feat(mobile): drag iOS threads from their context menus

[juliusmarminge]

Long-press thread menus need to coexist with direct dragging, including moves between Pinned and Active. This adds native iOS drag/drop to the context-menu host view and routes drops through the lifecycle-aware arrangement actions from #10496.

All live thread rows can initiate a drag, including expanded Snoozed and Settled rows. Pinned and Active rows and headers accept drops. The Settled divider accepts settling drops and remains available when empty. Headers remain available when a section is empty, in both Home and the navigation sidebar. Moving into a section pins, unpins, wakes, or un-settles the source before saving its position. Starting a drag suppresses row taps. Android uses the shared Arrange screen.

During a native drag, snapshots of visible rows slide aside to open a full insertion gap. UIKit drop targets keep their original coordinates so the animated rows cannot destabilize hit testing. Pictures and original opacity are restored on completion or cancellation; recycled cells are refreshed during edge scrolling. Animation runs only during the gesture and respects Reduce Motion. A live native action badge follows the dragged card and updates as it crosses sections. Source-provided action labels use the same mapping as Arrange; invalid targets hide the badge, and unsupported settlement is rejected. The thread snapshot stays stable.

Validation: 78 focused tests and mobile TypeScript pass. The shared drop actions passed real Android Arrange gestures for reorder, pin, unpin, empty-section pinning, wake, and un-settle, with persisted server state checked. The first layer contains screenshots and a recording.

Draft limitation: the updated native iOS gesture still needs an on-device pass. The UIKit preview and live badge helper passes Swift typechecking against the iOS Simulator SDK, but Xcode 27 automation cannot inject the required drag movement. Shared Arrange motion passed a real Android emulator run; the first layer contains the recording.

This native patch requires a new binary. Standalone builds without Metro are queued from the final stack: iOS, Android.

Model: GPT-6. Harness: Codex.

[github-actions]

Thread transfer impact

✅ Thread transfer remains within every enforced ceiling.

ℹ️ No successful main baseline artifact is available yet. This run establishes the initial measurement.

Provider	Metric	Main baseline	This PR	Impact	PR ceiling
Codex	Total thread wire	—	13.5 KiB	—	15.1 KiB	✅
Codex	Thread snapshot wire	—	7.0 KiB	—	7.3 KiB	✅
Codex	Live turn WebSocket wire	—	6.5 KiB	—	7.8 KiB	✅
Codex	Live turn WebSocket decoded	—	57.0 KiB	—	66.4 KiB	✅
Codex	Live turn messages	—	8	—	21	✅
Claude	Total thread wire	—	13.6 KiB	—	15.1 KiB	✅
Claude	Thread snapshot wire	—	7.1 KiB	—	7.3 KiB	✅
Claude	Live turn WebSocket wire	—	6.5 KiB	—	7.8 KiB	✅
Claude	Live turn WebSocket decoded	—	57.8 KiB	—	66.4 KiB	✅
Claude	Live turn messages	—	9	—	21	✅

Baseline: unavailable · PR result: 2f3ee76 · Source CI: success

Scenario and decoded snapshot size

10 historical turns, 5 command tools per turn, 878.9 KiB retained MCP result per historical turn, and a 1.05 MiB retained result in the measured turn.

• Codex decoded thread snapshot: 113.9 KiB
• Claude decoded thread snapshot: 114.6 KiB

Updated in place by a trusted workflow. PR artifacts are strictly validated and never executed.