perf: reuse Apple runner cache across version bumps

[thymikee]

Summary

Reduce Apple runner cold-build and first-use work while keeping cache invalidation reliable.

Shipped changes:

Area	Before	After
Runtime runner Swift surface	Runtime code and Swift unit-test helpers compiled together	Unit-test-only Swift is behind AGENT_DEVICE_RUNNER_UNIT_TESTS
Asset catalog	Tiny branding assets still ran actool during cold runner builds	Assets.xcassets and bundled branding images removed
Runner cache key	Invalidated on package version bumps	Reused across package version bumps; still keyed by toolchain, SDK, source fingerprint, platform, and build flags
Maintainer/CI script destination	generic/platform=iOS Simulator, which can build arm64 + x86_64	Concrete available simulator when possible, with 3s fallback to generic
Swift unit coverage	Coupled to build:all	Separate CI compile job for the guarded Swift unit-test surface

Measurements

Measured on Xcode 26.2 with fresh DerivedData per run.

Scenario	Runs	Before median / mean	After median / mean	Delta
User runner xcodebuild build-for-testing, iPhone 17 Pro Max simulator	5 before + 5 after	8.06s / 7.82s	7.76s / 7.75s	-3.7% median
First use: open settings then first snapshot -i, no runner cache	1 before + 1 after	31.5s	21.4s	-32% wall time
pnpm build:xcuitest iOS script destination	5 before + 5 after	8.210s / 8.042s	5.767s / 5.825s	-29.8% median

Notes:

• The user runner build-only win is modest because Swift frontend work and Xcode warm caches dominate after the first run, but the first-use CLI path is materially faster.
• The script destination win applies to maintainer/CI build:xcuitest; runtime/user builds already used concrete simulator destinations.
• Package dry run stayed small: 161 files, 568 KB packed / 1.97 MB unpacked, with no asset catalog or branding images.

Rejected Experiments

I also tested wholemodule + -num-threads, debug info off / Swift -gnone, explicit modules off, testability off, incremental off, lazy recording compilation, and related combinations. None met the shipping bar for runtime builds: either the median win was noise-level, wall time got worse, or the change added complexity/semantic risk.

Validation

• pnpm check:quick: passed
• pnpm exec vitest run src/platforms/ios/__tests__/runner-client.test.ts src/platforms/ios/__tests__/runner-xctestrun.test.ts src/platforms/ios/__tests__/runner-icon.test.ts: 79 tests passed
• pnpm build:xcuitest: passed for iOS and macOS
• npm pack --dry-run --ignore-scripts --json --cache /private/tmp/agent-device-npm-cache: verified package contents and asset removal
• AGENT_DEVICE_XCUITEST_INCLUDE_UNIT_TESTS=1 ... pnpm build:xcuitest:macos: passed earlier and showed -D AGENT_DEVICE_RUNNER_UNIT_TESTS

Latest pushed SHA: 5cce08d8d.

[github-actions]

Size Report

Metric	Base	Current	Diff
JS raw	1.4 MB	1.4 MB	+167 B
JS gzip	445.5 kB	445.6 kB	+52 B
npm tarball	584.7 kB	545.9 kB	-38.8 kB
npm unpacked	2.0 MB	1.9 MB	-38.9 kB

Startup median (7 runs, lower is better):

Scenario	Base	Current	Diff
CLI --version	28.6 ms	28.2 ms	-0.4 ms
CLI --help	48.7 ms	50.1 ms	+1.4 ms

Top changed chunks:

Chunk	Raw diff	Gzip diff
dist/src/9722.js	+167 B	+52 B

[github-actions]

PR Preview Action v1.8.1
🚀 View preview at https://callstack.github.io/agent-device/pr-preview/pr-900/
Built to branch gh-pages at 2026-06-27 12:53 UTC. Preview will be ready when the GitHub Pages deployment is complete.

[thymikee]

Reviewed the latest head, including the new concrete-simulator destination commit.

I do not see a blocker. The default iOS/tvOS script path now prefers an available concrete simulator id, but keeps the generic simulator fallback when simctl lookup fails. Cache metadata still normalizes the destination back to the simulator family, so choosing a specific UDID should not churn runner cache keys. The unit-test Swift flag is now explicitly represented in both the shell build path and metadata comparison, so runtime and unit-test runner variants stay separated. Current CI is green, including Swift Runner Unit Compile, typecheck, unit, integration, smoke, and iOS runner compatibility.

Residual risk: this is still Apple runner build/cache behavior, so I would treat the PR body’s local Xcode measurements and first-use validation as the device-facing evidence rather than relying on fixture tests alone. With that evidence plus green CI, this is ready for maintainer merge judgment.