CI: full battery in GitHub Actions - #21
Merged
Merged
Conversation
The g-meet regression (d731e5d, fixed in #17) survived 11 days on main because the landing claimed only the smoke suite green and the repo has no CI. This workflow runs the unmodified battery (./run_tests.sh) in a debian:trixie container - the same distribution and apt versions as the dev machine (SBCL 2.5.2, Agda 2.6.4.3, agda-stdlib 2.1, GHC 9.6.6). The battery treats a skipped check as a failure by design, so the job installs the full toolchain instead of weakening the battery. The oracle builds itself on a cold runner (MAlonzo via GHC); 45-minute cap covers the first compile. Fixes #20
arkh-node
force-pushed
the
ci/github-battery
branch
from
August 30, 2026 20:17
0ea5d77 to
3cfd03a
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Fixes #20.
What
.github/workflows/battery.yml— one job,debian:trixiecontainer (same distribution and apt versions as the dev machine), running the unmodified./run_tests.shon every push tomainand every PR.Why the container matches the dev machine
The battery's dependency contract is strict by design: missing agda/ghc is a failure ("пропущенная проверка не должна быть неотличима от пройденной"), the oracle builds itself via MAlonzo→GHC on a cold clone.
debian:trixiegives byte-identical toolchain versions (SBCL 2.5.2, Agda 2.6.4.3, agda-stdlib 2.1, GHC 9.6.6) — no version drift between local and CI.Agda stdlib registration is explicit (apt lands it in
/usr/share/agda-stdlib; the container HOME is empty), andLANG=C.UTF-8is set because the battery greps for Cyrillic markers.Proof
The PR's own battery run (this workflow) is the first live proof. Merge only after it is green; a green run on
mainafter merge verified via API.What this does NOT prove