test(root): recover the clock-bound commit stranded by #867

[mobeenabdullah]

#867 merged without its final commit, and main carries an assertion that already failed CI.

What happened

git log <headRefOid>..<ls-remote tip> names one commit absent from the merge:

be8a0a0ba test(root): bound the rewound clock rather than pinning it

Confirmed by CONTENT, with a control proving the search works: toBeLessThan(-29000) is present on the branch head and absent from merge commit 88b44fce8. main still reads:

expect(applied?.currentTime).toBe(-30000);

Why that matters rather than being cosmetic

That exact-equality assertion is the one that failed Browser tests on #867's own CI:

Expected: -30000
Received: -29983.334

The animation's clock advances between the assignment and the read — one frame, 16.7ms. It held locally and lost that frame on the runner. So main currently carries a canvas test that fails whenever the round trip costs a frame, which is most of the time on a loaded machine.

The recovered commit replaces it with a bound. What makes the fixture the case it claims to be is that the effect is rewound far into the past — stable. The exact value is not, and asserting it measures how long the round trip took rather than the rewind.

Verification

• Cherry-picked onto main, one file, +6/-1.
• Guard suite: 32 passed.
• The same stranded-tail check that found this reports the recovered branch clean.

This is the second tail this lane has stranded — #825 did it too, recovered as #854 — and both times the cause was the same: pushing a fix while a merge was being computed. Worth noting on the merge procedure rather than only in a PR body.

Summary by CodeRabbit

• Bug Fixes
  • Improved animation rewind test reliability by allowing for minor timing differences when validating the current playback position.

[mobeenabdullah]

@codex please review this PR

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro Plus

Run ID: 22813293-b1b5-4fbd-8091-c5f2ebfad2c1

📥 Commits

Reviewing files that changed from the base of the PR and between 88b44fc and 03b0639.

📒 Files selected for processing (1)

• e2e/tests/canvas/geometry-settle-matches-the-canvas.test.ts

Included review availability: Your plan includes up to 1 review per rolling hour; 0 remain after this review.

---

📝 Walkthrough

Walkthrough

The canvas animation test now checks that a rewound animation remains below -29000 instead of requiring an exact -30000 value.

Changes

Canvas animation tests

Layer / File(s)	Summary
Relax rewound time assertion e2e/tests/canvas/geometry-settle-matches-the-canvas.test.ts	The test accepts elapsed time during rewinding by asserting currentTime < -29000.

Estimated code review effort: 1 (Trivial) | ~2 minutes

Merge Risk: ⚪ Minimal · up to 03b06

The change makes the canvas test resilient to the animation clock advancing between writes and reads without changing product behavior; no actionable merge-blocking risk remains after normal checks and review.

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly identifies the recovered clock-bound test commit and references the related pull request.
Description check	✅ Passed	The description explains the issue, rationale, affected assertion, related pull request, and verification, but it does not follow the template headings or checklist format.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch fix/recover-rewound-clock-bound

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[pkg-pr-new]

Open in StackBlitz

@nextlyhq/adapter-drizzle

npm i https://pkg.pr.new/@nextlyhq/adapter-drizzle@03b0639

@nextlyhq/adapter-mysql

npm i https://pkg.pr.new/@nextlyhq/adapter-mysql@03b0639

@nextlyhq/adapter-postgres

npm i https://pkg.pr.new/@nextlyhq/adapter-postgres@03b0639

@nextlyhq/adapter-sqlite

npm i https://pkg.pr.new/@nextlyhq/adapter-sqlite@03b0639

@nextlyhq/admin

npm i https://pkg.pr.new/@nextlyhq/admin@03b0639

@nextlyhq/admin-css

npm i https://pkg.pr.new/@nextlyhq/admin-css@03b0639

@nextlyhq/blocks-engine

npm i https://pkg.pr.new/@nextlyhq/blocks-engine@03b0639

@nextlyhq/blocks-react

npm i https://pkg.pr.new/@nextlyhq/blocks-react@03b0639

@nextlyhq/builder

npm i https://pkg.pr.new/@nextlyhq/builder@03b0639

create-nextly-app

npm i https://pkg.pr.new/create-nextly-app@03b0639

nextly

npm i https://pkg.pr.new/nextly@03b0639

@nextlyhq/plugin-form-builder

npm i https://pkg.pr.new/@nextlyhq/plugin-form-builder@03b0639

@nextlyhq/plugin-page-builder

npm i https://pkg.pr.new/@nextlyhq/plugin-page-builder@03b0639

@nextlyhq/plugin-sdk

npm i https://pkg.pr.new/@nextlyhq/plugin-sdk@03b0639

@nextlyhq/plugin-seo

npm i https://pkg.pr.new/@nextlyhq/plugin-seo@03b0639

@nextlyhq/storage-s3

npm i https://pkg.pr.new/@nextlyhq/storage-s3@03b0639

@nextlyhq/storage-uploadthing

npm i https://pkg.pr.new/@nextlyhq/storage-uploadthing@03b0639

@nextlyhq/storage-vercel-blob

npm i https://pkg.pr.new/@nextlyhq/storage-vercel-blob@03b0639

@nextlyhq/ui

npm i https://pkg.pr.new/@nextlyhq/ui@03b0639

commit: 03b0639

[chatgpt-codex-connector]

Codex Review: Didn't find any major issues. Can't wait for the next one!

