Commit graph

18 commits

Author SHA1 Message Date
Scott Reimers
60463d1817 Phase 2d (0.6.1-beta): route manifest + blob ops through file_holders
Switch ALL propagation-decision reads to the flat holder set.

push_manifest_to_downstream now targets file_holders instead of
blob_downstream. ManifestPush receive-side relay likewise — known
holders fan out to up to 5 most-recent peers instead of a directional
tree.

Blob delete notices: single flat fan-out to file_holders; the legacy
upstream_node tree-healing field is emitted as None (wire-stable via
serde default) and ignored on receive — the post-0.6 flat model
doesn't need sender-role distinction. send_blob_delete_notices keeps
its Option<&Upstream> parameter as an unused placeholder for signature
stability with the call sites in this commit.

Other reads migrated:
- blob fetch cascade: step 2 now tries "known holders" (up to 5)
  instead of a single upstream
- manifest refresh: downstream_count reported from file_holder_count
- web/http post holder enumeration
- Worm search post/blob holder fallback (both connection.rs paths)
- DeleteRecord fan-out rewires to file_holders
- Under-replication replication check: < 2 holders

Storage additions:
- get_file_holder_count(file_id)
- remove_file_holder(file_id, peer_id)

Legacy upstream/downstream writes are still happening from Phase 2b;
those + the tables themselves go in 2e.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-21 21:09:45 -04:00
Scott Reimers
3a0d2e93ab Phase 2c (0.6.1-beta): route engagement diffs through file_holders
propagate_engagement_diff now targets the post's flat holder set (up to
5 most-recent) instead of the post_downstream directional tree. The
holder set naturally subsumes the old upstream+downstream partition, so
the separate "also send to upstreams" loops at each engagement call
site are removed (reactions, comments, comment edit/delete, receipt
slots, comment slots).

handle_blob_header_diff on receive:
- records the sending peer as a file holder (an engagement exchange is
  proof the peer holds the post)
- re-propagates to the holder set minus the sender

Writes to post_upstream / post_downstream still occur from Phase 2b
(dual-write); those and the legacy tables will be removed in 2e.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-21 21:00:53 -04:00
Scott Reimers
0b2b4f5a68 Phase 2b (0.6.1-beta): dual-write file_holders on all propagation events
Populate the flat holder set alongside every existing post_upstream /
post_downstream / blob_upstream / blob_downstream write so that read
paths can be switched over in the next commit without losing continuity.

Events wired:
- Pull sync receive (3 paths in connection.rs)
- PostPush receive (public posts only after Phase 1)
- PostFetch via notification (discovery pull)
- PostDownstreamRegister
- Replication accept (downstream) + replication-driven pull (upstream)
- Attachment upstream recorded after replication blob fetch
- ManifestPush receive (remote is a CID holder)
- ManifestPush send (downstream peer becomes CID holder)
- Blob fetch fallback (upstream lateral sources)

Direction is tracked as Received vs Sent. Not load-bearing for routing;
retained for future use. LRU cap of 5 enforced on every touch.

Legacy upstream/downstream writes remain in place; they'll go away
together with the table drops at the end of this phase.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-21 20:56:28 -04:00
Scott Reimers
e6265b52b6 Phase 1 (0.6.0-beta): remove direct PostPush for encrypted posts
Encrypted posts now propagate only via the CDN (ManifestPush + neighbor
header updates), eliminating the sender→recipient traffic signal on the
wire. Encrypted DMs are indistinguishable from any other encrypted post.

- Remove push_post_to_recipients entirely from network.rs
- Remove call sites in create_post and re-encrypt-on-revoke
- PostPush handler now ignores non-public visibility (kept for public
  audience push path)

Known gap: non-follower DMs won't reach until Phase 3 (merged pull +
recipient-match). Followers receive via the existing CDN path — new
posts trigger neighbor-manifest updates, ManifestPush fans out to
downstream holders, recipients pull missing post IDs from followed
authors.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-21 20:46:34 -04:00
Scott Reimers
288b53ffb1 Feed pagination, duplicate identity detection, pkarr leak fix, Android SAF
Feed pagination:
- Cursor-based pagination: get_feed_page/get_all_posts_page (20 posts/page)
- Batched engagement queries (3 bulk SQL queries instead of 4 per post)
- IntersectionObserver for infinite scroll (sentinel at midpoint)
- Viewport-based media loading (blobs only load when post enters view)
- Pre-fetch next page immediately after current page renders

Duplicate identity detection:
- Anchor detects when a NodeId is already mesh-connected during initial
  exchange and sets duplicate_active flag in response
- Client skips sync tasks when duplicate detected
- Frontend shows red warning banner

Privacy:
- Fixed pkarr leak: clear_address_lookup() removes default dns.iroh.link
  publishing. Only mDNS (local network) discovery enabled.

