Mesh and geohash timelines are now store conversations. All public
mutation sites flow through store intents; PublicMessagePipeline keeps
its 80ms UI batching but commits batches via store appends with each
buffered entry carrying its destination conversation (a mid-batch
channel switch now flushes instead of dropping the buffer).
ChatViewModel.messages becomes a cached get-only view of the active
conversation, invalidated through the change subject. The mesh
late-insert threshold is consciously removed: it only ever ordered the
non-rendered messages copy, so strict timestamp insertion makes the
working set agree with rendered order. PublicTimelineStore and the
per-message full-array legacy sync are deleted; the coalescing bridge
mirrors public conversations for the remaining legacy readers.
pipeline.publicIngest: 6.6k -> 9.5k msg/s (+45%); private steady;
store.append 237k/s.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
docs/CONVERSATION-STORE-DESIGN.md records the approved design: a
single-writer ConversationStore of per-conversation ObservableObjects
(per-conversation publishing, incremental ID index, folded caps, typed
change subject) replacing today's four-store/three-bridge topology,
with a five-step migration plan and explicit deletions/non-goals.
New pipeline benchmarks measure the CURRENT architecture end-to-end so
every migration step is judged against real before-numbers:
pipeline.privateIngest ~9.7k msg/s, pipeline.publicIngest ~6.8k msg/s
(200-message passes, stable within 1.5%).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Vendored arti.xcframework rebuilt from source (Rust 1.96.0, normalized
archive metadata for reproducible hashes). New ARTI-BINARY-PROVENANCE.md
records toolchain, rebuild steps, and a SHA256 manifest for every file
in the xcframework. A new CI workflow turns that policy into a gate:
PRs must keep the binary matching the manifest, and binary changes must
ship with source/lockfile/build-script evidence.
Also raises TorManager.awaitReady's default timeout from 25s to 75s to
match the bootstrap monitor deadline - a shorter wait reported "not
ready" while Arti was still legitimately bootstrapping, silently
stranding queued relay work.
Privacy policy, Tor integration doc, and privacy assessment updated to
match the current implementation.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* Refactor BLE transport event handling
* Make image output paths unique
* Keep queued Nostr read receipts alive
* Refine BLE ingress fanout
* Rediscover BLE service after invalidation
* Extract BLE notification retry buffer
* Extract BLE inbound write buffer
* Extract BLE fragment assembly buffer
* Tidy secure log handling from device run
* Allow self-authored RSR ingress replies
* Harden read receipt queue test timing
---------
Co-authored-by: jack <jackjackbits@users.noreply.github.com>
* feat: Implement Request Sync Manager (V2 Sync)
- Add RequestSyncManager to track and attribute sync requests
- Update BitchatPacket and BinaryProtocol to support IS_RSR flag (0x10)
- Update RequestSyncPacket with new TLV fields (sinceTimestamp, fragmentIdFilter)
- Update GossipSyncManager to use unicast sync requests and mark responses as RSR
- Update BLEService to enforce timestamp validation for normal packets and exempt valid RSRs
- Add documentation for the new sync manager mechanism
* fix: Resolve compilation errors in V2 Sync implementation
- Remove duplicate restartGossipManager in BLEService
- Add missing TransportConfig constants for sync
- Add 'sync' log category to BitLogger
- Add missing BitLogger import in GossipSyncManager
* fix: Update tests for V2 Sync changes
- Add requestSyncManager parameter to GossipSyncManager init in tests
- Implement getConnectedPeers stub in RecordingDelegate
- Remove unused variable warning in SubscriptionRateLimitTests
---------
Co-authored-by: a1denvalu3 <>
Co-authored-by: jack <212554440+jackjackbits@users.noreply.github.com>
* feat(tor): Tor-by-default scaffold and integration
- Add TorManager with static/dlopen start, torrc generation, SOCKS probe
- Add TorURLSession; route Nostr/Web fetches via SOCKS proxy
- Add chat system messages for Tor status; show progress (macOS) and ready
- Disable ControlPort bootstrap monitor on iOS; keep it on macOS
- Make Tor waits non-blocking; avoid main-actor stalls on startup
- Queue & flush Nostr subscriptions on relay connect; skip duplicates
- Always rewrite torrc at launch to fix iOS container path mismatches
- Link libz; add project wiring for tor-nolzma.xcframework
- Minor fixes: SOCKS probe resumeOnce guard, entitlement for network.server (macOS)
* iOS: deterministic Tor recovery + 100% gating; BLE-first; session rebuild
- Restart/wake Tor on foreground via ControlPort (ACTIVE/SHUTDOWN),
avoid restarts during bootstrap; add NWPathMonitor to trigger checks
- Use NWConnection control polling for GETINFO; remove blocking CFStream
readers to avoid QoS inversions; compute readiness from SOCKS + 100%
