[HemSoft]

Manual iOS validation rejected two existing subscription-utilization UI tests missing from the release manifest, stopping before XCTest. Include both in the supplemental phase, preserve the existing51 journeys and require53 per family plus the same three preparation/reset invocations. Update the strict manifest/source-discovery contract and developer documentation.

Also retain the last native auto-renew-off predicate observation if its existing25-second wait fails. The failure-only capture preserves short-circuit evaluation, caps fixed-product status output at3 rows with explicit counts, and retains the primary failure if diagnostics fail. No production source, workflow, automatic workload, timeout, test selector, acceptance assertion or marketing exclusion is weakened.

Refs #523. Extracts the existing manifest correction from PR #521, whose source and separate cost/full-qualification evidence remain unchanged. Reviewed/tested head: 4e4a3c34173eb7762b7431328e86c2af3c659c30.

Validation

• Connected Codex review is clean for this exact head. Zero review threads. Required automatic SwiftLint passed.
• Hosted focused validation40345384586, attempt1, completed success on October26 at12:39:50PM EDT. All five quality jobs passed: SwiftLint, strict concurrency, iOS tests/risk, watchOS tests/risk and SwiftPM smoke.
• Retained native receipts verify ordinary iOS 1166/1166 and native phase counts 1/25/1/1, zero skips/expected failures, complete exit0 and owned shutdown. Watch tests 112/112 passed. Risk reports cover2315iOS and161watch functions with empty error lists. All555 retained hosted source-file hashes match the clean reviewed checkout.
• Focused account-menu job passed all four baseline/candidate iPhone/iPad controls. Both source inputs are this SHA, so this establishes correctness only, not improvement. Producer logs and the unchanged comparison validator establish four one-test passes; the1.1GB raw UI-media artifact remains on GitHub and is not claimed visually inspected.
• The original exact-source manifest failure reproduces on main. Repair passes 83 local Python controls covering release/manifest failure handling, function risk and the four unchanged CI-policy assertions. Diff check and pinned strict-concurrency Swift build pass. No UI/product behavior changes, so new rendered product screenshots are not applicable.

Limits and retained adverse evidence

Previous validation40318765544 failed the natural-expiration auto-renew-off wait. One isolated and one ordered local control did not reproduce it; the new hosted run passed it. Its cause remains unexplained, and this diagnostic update is not a claimed causal product fix. Raw failures, local controls and source ledgers remain under /tmp/codexbar-main-ci-20261026/.

A temporary supported SDK-error injection verified the final failure capture and original exit65, but its separate reset hit simulator Busy/preflight and internal simctl lookup errors. Shutdown passed; that simulator is quarantined and the whole local control is unqualified. The injection was removed and exact final compiled Swift bytes restored before commit.

This focused run intentionally skipped the full53-journey release UI gate. It does not qualify an App Store release or PR #521's full sharding/cost gates. Current standalone Usage history performance is also unqualified; its historical noisy failure is retained. No Home sync, Store/release action, polling/timer or CI-notification change. The external Chief supplied the workflow-completion wake; the PR watch alone did not establish it.

[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 2026-10-26T15:00:00.902533Z	4e4a3c3	Manual request

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.

[HemSoft]

@codex review

[HemSoft]

@codex review

Review exact head 4e4a3c34173eb7762b7431328e86c2af3c659c30. The manifest repair now reaches native tests. The hosted natural-expiration wait failed, but one isolated and one ordered local control did not reproduce it. This update preserves its last predicate observation on failure, with unchanged waits, native assertions, short-circuit behavior and primary failure. It is diagnostic evidence work, not a claimed product repair. A supported SDK-error control exercised the capture and retained XCTest65; its separate reset launch failed and is explicitly unqualified. The temporary SDK injection is absent from this head. The old-head review has remained Running for over an hour; this is the single request for the new head.

[chatgpt-codex-connector]

Codex Review: Didn't find any major issues. What shall we delve into next?

Reviewed commit: 4e4a3c3417

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".