Regression in inferring universe levels between v4.9.1 and v4.10.0-rc1

[mik-jozef]

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

The following code typechecks with Lean 4.9.1, but not with 4.10.1-rc1 and later:

structure Asdf: Type u where

def foo (T: Type u) : Asdf.{u} := {}

def bar
  (T: Type u)
  (holdsForAll: ∀ n: Asdf, True)
:
  True
:=
  let n := foo T
  
  holdsForAll n

The error:

application type mismatch
  holdsForAll n
argument
  n
has type
  Asdf : Type u
but is expected to have type
  Asdf : Type u_1

An easy workaround is to propagate the universe explicitly:

(holdsForAll: ∀ n: Asdf.{u}, True)

Expected behavior: The code produces no errors.

Versions

Lean versions affected: 4.10.0-rc1, lean nightly as of wriging
OS: Ubuntu 22.04.4 LTS

Impact

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

[Kha]

This is an intentional change: #4482