- Rebuild TorURLSession on resume; reset Nostr connections to rebind
- Gate all internet after full bootstrap; keep BLE mesh startup fast
- Fix Swift 6 capture issues; hop UI updates to @MainActor
- Remove Tor progress spam; persist initial "starting tor..." system message
* UI: show Tor system messages only in geohash channels (not mesh)
- Gate "starting tor..." and readiness/timeout messages to geohash view
- Add helper addGeohashOnlySystemMessage() to avoid posting to mesh timeline
- Persist system messages in geohash backing store via addPublicSystemMessage()
* Relays: treat repeated -1011 handshake failures as permanent; skip reconnects
- Classify NSURLErrorBadServerResponse as permanent and stop retrying
- Filter permanently-failed relays from subscribe/connect attempts
- Avoid reconnect scheduling for permanently failed relays
* Embed Tor via tor_api; deterministic restart + Nostr gating; add Tor notifications
- Run Tor via tor_api in a dedicated thread with OwningControllerFD
- Cleanly stop Tor on background; restart on .active (single instance)
- Avoid fallback to tor_main/dlopen; add is-running check to prevent duplicates
- Fix argv lifetime in C glue to avoid strcmp crash on start
- Gate Nostr connect/subscribe/send until Tor is fully ready
- Rebuild URLSession + reset relays after Tor readiness (scene-based)
- Remove TorDidBecomeReady double-reset and appDidBecomeActive resubscribe
- Add TorWillRestart/TorDidBecomeReady notifications and chat system messages
- Debounce path-change restarts; ACTIVE poke first; coalesce subs; cancel stale reconnect timers
- Project: add CTorHost.c and TorNotifications.swift to targets; fix libz.tbd path
* Defer Nostr setup logs until Tor is ready; fix subscribe coalescing and reconnect generation
- Move "Connecting to Nostr relays" log after awaitReady()
- Log "Queuing subscription" when Tor not ready; only coalesce when handler exists
- Clear coalescer on unsubscribe
- Cancel stale reconnect timers using connectionGeneration
- Remove app-level TorDidBecomeReady reset to avoid duplicate reconnects
- Debounce path-change restarts
* Gate Nostr init/subscription logs until Tor is ready
- ChatViewModel: await Tor readiness before initializing Nostr and logging
- Only log GeoDM subscription when Tor is ready to avoid early noise
* Make Nostr connect single-sourced; defer DM subscription until connected
- Remove duplicate connect call from ChatViewModel; let scene-based flow connect
- Setup DM subscription once on first connection via sink
- Reduce early subscription send/cancel noise after Tor restarts
* On launch, queue Nostr subscriptions without initiating connects; let centralized connect handle it
- In subscribe(), if no connections exist, just list relays and queue subs
- Avoids early send/cancel churn before connect() runs post-Tor-ready
* Always queue subscriptions and flush on connection; avoid immediate sends
- Prevents early send/cancel churn at launch and during reconnects
- If relays are already connected, flush immediately; otherwise pending until connected
* UI: scope Tor restart messages to geohash channels; skip initial foreground restart on cold launch to avoid confusing system message in #mesh
* geo: disable background sampling + notifications\n- Gate sampling to foreground only (beginGeohashSampling, watchers)\n- Suppress geohash activity notifications unless app is active\n- Stop sampling explicitly on background scene phase
* Update BitchatApp.swift
Co-authored-by: asmo <asmogo@protonmail.com>
* Update BitchatApp.swift
Co-authored-by: asmo <asmogo@protonmail.com>
* Update BitchatApp.swift
Co-authored-by: asmo <asmogo@protonmail.com>
* Update bitchat/BitchatApp.swift
Co-authored-by: asmo <asmogo@protonmail.com>
* Update bitchat/BitchatApp.swift
Co-authored-by: asmo <asmogo@protonmail.com>
* fix(iOS App): resolve merge artifacts in scenePhase handler\n- Remove duplicate didEnterBackground state\n- Fix switch/if braces and logic for foreground restart gating
---------
Co-authored-by: jack <jackjackbits@users.noreply.github.com>
Co-authored-by: asmo <asmogo@protonmail.com>
* chat: de-dup private chats across ephemeral/stable IDs; prefer most advanced delivery status\n\n- Fixes LazyVStack duplicate ID warnings and blank row in PM\n- Merges messages by id from ephemeral and Noise-key stores\n- Chooses read > delivered > partiallyDelivered > sent > sending > failed (newer wins on tie)\n- Ensures status icon updates immediately without waiting for another send\n- Adds exhaustive handling for DeliveryStatus in ranking
* logging: reduce noisy info logs to debug; keep errors/warnings\n\n- Downgrade routing/ACK/subscription/connect logs to debug\n- Retain security/fingerprint/keychain info logs\n- Keep errors and warnings intact\n\ndocs: add docs/privacy-assessment.md covering BLE privacy, routing TTL/jitter, Nostr E2E gift wraps, ACK throttling, and logging posture
---------
Co-authored-by: jack <jackjackbits@users.noreply.github.com>
