refactor(ios): drop the transport seam usbmux-first made dead

[thymikee]

Answering "is there code we can safely drop now that usbmux is primary?" — yes, this much.

What went dead

IosPhysicalDeviceControl exposed:

resolveRunnerTransport(device, timeoutBudgetMs): Promise<
  { kind: 'network'; tunnelIp: string | null } | { kind: 'usbmux' }
>

That predates #1517, when the control still chose the transport. Since #1517 the route resolver chooses: it returns a usbmux route for an XCTest device, and for a CoreDevice device too, reaching the control only after usbmux has reported the device unattached. At that point the sole question left is "what is the tunnel address".

So the usbmux arm has no live producer or consumer:

• the resolver's if (transport.kind === 'usbmux') branch was unreachable — the only control that gets that far is CoreDevice, which always answers network
• the XCTest implementation returning { kind: 'usbmux' } was called only from a test; production returns a usbmux route before it

Change

• collapse the union to IosPhysicalDeviceTunnel = { tunnelIp: string | null } — one shape, so no discriminator to carry
• rename resolveRunnerTransport → resolveTunnel, which is what it now does; the old name advertised a decision this seam no longer makes
• delete the unreachable branch in the resolver
• XCTest now rejects a tunnel lookup, matching how it already rejects app inventory, process lookup and runner file copy — it has no CoreDevice tunnel, which is exactly why the resolver never asks it for one

Behaviour is unchanged: every path that ran before runs the same way. This removes a false affordance rather than a feature — the interface implied XCTest devices resolve a transport, when the transport question was settled upstream.

Testing

The one test that exercised the dead path now asserts the rejection instead of a value production never produced, matching its sibling tests for the other unsupported XCTest operations. physical-device-control and runner-transport suites pass (19 tests); typecheck and lint clean.

On the gate run

check:affected reports failures in help-conformance-bench, install-source (unzip exited with code 9), owner-identity, device-claims and android-recording. Two of those are assertion failures, which I would normally treat as blocking rather than flake — so I checked rather than assumed: all 51 of those tests pass in isolation, both with and without this change, and none of them are reachable from a diff that touches only Apple physical-device control, the Apple route resolver, and one Apple test. The host has been driving real devices and xcodebuild for hours; unzip failing with a subprocess code and a process-liveness check flipping are both load-sensitive.

Worth a separate look if they recur on a quiet machine — I would rather flag that than bury it.

[github-actions]

Size Report

Metric	Base	Current	Diff
JS raw	1.92 MB	1.92 MB	-2.8 kB
JS gzip	615.5 kB	614.6 kB	-890 B
npm tarball	733.5 kB	732.6 kB	-926 B
npm unpacked	2.57 MB	2.57 MB	-2.8 kB

Startup median (7 runs, lower is better):

Scenario	Base	Current	Diff
CLI --version	23.0 ms	22.1 ms	-0.9 ms
CLI --help	47.0 ms	46.3 ms	-0.6 ms

Top changed chunks:

Chunk	Raw diff	Gzip diff
dist/src/physical-device-control.js	+23.1 kB	+7.6 kB
dist/src/screenshot-result.js	-3.3 kB	-910 B
dist/src/runner-disposal.js	-52 B	-18 B
dist/src/internal/daemon.js	+41 B	+16 B
dist/src/session.js	+363 B	-13 B

[thymikee]

Re-reviewed at 08be0f5: code clean and all CI green. The resolver still owns transport choice: XCTest never reaches tunnel lookup, while CoreDevice reaches resolveTunnel only after the typed usbmux-unattached verdict. Existing route tests cover usbmux-first, CoreDevice same-attempt fallback, and XCTest’s terminal cable hint; the new rejection test makes the removed test-only seam explicit. Remaining readiness evidence: please attach an exact-head physical CoreDevice replay for the cabled usbmux path and Wi-Fi tunnel fallback, or document the concrete device-availability blocker.