Zulip Chat Archive

Stream: general

Topic: aesop can prove FLT


Eric Wieser (Mar 19 2025 at 01:18):

import Mathlib

-- we don't mention this in the proof, so we can ignore it
variable {not_used : False}

omit not_used in
def ohno : PLift FermatLastTheorem :=
  by aesop

Eric Wieser (Mar 19 2025 at 01:18):

Should by have the "don't capture any more variables" semantics that theorem has?

Alex Meiburg (Mar 19 2025 at 01:25):

No need for Mathlib here:

-- we don't mention this in the proof, so we can ignore it
variable {not_used : False}

omit not_used

def ohno : PLift FermatLastTheorem := by
  exact? --suggests `exact not_used.elim`, which also works

This seems like omit doesn't do what it says on the box, it doesn't remove variables for defs.

Eric Wieser (Mar 19 2025 at 01:26):

You need mathlib for docs#FermatLastTheorem

Aaron Liu (Mar 19 2025 at 01:30):

I've not been able to get omit to do what I want it to do yet. Every time I use it to omit something, the something doesn't get omitted

Alex Meiburg (Mar 19 2025 at 01:43):

Eric Wieser said:

You need mathlib for docs#FermatLastTheorem

Yeah, I know it's autoimplicit, but that's kind of not the point right? The point is that we pull in this variable that 'should' be omitted

Sebastian Ullrich (Mar 19 2025 at 06:24):

That's a variant of lean#5595: omit currently only undoes include


Last updated: May 02 2025 at 03:31 UTC