Reduce simulator lifecycle polling overhead in multi-device workflows

[shaun0927]

Tracked by #713.

Background

Parallel simulator automation is a core OpenSafari value. Repeated full simctl list devices -j polling during boot/shutdown can become expensive on machines with many runtimes or devices.

Objective

Make simulator lifecycle waits cheaper and more deterministic for parallel device workflows.

Scope

• Use narrower simulator status commands where available.
• Add short-lived device state caching during polling loops.
• Share state lookups between manager and pool code.
• Keep behavior compatible with supported Xcode versions.

Target files and paths

• src/simulator/manager.ts
• src/simulator/sim-pool.ts
• src/simulator/pool.ts
• src/simulator/simctl.ts
• tests/unit/*simulator*.test.ts

Implementation steps

1. Audit current boot/shutdown polling callers.
2. Introduce a simulator state reader with a short TTL for repeated polling.
3. Use simctl bootstatus <udid> -b for boot readiness when it is available.
4. Keep fallback to current polling for unsupported environments.
5. Add tests for cache hit, cache invalidation, and fallback paths.

Acceptance criteria

• Boot wait does not repeatedly parse the full device list when a narrower command is available.
• Shutdown wait uses a bounded polling strategy with clear timeout behavior.
• Parallel pool operations share state reads during a polling tick.
• Fallback behavior works when bootstatus is unavailable.

Verification

• npm run lint -- --quiet
• Targeted simulator manager/pool unit tests
• Manual macOS smoke: boot and shutdown one simulator, then two simulators in parallel.

Dependencies / recommended order

Run after the proxy/WebKit speed issues or in parallel by a separate PR because it touches simulator lifecycle only.

Risks and cautions

• Do not assume all Xcode versions support the same simctl commands.
• Do not cache state across lifecycle mutations without invalidation.
• Do not hide timeout errors.

Non-goals

• Do not redesign device preset selection.
• Do not change public tool names or parameters.
• Do not optimize native input routing in this issue.

Ambiguity reducers

• “Short TTL” means cache lifespan measured in a single polling interval, not long-term device inventory caching.
• Fallback must be explicit and tested.
• Lifecycle actions invalidate relevant cache entries.

Validity review

Proceed. This directly supports parallel multi-device automation and CI speed.

Implementation detail refinements

• Wrap bootstatus use in a capability-detected helper.
• Record the command path in debug logs without excessive noise.
• Prefer dependency injection for Simctl in tests.

Verification checklist reinforcement

• Boot path uses bootstatus or documented fallback.
• Cache invalidates after boot/shutdown/delete.
• Unit tests cover unavailable-command fallback.
• Parallel boot test avoids duplicate full-list calls in mock assertions.

[shaun0927]

검증 완료(2026-05-16 KST): 최신 origin/develop / OpenSafari v0.6.0 기준으로 해당 이슈의 실행 가능한 체크리스트를 확인했습니다.

Evidence:

• PR-linked implementation is merged into develop.
• npm test -- --runTestsByPath ... targeted suites: 10 suites / 103 tests passed for HTTP auth, high-risk HTTP gating, auth persistence, audit logging, network interception, zombie cleanup, and device_boot WebKit retry.
• Broader targeted suites: 23 suites / 557 tests passed for lazy tool registration, proxy readiness, WebKit performance/refactors, socket discovery cache, simulator lifecycle/refactors, native input backends, DOM input scripts, coordinate-space, and AX diagnostics.
• npm run lint -- --quiet ... passed for the changed/validated paths.
• npm run build passed for v0.6.0.
• Real OpenSafari HTTP smoke: serve --http bound to 127.0.0.1; /health returned 200; unauthenticated /mcp returned 401; authenticated initialize succeeded; disallowed CORS origin returned 403.
• Real OpenSafari device smoke for Recover device_boot when WebKit client registration fails #756: device_boot on iPhone 16 iOS 26.4 returned a booted device and proxy; subsequent diagnose retained the simulator in device even while WebKit was unavailable.

체크리스트 중 외부 앱/장기 수동 검증 성격이 아닌 항목은 모두 체크 가능하다고 판단하여 완료로 닫습니다.