test: ds-shots verification — DO NOT MERGE (TASK-22044)

[kushagrasarathe]

Throwaway verification PR for the ds-shots changes merged in #2897 and #2907. It nudges the settings-language screen 8px so exactly one fixture moves.

Expected on this PR's Tests run:

1. ds-shots completes in ~5–7 min (2-width capture, warm Next cache via restore-keys)
2. Capture runs only projects 320 and 430 (~60 shots)
3. Artifacts: visual-diff-report-<pr#>-1 (small, baseline SHA stamped) and visual-diff-<pr#>-1 (PNGs)
4. report.json shows settings-language@320/430 changed, everything else identical, no removed-width noise
5. Running scripts/ds-shots-publish.mjs against the report renders the sticky comment the publisher will post once the workflow reaches main (workflow_run fires from the default branch only — the comment itself CANNOT appear on this PR yet, that is the documented deploy gotcha)

This PR is never merged. It gets closed after the run is inspected and the branch deleted.

Screenshots: N/A (throwaway test change).

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated
peanut-wallet	Ready	Preview	Sep 1, 2026 4:11pm UTC

[notion-workspace]

ds-shots: publish the visual-diff result as a PR comment (workflow_run publisher)

[github-actions]

Code-analysis diff

Painscore total: 7153.05 → 7153.15 (+0.1)
Findings: 0 net (+0 new, -0 resolved)

[github-actions]

🧪 UI test report — ✅ all green

Suites

• ✅ unit: 5019 ran, 0 failed, 0 skipped, 1.6m

📊 Coverage (unit)

metric	%
statements	73.5%
branches	58.9%
functions	64.7%
lines	74.5%

⏱ 10 slowest test cases

time	test
4.3s	src/components/Card/share-asset/__tests__/shareAssetLayout.test.ts › never places two stickers in heavy overlap (broad seed sweep)
2.1s	src/components/Profile/views/__tests__/ResidenceChangeModal.test.tsx › swaps from the server value with no device mirror at all
1.5s	src/components/Profile/views/__tests__/ResidenceChangeModal.test.tsx › the change cooldown shows its date and blocks changing to another country, not re-saving
1.5s	src/components/Profile/views/__tests__/ResidenceChangeModal.test.tsx › saves the declared residence, refetches, and closes
1.4s	src/components/Profile/views/__tests__/ResidenceChangeModal.test.tsx › prefers the server value over a stale device mirror
1.2s	src/components/Profile/views/__tests__/ResidenceChangeModal.test.tsx › moving to a country in neither slot leaves the second document alone
1.0s	src/components/Profile/views/__tests__/ResidenceChangeModal.test.tsx › promoting the second document country swaps the pair instead of dropping one
1.0s	src/hooks/query/__tests__/user.test.tsx › does NOT clear a token that rotated mid-request (stale 401 racing a fresh login)
1.0s	src/components/Setup/Views/__tests__/Residence.test.tsx › reveals the second selector via the multi-doc link
1.0s	src/components/Setup/Views/__tests__/Residence.test.tsx › returns to the selector from the heads-up

_{📍 Inline annotations are in the **Unit test report** check above. Coverage artifact: `coverage-unit`. Generated by `.github/workflows/tests.yml`.}

[coderabbitai]

Important

Draft PR not reviewed

Draft PRs are not automatically reviewed by default.

• Trigger a manual review

To automatically review draft PRs, update your CodeRabbit configuration:

reviews:
  auto_review:
    drafts: true

---

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

[chip-peanut-bot]

Chip review — no blocking findings — this is not an approval

Clean for the stated throwaway verification purpose: the sole change nudges only the settings-language fixture by 8px. The PR should remain unmerged as its description states.

Checked clean

• Correctness: pt-4 to pt-6 uses the repository's stock 4px spacing scale, producing the intended 8px vertical shift.
• Scope and adversarial paths: LanguageView serves /settings/language and is registered as the settings-language fixture; no logic, state, authorization, or data path changed.
• Security and slop: the existing utility-class replacement adds no trust-boundary, dependency, duplication, dead-code, or architecture change.
• CI at the exact head: every completed required gate passed, including unit, typecheck, eslint, format, ds-lint, ds-shots-filter, analyze, and ci-success; ds-shots was still in progress when reviewed.

Second opinion by moonshotai/kimi-k3: 0 finding(s), marked with the model name. It reads the diff only, so treat its findings as advice.

Third opinion by claude-opus: 0 finding(s), marked with the model name. It answers only product truth, missing tests and the cross-repo contract, so treat its findings as advice.

Exact head: 4cddecfc7184 · Context: repo · Took 3m (queued 6m)

[github-actions]

🖼 Visual diff — 2 screens moved

5 of 120 shots changed · 115 identical · baseline 85c34ad → head 4cddecf

worst %	screen	widths
4.43%	settings-language	320, 375, 393, 430
2.69%	profile	393

job summary · before/after/diff images — artifact

_{Fixture screenshots, no backend. Advisory — this check never blocks a merge. Posted from the default branch by ds-shots-comment.yml; the report it renders is untrusted data.}