Fix RwFit initialization, history sync, and manual readings

[saksham2001]

Summary

Fix RwFit rings pairing without importing data or completing manual readings. Align the native Swift integration with the vendor's published SDK: preserve response payloads, verify initialization and capabilities, decode live readings correctly, and page history using confirmed responses instead of treating timeouts as success.

History is durably saved before consumption, with a recoverable journal for sleep sessions spanning pages. Individual stream failures report partial sync without blocking other supported streams. Add actionable sync errors, accurate freshness, complete opt-in packet capture, and regression coverage. No vendor SDK binary is bundled.

Related issues

Related to #135 (vendor SDK references). Addresses reported SR16 and SY01 sync failures; compatibility with these models still requires hardware validation.

Type of change

• 🐛 Bug fix (non-breaking change that fixes an issue)
• ✨ New feature (non-breaking change that adds functionality)
• 📟 New / improved wearable support (BLE driver layer)
• 🤖 Coach / LLM change (tools, prompts, orchestration)
• 🎨 UI / DesignSystem change
• 🧹 Refactor / chore (no behavior change)
• 📝 Docs only
• ⚠️ Breaking change (existing data, settings, or APIs change)

Behavior change: successful modern history sync consumes transferred records from the ring. RwFit may therefore be unable to import those records afterward. Local imports and sleep-page staging must succeed before consumption; ambiguous delete timeouts are not retried.

How was this tested?

• Added / updated unit tests (PulseLoopTests)
• Ran the test suite locally (⌘U in Xcode)
• Tested on a physical device with a real ring — model: SR16 / SY01 validation pending
• Tested with demo data (-seedDemo YES, no hardware)
• N/A (docs / non-code change)

Ran the CLI equivalent with xcodebuild test on the PulseLoop scheme using iPhone 16 Pro / iOS 18.6: 1,146 tests passed, zero failures. Coverage includes framing and fragmentation, authenticated readiness, manual-command ordering, history paging, partial outcomes, persistence rollback, sleep recovery, and transport timeouts. SwiftLint passed with warning-level findings; git diff --check passed.

Simulator results do not verify BLE behavior on actual SR16 or SY01 hardware. Both require initial sync, HR/SpO₂ readings, reconnect, interrupted-transfer testing, and an opt-in packet export.

Privacy & data

• This change does not send health data off-device without explicit user action.
• No secrets, API keys, or personal data are committed.
• N/A

Raw diagnostic packets remain opt-in in release builds. Sleep transfer pages are stored locally in a bounded, protected, per-device journal, retained for recovery until imported, and removed when app data is cleared. Test fixtures are synthetic; the user's diagnostic export is not committed.

Screenshots / recordings

Not captured. UI changes add a sync-failure message with Retry and use an indeterminate RwFit measurement indicator while command-start timing is variable.

Checklist

• My code follows the project's style (SwiftLint passes).
• I ran the tests and they pass.
• I updated docs / README where relevant.
• I read the Contributing guide.