remove via_vm_compute

theories/lithium/infrastructure.v

@@ -47,19 +47,3 @@ Class CanSolve (P : Prop) : Prop := can_solve: P.
 Definition shelve_hint (P : Prop) : Prop := P.
 Typeclasses Opaque shelve_hint.
 Arguments shelve_hint : simpl never.
-
-(** * [via_vm_compute]
-  [via_vm_compute f a] is equivalent to [f a] but tells the automation that it should
-  be reduced using [vm_compute]
- *)
-Definition via_vm_compute_def {A B} (f : A → B) (x : A) : B :=
-  f x.
-Definition via_vm_compute_aux : seal (@via_vm_compute_def). by eexists. Qed.
-Definition via_vm_compute := unseal via_vm_compute_aux.
-Definition via_vm_compute_eq' : @via_vm_compute = @via_vm_compute_def :=
-  seal_eq via_vm_compute_aux.
-Arguments via_vm_compute {_ _} _ _.
-
-Lemma via_vm_compute_eq {A B} (f : A → B) x:
-  via_vm_compute f x = f x.
-Proof. rewrite via_vm_compute_eq'. done. Qed.

theories/lithium/interpreter.v

@@ -704,15 +704,6 @@ Ltac liSideCond :=
     lazymatch P with
     | shelve_hint _ => split; [ unfold shelve_hint; li_shelve_sidecond |]
     | _ => first [
-      match P with
-      | context C [via_vm_compute ?f ?a] =>
-        rewrite (via_vm_compute_eq f a);
-        (* TODO: using [vm_compute (f a)] leads to very slow QED. Why is this the case? *)
-        (* vm_compute (f a) *)
-        let x := eval vm_compute in (f a) in
-        let P' := context C [ x ] in
-        change_no_check (P' ∧ Q)
-      end |
       progress normalize_goal_and |
     lazymatch P with
     | context [protected _] => first [