2 (/ 1 (^ (cos (* 2 x)) 2)))

  actual:   `"(* 2 (/ 1 (^ (cos (* 2 x)) 2)))"`,
  expected: `"(* 2 (+ 1 (^ (tan (* 2 x)) 2)))"`

now it works, thank you!

Scala tests are flawed, identical c# code works just fine