Turning `abbrev` into `def` defeats type inference

Prerequisites

Please put an X between the brackets as you perform the following steps:

• Check that your issue is not already filed:
  https://github.com/leanprover/lean4/issues
• Reduce the issue to a minimal, self-contained, reproducible test case.
  Avoid dependencies to Mathlib or Batteries.
• Test your test case against the latest nightly release, for example on
  https://live.lean-lang.org/#project=lean-nightly
  (You can also use the settings there to switch to “Lean nightly”)

Description

Turning an abbrev into a def may regress elaboration.

Context

The definition of SPred below will be upstreamed into Core soon (#8745); it would be better if it could be a def.

Steps to Reproduce

Elaborate

def SPred (σs : List Type) := match σs with
  | [] => Prop
  | σ :: σs => σ → SPred σs

class WP (m : Type → Type) (σs : outParam (List Type)) where

instance : WP (EStateM ε σ) [σ] where

def Triple [WP m σs] (x : m α) (P Q : SPred σs) := True

theorem test {prog : EStateM ε σ α} (P : σ → Prop) :
    Triple prog (fun s' => s' = s) P → P s := sorry

Expected behavior: Successfully elaborate. This is the behavior if I turn def SPred into abbrev SPred.

Actual behavior: Fails to elaborate (fun s' => s' = s).

Versions

Lean 4.21.0-rc3

Impact

Add 👍 to issues you consider important. If others are impacted by this issue, please ask them to add 👍 to it.

Start by running the minimal Lean 4.21.0-rc3 reproduction and compare elaboration with SPred declared as def versus abbrev. Trace the failed elaboration of (fun s' => s' = s); done means the def version elaborates successfully without changing the reported example.