[ElliotDrel]

Problem

Follow-up to #26127 for #8732, targeting its branch fix/issue-8732. The busy-renderer regression used a real 250 ms native timer even though the renderer and navigation are mocked. This addresses #26127 (comment) without changing production behavior.

Change

Run the two renderer cases with the Effect test clock. A promise signals when the mocked title read starts, so the native title deadline is already registered before the busy case advances fake native timers by 250 ms. Join the invocation and retain the same committed-page, title and loading assertions. Always restore native timers afterward.

This is a focused test-maintenance correction for an existing implementation. The upstream branch is not writable by this account; the authored correction is test-only. The later main merge temporarily widens this PR comparison until the upstream dependency branch is updated.

Verification

Agent verification on Windows at df614ee:

• Focused renderer cases: 2 passed.
• Complete ServerBrowser.test.ts: 32 passed.
• Targeted lint and formatting passed; server tsc --noEmit passed; diff check passed.
• The runner emitted a separate web pre-transform warning for missing @formkit/auto-animate; all selected server tests executed and passed. No web verification is claimed.
• Historical CI at df614ee failed in the inherited usage-retention fixture; aggregate Check propagated that failure. Maria merged main 8b44f42 into the follow-up as 3ebf8c0, including the existing fix from test(server): pin clock in saved-usage window test #28603. On updated head 3ebf8c0, the aggregate CI Check, Agent Context Budget and Mobile Fingerprint Check succeeded; preview workflows were skipped. Label PR size remains the sole failed check because the comparison includes main history (details below). The earlier action-required statuses were authorization gates, not executed failures. Codex reported no findings and Macroscope approved the original test correction.
• No real-client or native-platform verification; production code is unchanged.

Implemented by gpt-6.1-sol through the Codex harness in T3 Code.

Metadata-check limitation: Label PR size failed with spawnSync git ENOBUFS while comparing the main-integrated head to the older dependency branch (31,303 files). The target branch needs alignment; this is not a product-test failure. See #30699 (comment) .

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-11-04T07:31:06.348829Z	df614ee	PR opened

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" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[macroscopeapp]

Approvability

Verdict: Approved at df614ee

Macroscope's review found this PR approvable — The changes are confined to repository-ignored paths and are described as a test-timing adjustment. Within the explicitly supplied review scope, no material risk requiring human judgment is evident.

Notes:

• No code objects were reviewed. Approvability was decided on eligibility alone.

^{You can add or adjust custom eligibility rules. Learn more.}

[ElliotDrel]

Note

🤖 gpt-6.1-sol responding on behalf of Elliot

The size-label failure is spawnSync git ENOBUFS, not a failed product test: https://github.com/pingdotgg/t3code/actions/runs/41796295273/job/126461119089 . It runs git diff --numstat from the still-old target 989a080 to the main-integrated head 3ebf8c0; GitHub currently reports 31,303 changed files in that comparison. The buffered diff exceeds the script's limit.

The intended correction remains the single test-file commit df614ee. Updating the upstream dependency branch fix/issue-8732 to include current main would realign the stacked comparison and include the existing usage-fixture fix. I have not raised a buffer limit, altered labeling policy, or retried the deterministic failure. Please align the target branch before incorporation; the current large comparison is imported main history, not the correction's scope.