Reviewed commit: 03b0639ae8

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[mobeenabdullah]

@nextly-bot review

[github-actions]

Claude finished @mobeenabdullah's task in 18m 13s —— View job

---

Review complete — round 1, no new findings

Ran the full review protocol on 03b0639 and posted the result as a PR review (pullrequestreview-4946514254).

• Phase 0 — pre-flight (PR open, not draft, head 03b0639ae8, base main @ 88b44fce8)
• Phase 1 — load the law (AGENTS.md, comment-convention checker, pr-title.yml allowlist)
• Phase 2 — round awareness (no prior bot rounds, zero review threads)
• Phase 3–4 — understand task, expand context (full enclosing file, sibling tests, poc-driver)
• Phase 5 — the hunt
• Phase 6 — adversarial verification
• Phase 7–8 — compose and post the review

Verdict: clean. The recovery claim checks out content-wise — git diff be8a0a0ba..03b0639ae on the changed file is empty, and be8a0a0ba is absent from origin/main — so this PR restores exactly what #867's merge stranded.

The substantive question I attacked was whether the new bound weakens the fixture, since the test exists to prove the live-span comparison catches an API-moved schedule. It doesn't: at any currentTime < -29000 the live span is ≥49000ms against a declared 20000ms, still orders of magnitude past the 1ms tolerance, so the /running effect claims/ refusal still fires. In the other direction, the bound buys ~1000ms of clock slack against a ~17ms round trip, and the assertion still fails if the rewind never applied (currentTime would coerce to 0).

Process sweep is also clean: title/scope valid, no changeset (correct for a test-only PR), comment passes the repo's comment-convention gate, no AI attribution, all CI green including Browser tests.

[github-actions]

Nextly Review Bot: round 1 - clean

Verdict: The PR does exactly what it claims — it restores, byte-for-byte, the commit that #867's merge stranded — and in doing so replaces the only exact-equality clock assertion in the file, one main's own Browser tests job had already failed on a single frame of drift. Safe to merge.

New findings: 0 P0/P1, 0 P2, 0 P3

Prior rounds: none — this is round 1. Zero review threads exist on the PR and no earlier pr-review-agent marker was posted, so there is nothing to re-verify or duplicate.

Not inline-anchorable: none

Pre-existing (not this PR): none noted. The defect this PR fixes (main carrying toBe(-30000) at 88b44fce8) is the thing being repaired, not a bystander.

Checked:

• Content-verified the recovery claim. git diff be8a0a0ba..03b0639ae -- e2e/tests/canvas/geometry-settle-matches-the-canvas.test.ts is empty (identical blobs, 833ebe273), and git merge-base --is-ancestor be8a0a0ba origin/main fails — the stranded commit is genuinely absent from main and this PR restores it exactly. The PR's +6/-1 matches the diff stat, and its "32 passed" matches the 32 top-level tests in the file.
• Attacked the fixture's purpose (the bound must not weaken the test). Traced geometrySpanMs to its live-span formula at lines 406-421: min(delay + activeDuration, endTime) - currentTime. At any currentTime < -29000 the live span is ≥49000ms against a declared 20000ms, still far past the 1ms TIMING_TOLERANCE_MS at line 611, so the /running effect claims/ refusal at line 1566 still fires. The bound cannot mask the mechanism it exists to exercise.
• Checked the false-failure direction. The one-sided bound buys ~1000ms of clock slack against a ~17ms assignment-to-read round trip (a 60x margin), a strict widening from the previous 0ms slack. animation.ready would have to stall for a full second to fail — and the assertion still fails if the rewind never applied at all, since Number(animation.currentTime ?? 0) coerces an unapplied rewind to 0, which is not < -29000.
• Process sweep. Title test(root): ... is a valid type and a scope on the pr-title.yml allowlist, lowercase, no trailing period. A test-only PR correctly carries no changeset (git diff origin/main HEAD -- .changeset is empty). The new comment describes the code's timing behavior rather than a conversation and passes the repo's check:comments gate (CI job green). No AI attribution in the commit or PR body. All 7 workflow runs on this head are green, including Browser tests.

An empty round here is the merge signal for this PR: the change is one assertion and its rationale, both of which hold up.