This PR adds a case to `Level.geq` that is present in the kernel's level
`is_geq` procedure, making them consistent with one another.
This came up during testing of `lean4lean`. Currently `Level.geq`
differs from `level::is_geq` in the case of `max u v >= imax u v`. The
elaborator function is overly pessimistic and yields `false` on this
while the kernel function yields true. This comes up concretely in the
`Trans` class:
```lean
class Trans (r : α → β → Sort u) (s : β → γ → Sort v) (t : outParam (α → γ → Sort w)) where
trans : r a b → s b c → t a c
```
The type of this class is `Sort (max (max (max (max (max (max 1 u) u_1)
u_2) u_3) v) w)` (where `u_1 u_2 u_3` are the levels of `α β γ`), but if
you try writing that type explicitly then the `class` command fails.
Omitting the type leaves the `class` to infer the universe level (the
command assumes the level is correct, and the kernel agrees it is), but
including the type then the elaborator checks the level inequality with
`Level.geq` and fails.
---------
Co-authored-by: Kyle Miller <kmill31415@gmail.com>
Many of our tests in `tests/lean/run/` produce output from `#eval` (or
`#check`) statements, that is then ignored.
This PR tries to capture all the useful output using `#guard_msgs`. I've
only done a cursory check that the output is still sane --- there is a
chance that some "unchecked" tests have already accumulated regressions
and this just cements them!
In the other direction, I did identify two rotten tests:
* a minor one in `setStructInstNotation.lean`, where a comment says `Set
Nat`, but `#check` actually prints `?_`. Weird?
* `CompilerProbe.lean` is generating empty output, apparently indicating
that something is broken, but I don't know the signficance of this file.
In any case, I'll ask about these elsewhere.
(This started by noticing that a recent `grind` test file had an
untested `trace_state`, and then got carried away.)
