Repository navigation
ci: drop judge-vision and run the restart journey nightly only - #6542
Merged
Merged
Conversation
judge-vision waited for every proof job before judging screenshot claims and running the advisory design review, and thread-continues-after-restart (~8.5 min) was the longest proof. Together they set PR time to green. The restart journey still runs nightly and on PRs that change it.
Contributor
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
benjaminshafii
enabled auto-merge (squash)
October 27, 2026 04:55
Contributor
benjaminshafii
disabled auto-merge
October 27, 2026 04:56
This branch was successfully deployed
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
What is this about?
Cuts the two longest pieces off PR change proof: the
judge-visionjob is removed, andthread-continues-after-restartno longer runs on every PR.What problem does it solve?
"Every check green" takes ~10.7 min. Its critical path is
thread-continues-after-restart(~8.5 min, it relaunches the desktop app twice), thenjudge-vision, which can only start after every proof job and adds 1-2 min.What was the situation before?
Every non-docs PR ran all core journeys including the restart journey, then judged screenshot claims and ran the advisory design review in
judge-vision.How is this implemented?
.github/workflows/pr-proof.yml: thejudge-visionjob is gone. Screenshot claims in changed specs stay pending in the record; no core journey on the PR path has any..github/scripts/pr-proof.mjs:NIGHTLY_ONLYkeepsthread-continues-after-restartout of the automatic PR set. The nightly Core journeys workflow still runs it, and a PR that changes the spec still runs it. Covered bypr-proof-selection.test.mjs.What else changed?
pr-proof-artifacts.test.mjsand the workflow assertion inadvisory-design-cap.test.ts;pr-proof-review-build.test.mjsnow splits at the next job.evals/design-review/README.mdand thefix-design-notesskill say design review now runs locally.Release note
none
Evidence
No E2E; the Unit tests step that runs the PR-proof script tests passes 19/19,
pnpm evals:checkis clean. This PR's own PR change proof run shows the shorter path.