[jackeyfaker77]

Summary

The event-wait tests can exhaust their fixed setImmediate polling budget before asynchronous Host work reaches the expected checkpoint. In CI this caused the late-authorization and slow-cooldown-persistence tests to fail before their third authorization observation.

Give the test helper a 5-second monotonic wall-clock deadline with 5 ms sleeps. The injected observation clock remains fixed, so existing timeout, Retry-After and cancellation assertions retain their exact semantics. A regression test verifies that delayed Host work completes without advancing that clock or firing its timers. Production code is unchanged.

Extracted from #6089 following its scope review, so this test fix can land independently of Write/Edit evidence. Refs #6113.

Failure evidence: https://github.com/apache/maka/actions/runs/39031806684/job/117393589335

Verification

• On this standalone branch based on main: clean npm ci, full root build, typecheck, lint, format, both desktop/UI Knip checks, ASF headers and whitespace checks passed.
• Complete event-wait-manager suite: 251 passed, no failures or skips.
• For the identical patch originally validated in fix(runtime): guard Write/Edit against content changes before execution #6089: the new regression failed with the old helper; the two original CI failures passed in 10 independent processes (20/20 cases).
• Independent Codex review found no P0-P2 issues or new P3 suggestions; the reviewer independently passed the regression and both original CI failure cases (3/3). This is AI review, not maintainer approval.
• Full-monorepo tests and native Linux/macOS execution were not run locally.

AI use

• No generative tool made a substantive contribution
• Generative tooling made a substantive contribution

Tool(s) and scope: OpenAI Codex assisted with diagnosis, implementation, tests, extraction, independent agent review and this description. AI review does not replace human review. The commit includes a Generated-by: OpenAI Codex trailer.

Checklist

• Tests cover the change and fail without it
• Lint, format, typecheck and the affected suites pass locally

Does this PR entail a change in behavior?

• Yes — described under Summary above
• No (production behavior is unchanged; only test synchronization changes)

[Astro-Han]

Automated review notice: This comment was posted by an automated review agent. It is not an independent human review and does not replace one.Reviewed 07bad455b02070e77bac788ffb9611507941d9d3 (one test file, +26/-2). This is the same change I reviewed inside #6089 (50565abd). I found no P0–P2 issue. CI test passes, and the PR is mergeable.

eventually() in event-wait-manager.test.ts now polls on a 5 s monotonic deadline with 5 ms sleeps, instead of 100 rounds of setImmediate. Real Host I/O therefore gets wall-clock time to finish, while the injected observation clock stays fixed. Timeout, Retry-After and cancellation assertions keep their exact meaning. The added self-test checks that delayed host work completes without advancing the observation clock or firing its timers. This fixes the flaky late-authorization and slow-cooldown-persistence cases described in the PR.

[Astro-Han]

Thanks for the contribution and for addressing the review feedback. Merging now.

[Astro-Han]

Approved at @Astro-Han's explicit request: no blocking findings in the automated review of this head, and CI is green.