Zulip Chat Archive
Stream: general
Topic: Error formatting
Yaël Dillies (May 17 2025 at 18:34):
I believe this has been discussed before, but I can't find the relevant discussions: Can we make rid lake's output of the irrelevant parts of the path to files?
Yaël Dillies (May 17 2025 at 18:34):
For example, this job prints
Warning: /home/lean/actions-runner/_work/mathlib4/mathlib4/Mathlib/Algebra/GroupWithZero/Int.lean:7:1: warning: import #[Mathlib.Algebra.Group.Int.Defs] instead
/home/lean/actions-runner/_work/mathlib4/mathlib4/Mathlib/Algebra/GroupWithZero/Int.lean:
remove #[Mathlib.Algebra.Group.Int.TypeTags]
add #[Mathlib.Algebra.Group.Int.Defs]
/home/lean/actions-runner/_work/mathlib4/mathlib4/Mathlib/Algebra/Order/Ring/IsNonarchimedean.lean:
remove #[Mathlib.Algebra.GroupWithZero.Action.Defs]
/home/lean/actions-runner/_work/mathlib4/mathlib4/Mathlib/Algebra/Tropical/Basic.lean:
remove #[Mathlib.Algebra.GroupWithZero.Action.Defs]
/home/lean/actions-runner/_work/mathlib4/mathlib4/Mathlib/RingTheory/Valuation/Discrete/Basic.lean:
remove #[Mathlib.Algebra.GroupWithZero.Int]
Yaël Dillies (May 17 2025 at 18:36):
There are several things that are wrong with it:
-
The first error is on the same line as
Warning:, meaning it shows differently to the other ones, ie
like this -
The repeated
/home/lean/actions-runner/_work/mathlib4/mathlib4/drown the important information
Yaël Dillies (May 17 2025 at 18:38):
It's possible part of this is an issue with https://github.com/leanprover-community/gh-problem-matcher-wrap
Eric Wieser (May 17 2025 at 19:18):
It looks like the --gh-style flag isn't doing as well as it could
Eric Wieser (May 17 2025 at 19:22):
Which confusingly, seems to currently mean --lean-style, which is then handled by the matcher you link
Eric Wieser (May 17 2025 at 19:22):
As opposed to directly using the actual github style
Kim Morrison (May 17 2025 at 21:50):
@Mac Malone is looking into fixing the absolute paths, if I understand correctly. Hopefully this will be better in the next release.
Last updated: Dec 20 2025 at 21:32 UTC