test(rivetkit): cover async websocket close db access

[NathanFlurry]

Description

Please include a summary of the changes and the related issue. Please also include relevant motivation and context.

Type of change

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• This change requires a documentation update

How Has This Been Tested?

Please describe the tests that you ran to verify your changes.

Checklist:

• My code follows the style guidelines of this project
• I have performed a self-review of my code
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• My changes generate no new warnings
• I have added tests that prove my fix is effective or that my feature works
• New and existing unit tests pass locally with my changes

[NathanFlurry]

• fix(pegboard): align drain grace defaults with stop threshold #4958 : 2 dependent PRs (#4960 , #4966 )
• fix(rivetkit): preserve bridge error messages #4959
• fix(rivetkit): fire abort signal on shutdown finalize #4957
• fix(rivetkit): use startup kv preload #4956
• fix(client): flatten skipReadyWait option #4955
• ci(publish): add manual preview publish flow #4954
• fix(depot): add ltx decode diagnostics #4953
• fix(universaldb): preserve plain set values #4952
• fix(universaldb): align versionstamp operand encoding #4951
• fix(rivetkit): export database provider types #4950
• wip: kitchen sink agentic loop finish #4949
• chore(rivetkit): note duplicate admission logic #4948
• test(rivetkit): cover bypass fetch header forwarding #4947
• feat(kitchen-sink): add actor probe controls #4946
• fix(gateway): split actor response wait errors #4945
• fix(guard): handle bypass preflight before query parse #4944
• feat(kitchen-sink): add mock agent event log copy #4943
• fix(kitchen-sink): default mock agent to local endpoint #4942
• test(rivetkit): enable driver lifecycle regressions #4941
• test(rivetkit): verify cached websocket db refs exceed grace #4940
• fix(rivetkit-core): keep sleeping actors reachable during grace #4939
• test(rivetkit): cover async websocket close db access #4938 👈 (View in Graphite)
• fix(rivetkit): keep actors awake until keepAwake work finishes #4937
• fix(rivetkit): persist native onSleep state on errors #4936
• fix(rivetkit-core): allow dispatch during sleep grace #4935
• fix(rivetkit): make duplicate sleep requests idempotent #4934
• main

This stack of pull requests is managed by Graphite. Learn more about stacking.

[claude]

Code Review

Overview

This PR re-enables three previously skipped tests (TODO #4705) that verify async WebSocket close handler DB access works correctly during sleep shutdown. The changes are a meaningful quality improvement.

---

What Changed

1. Three test.skip calls converted to active tests.
2. Actor names switched to crypto.randomUUID() suffixes for isolation.
3. waitForAction(actor.getStatus, ...) polling replaced with time-based waitFor(...).
4. Richer assertions: checks event ordering (close-start -> close-finish -> sleep -> wake) instead of just counting wake/sleep cycles.

---

Positive Observations

Polling removal is the right fix. The previous approach called actor.getStatus (an actor action) inside a polling loop to wait for sleep to complete. Per project guidelines, each action call counts as activity and can reset the sleep deadline, making those tests inherently self-defeating. Switching to a time-based waitFor is the correct resolution.

UUID-based actor names correctly prevent cross-test actor identity collisions, which was likely a source of the original flakiness.

Event ordering assertions are semantically stronger than the old sleepCount/startCount checks. Verifying close-start < close-finish < sleep < lastIndexOf("wake") directly encodes the contract being tested.

lastIndexOf("wake") is the right call - onWake fires both at initial startup and after sleep resumes, so there are two "wake" entries and the test correctly targets the second one.

---

Issues

Time-based waits with no event signal. The waitFor(driverTestConfig, RAW_WS_HANDLER_DELAY + RAW_WS_HANDLER_SLEEP_TIMEOUT + 1_000) calls (1350 ms total) are a pure wall-clock sleep with no event that fires when the sequence completes. This works because the budget is generous relative to the delays, but it is inherently load-sensitive. On an overloaded CI runner, the sleep completion and DB write could land after the waitFor expires, causing a spurious failure. A comment explaining the trade-off (polling was avoided because each action resets the sleep deadline) would help future maintainers.

Duplicate test body. The two test cases (onclose property vs addEventListener("close", ...)) are structurally identical apart from the actor name. If the ordering assertions or wait duration change, both need to be updated in sync. A shared helper accepting the actor handle would keep them DRY.

No comment on waitFor explaining why time-based. The check:wait-for-comments script only enforces comments on vi.waitFor, not the project's waitFor helper. Given how non-obvious the rationale is, a short comment above each waitFor call would prevent future maintainers from reverting to the polling pattern.

---

Minor Nits

• The removed sleepCount === 1 / startCount === 2 assertions are replaced by event ordering checks, which is strictly stronger for this use case. The coverage shift is intentional and correct.
• The 30,000 ms test timeout is appropriate for the 1350 ms wait plus real-timer overhead headroom.

---

Summary

The core approach is sound and the tests are now functionally correct. The main risk is latent time-sensitivity in the waitFor calls - a comment explaining why polling was avoided would protect against accidental regression.

[github-actions]

Preview packages published to npm

Install with:

npm install rivetkit@pr-4938

All packages published as 0.0.0-pr.4938.50d75f0 with tag pr-4938.

Engine binary is shipped via @rivetkit/engine-cli on linux-x64-musl, linux-arm64-musl, darwin-x64, and darwin-arm64. Windows users should use the release installer or set RIVET_ENGINE_BINARY.

Docker images:

docker pull rivetdev/engine:slim-50d75f0
docker pull rivetdev/engine:full-50d75f0

Individual packages

npm install rivetkit@pr-4938
npm install @rivetkit/react@pr-4938
npm install @rivetkit/rivetkit-napi@pr-4938
npm install @rivetkit/workflow-engine@pr-4938