T12-T14: platform liveness via geolocator; drop flutter_foreground_task

geolocator_android's ForegroundNotificationConfig raises a service with
foregroundServiceType=location and holds a wake lock for as long as the position
stream is subscribed, and the plugin declares that service in its own manifest.
That covers everything the native TrackingService used a foreground service and
a PARTIAL_WAKE_LOCK for, so flutter_foreground_task is gone -- and with it both
deprecation warnings that were recorded under T01.

Known parity gap for T27: geolocator's notification cannot carry actions, so the
native Pause/Resume buttons in the shade are not reproduced.

iOS: UIBackgroundModes plus allowBackgroundLocationUpdates keep the isolate and
the writer timer alive while updates flow. Two settings matter more than they
look -- pauseLocationUpdatesAutomatically is false because CoreLocation does not
reliably restart, and activityType is otherNavigation rather than
automotiveNavigation, which would snap fixes to the road network and silently
falsify the recorded path. Usage strings are specific, not generic.

Android deliberately does not request ACCESS_BACKGROUND_LOCATION: a foreground
service with a visible notification is exactly the supported case, and asking
would trigger the harder "Allow all the time" flow for no benefit.

T14 process-death resume is implemented and called at startup, covered by four
tests including the 111 km crash-gap guard.

Verified by building both platforms and launching on the iOS simulator: Drift
opened, Riverpod resolved, restore found no active trip, controls correct.
lib/main.dart is a harness so the pipeline can be exercised on hardware before
Phase 4 builds screens; T15 replaces it.

Phase 3 done: 145 tests passing, analyze clean.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

docs/port/PROGRESS.md

@@ -476,3 +476,103 @@ with the fix.
 **A regression test nobody has watched fail is not yet a regression test.**
 
 **145 tests passing, analyze clean.**
+
+---
+
+## T12 / T13 — Platform liveness · **implemented, unvalidated on a real ride**
+
+`lib/src/recording/geolocator_location_source.dart`, plus the Android manifest and iOS
+`Info.plist`. This is the **only** file in the app that knows the two platforms differ.
+
+### `flutter_foreground_task` dropped
+
+Dylan's call, on the T10 finding. `geolocator_android`'s `ForegroundNotificationConfig`
+raises a service with `foregroundServiceType="location"` and holds a wake lock for as
+long as the position stream is subscribed. The plugin declares its own
+`GeolocatorLocationService` in its manifest, which merges automatically — nothing to
+declare ourselves. Verified in the merged manifest.
+
+**Both deprecation warnings recorded under T01 are now gone from the build output.**
+
+> **⚠ Parity gap for T27:** geolocator's notification cannot carry *actions*, so the
+> native app's Pause/Resume buttons in the notification shade are not reproduced. Tapping
+> the notification opens the app. This is the only known capability lost so far.
+
+### Android
+
+Permissions mirror the native manifest. **`ACCESS_BACKGROUND_LOCATION` is deliberately
+not requested** — recording runs under a foreground service with a visible notification,
+which is exactly what `foregroundServiceType="location"` is for. Asking for background
+location would trigger the harder "Allow all the time" flow for no benefit.
+
+### iOS — two settings that matter more than they look
+
+```dart
+pauseLocationUpdatesAutomatically: false   // CoreLocation does not reliably restart
+activityType: ActivityType.otherNavigation // NOT automotiveNavigation
+```
+
+`automotiveNavigation` makes CoreLocation **snap fixes to the road network**. For a
+navigation app that is a feature; for a recorder whose entire purpose is the path actually
+travelled, it silently falsifies the data. `otherNavigation` is the correct choice for a
+motorcycle.
+
+`UIBackgroundModes: [location]` plus `allowBackgroundLocationUpdates` keeps the process —
+and therefore the Dart isolate and the writer timer — alive while updates flow. Usage
+strings are written specifically rather than generically, since vague text is the leading
+cause of Guideline 5.1.1 rejection.
+
+**None of this is validated.** Whether iOS suspends a stationary app mid-ride is
+answerable only by riding. It belongs to T25.
+
+---
+
+## T14 — Process-death resume · **complete**
+
+Implemented in `RecordingEngine.restoreAfterProcessDeath` and called from a post-frame
+callback at startup. Covered by four tests, including the 111 km guard documented under
+T11. The remaining acceptance criterion — force-killing mid-ride on real hardware —
+belongs with T25.
+
+---
+
+## The app runs
+
+Built and launched on the iOS simulator. Drift opened against app-private storage,
+Riverpod resolved the graph, `restoreAfterProcessDeath` correctly found no active trip,
+and the controls were in the right state (START enabled, PAUSE/STOP disabled while idle).
+
+`lib/main.dart` is a deliberate **harness**, not the real UI — it exists so the pipeline
+can be exercised on hardware before Phase 4 builds any screens, because the pipeline is
+precisely the part unit tests and emulators cannot validate. T15 replaces it.
+
+---
+
+## ⚠ Disk: the real environmental constraint
+
+Today's arc: **36 GiB free → 355 MiB** at the worst point, with roughly 7 GiB reclaimed
+along the way and still ending at 6.6 GiB.
+
+A full dual-platform build cycle costs roughly **10 GiB** — Gradle caches, a 2.8 GB NDK,
+CocoaPods, DerivedData, `build/`, and the installed simulator app. It is reclaimable, but
+it is not optional space.
+
+Standing footprint: `/opt/homebrew` 17 GB, `~/Library/Android` 7.5 GB, Flutter SDK 3.9 GB,
+iOS runtime ~8 GB.
+
+**Before T24 boots the Android emulator** (fixed 7.4 GiB free required):
+
+```bash
+flutter clean && rm -rf ios/Pods ~/Library/Developer/Xcode/DerivedData/*
+```
+
+Realistically this machine needs headroom freed outside the dev tooling before Phase 5.
+
+---
+
+## Phase 3 complete
+
+**145 tests passing, analyze clean, both platforms building and the app running on iOS.**
+
+Next: **Phase 4, the UI** — six screens, and the first widget tests this project has ever
+had.