Zulip Chat Archive
Stream: lean4
Topic: prelude check fails when rebasing onto nightly-with-mathlib
Thomas Murrills (Feb 23 2024 at 23:36):
I'm just experimenting with a draft PR to try to fix a bug in lean4, and things work fine on master—but when I rebase onto nightly-with-mathlib
to test, "Check for modules that should use prelude
/ check-prelude (pull_request) " fails (listing a bunch of files I haven't touched). Does anyone know what this means/how to fix it?
Matthew Ballard (Feb 23 2024 at 23:42):
lean#3374 and lean#3463 is what I noticed
Thomas Murrills (Feb 23 2024 at 23:49):
Ah, hmm. What's the best solution here? I imagine nightly-with-mathlib
will catch up sooner or later, but it seems about a week behind...
Matthew Ballard (Feb 23 2024 at 23:51):
What is the reason you on nightly with mathlib to start?
Matthew Ballard (Feb 23 2024 at 23:51):
There is a clean up from zeta reduction changes in simp
ongoing that has broken nightly
for a bit
Thomas Murrills (Feb 23 2024 at 23:52):
I'm currently hitting ":exclamation: Std/Mathlib CI will not be attempted unless your PR branches off the nightly-with-mathlib branch", and I'd like to attempt Std/Mathlib CI.
Matthew Ballard (Feb 23 2024 at 23:53):
Yeah, I was just wondering if you really wanted std and mathlib. Since you do, I think you have to wait
Thomas Murrills (Feb 23 2024 at 23:56):
Hmm. I wonder if I could cherry-pick the result of lean#3463 into (onto?) my branch in the meantime.
Matthew Ballard (Feb 23 2024 at 23:58):
Or you could disable the CI temporarily :smiling_devil:
Matthew Ballard (Feb 23 2024 at 23:58):
But mathlib won’t build still
Thomas Murrills (Feb 23 2024 at 23:59):
Oh, earlier you meant that (I'm mixing my branches:nightly-with-testing
nightly-with-mathlib
) itself breaks mathlib?
Matthew Ballard (Feb 23 2024 at 23:59):
Yeah, sorry if that was unclear
Matthew Ballard (Feb 24 2024 at 00:01):
See the most recent CI run
Thomas Murrills (Feb 24 2024 at 00:02):
Ah gotcha, part of me was hoping you did literally mean (lean4) nightly
:P
Henrik Böving (Feb 24 2024 at 10:58):
Thomas Murrills said:
I'm just experimenting with a draft PR to try to fix a bug in lean4, and things work fine on master—but when I rebase onto
nightly-with-mathlib
to test, "Check for modules that should useprelude
/ check-prelude (pull_request) " fails (listing a bunch of files I haven't touched). Does anyone know what this means/how to fix it?
https://lean-lang.org/lean4/doc/dev/index.html#prelude this is the reasoning
Last updated: May 02 2025 at 03:31 UTC