Add public API unit tests with ≥80% coverage (#30)

[itsniper]

Closes #30 (NFR-2.1).

Summary

Expands ReliaBLETests to give every public method and property on ReliaBLEManager — plus the public value types — at least one behavioral test, reaching ≥80% code-path coverage of Sources/ReliaBLE/.

What's covered

• ReliaBLEManager: init(config:), loggingService, state / currentState, authorizeBluetooth() (denied / restricted / allowed / cancellation), startScanning / stopScanning (lifecycle + not-powered-on no-op), peripheralDiscoveries, discoveredPeripherals (replay + live updates), connect(to:) (success + unknown).
• Value types: BluetoothState.description for every case, Peripheral / PeripheralDiscoveryEvent id-based equality & hashing, AdvertisementData typed extraction (full + empty), ReliaBLEConfig defaults, and the logging module (LoggingService, LogMessage, OSLogWriter).

Test design

All BLE paths run against Nordic's CBMCentralManagerMock through the existing three-target SPM trick (forceMock: true) — no real CoreBluetooth hardware.

Because BluetoothActor.shared is a process-lifetime singleton and the mock keeps global static state, the tests live in a single @Suite(.serialized). A one-shot SimulationConfig registers the simulated peripheral and pins authorization before any central is created; an ensureReady helper re-establishes a known baseline per test, so the suite is order-independent and stable across repeated runs.

Adds CoreBluetoothMock and Willow as test-target dependencies.

Coverage

Measured with swift test --enable-code-coverage + llvm-cov report on macOS:

Metric	Sources/ReliaBLE/
Line	91.76%
Region	87.89%
Function	92.31%

Per file: ReliaBLEManager 100%, ReliaBLEConfig 100%, Peripheral/AdvertisementData 100%, Logging 100%, BluetoothActor 87.7% line / 82.0% region, PeripheralDiscoveryEvent 94.1% line.

Remaining uncovered paths are the hard-to-simulate ones (central .resetting/.unsupported/.unauthorized invalidation, the @unknown default assertion, and "no central manager" guards unreachable once the singleton central exists).

Acceptance criteria

• Every public method/property on ReliaBLEManager has a behavioral test.
• swift test passes on macOS with zero failures (27 tests).
• Coverage of Sources/ReliaBLE/ ≥ 80%.
• Tests follow conventions: Swift Testing (@Test), @testable import ReliaBLEMock, no real CoreBluetooth hardware.

🤖 Generated with Claude Code

[Copilot]

Pull request overview

This PR expands the ReliaBLETests target to provide behavioral coverage for the public surface area of ReliaBLEManager and key public value/logging types, aligning with NFR-2.1’s ≥80% coverage goal while keeping BLE interactions fully mocked.

Changes:

• Adds a serialized Swift Testing suite that exercises ReliaBLEManager authorization, scanning/discovery, connection, and stream replay semantics using CBMCentralManagerMock.
• Adds direct behavioral tests for public value/logging types (BluetoothState.description, Peripheral / PeripheralDiscoveryEvent identity semantics, AdvertisementData, ReliaBLEConfig defaults, LoggingService / LogMessage / OSLogWriter).
• Updates ReliaBLETests target dependencies to include Willow and CoreBluetoothMock so tests can import and drive the mock + logging types directly.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated no comments.

File	Description
Tests/ReliaBLETests/ReliaBLEManagerTests.swift	Adds a serialized end-to-end mock-driven test harness covering ReliaBLEManager’s public API and public value/logging behaviors.
Package.swift	Adds Willow and CoreBluetoothMock as explicit ReliaBLETests dependencies to support the new test imports.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.