Android:
- SAF integration via tauri-plugin-android-fs: exports open native "Save As"
  dialog so users can save to Downloads/Drive/etc.
- Download/export paths use app data dir on Android (writable)
- File picker gated behind desktop cfg (blocking_pick not on Android)

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-18 15:35:23 -04:00
Scott Reimers
cba30a1bb3 AppImage video fix, proper import posts, first-run chooser, file pickers
- bundleMediaFramework: true — bundles full GStreamer plugin set in AppImage.
  Fixes WebKit hang on video/audio (missing appsink/autoaudiosink plugins).
  MSDK/VA plugins removed post-build to avoid Ubuntu DMA assertion crash.
- Import creates complete posts: BlobHeader, VisibilityIntent, pinned blobs,
  self-follow. Imported posts now indistinguishable from locally created ones.
- First-run chooser: Start Fresh or Import Identity on fresh install only.
  Profile setup shown for existing identities without display name.
- File pickers: native Browse buttons on export (folder) and import (ZIP)
  via tauri-plugin-dialog.
- Export path: relative paths resolved against home dir.
- Lightbox close: only on overlay/image click, not inner form content.
- Growth loop skips self as N2 candidate.
- Node shutdown on identity switch (prevents zombie background tasks).
- media-src CSP includes asset protocol.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-16 13:49:52 -04:00
Scott Reimers
ec731fdb4b First-run chooser, node shutdown on switch, file picker, export path fix
- First-run: show Start Fresh / Import chooser only when single auto-created
  identity with no profile (not on every boot without a display name).
- Identity switch: shut down old node's endpoint before starting new one.
  Fixes lockup after multiple switches (zombie background tasks).
- File picker: native Browse buttons on export (folder) and import (ZIP file)
  via tauri-plugin-dialog.
- Export path: resolve relative paths against home dir (was using process cwd).
- Lightbox close: only close on overlay/image click, not inner form content.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-06 01:58:02 -04:00
Scott Reimers
be253e8001 Our Info panel, hole punch race fix, NAT profiles in relay introduction
- Network Diagnostics: "Our Info" button shows addresses with NAT status,
  device role, UPnP, HTTP capability. Addresses stacked for mobile.
- Hole punch race: re-check for existing connection before and after relay
  introduction to avoid wasting minutes on redundant punch attempts.
- Relay introduction now carries requester/target NAT mapping+filtering
  so hole punch strategy uses fresh profiles instead of stale stored ones.
  Critical for phones that switch between WiFi/cellular/VPN.
- STUN fix: filter DNS results to IPv4 (was resolving to IPv6 first on
  dual-stack, causing silent send failure and "NAT unknown").
- Welcome screen: Ready button with loading bar for instant feed access.
- LAN addresses show just "LAN" (no misleading punchability label).

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-05 17:57:41 -04:00
Scott Reimers
fb1e92985c Audit fixes: key permissions, lock contention, Docker IP filter, doc updates
Security: identity.key written with 0600 permissions (Unix). Docker bridge
IPs (172.17-31.x) filtered from is_shareable_addr to prevent topology
disclosure in relay introductions.

Lock contention: ManifestPush relay and DeleteRecord CDN notify now gather
connections under lock, then send outside lock.

UI: syncBtn null guard prevents crash on hidden element.

Documentation: design.html version badge updated to v0.4.4. Self Last
Encounter threshold corrected from 3h to 4h. Multi-Device Identity section
rewritten for multi-identity-per-device (complete) + multi-device (planned)
+ post merge (planned). MEMORY.md updated to v0.4.4+ status.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-31 19:37:17 -04:00
Scott Reimers
1030dc21a7 Sync fixes, hole punch address family filtering, ManifestPush blob fetch
Per-peer sync (People tab Sync button) now resets last_sync_ms to 0 so
the responder sends ALL posts, not just posts newer than last sync.

ManifestPush post discovery now fetches blobs alongside discovered posts
while the connection is live, instead of waiting for the next prefetch cycle.

Hole punch: filter_reachable_families() checks endpoint bound sockets and
removes addresses the local endpoint can't reach (IPv4-only won't try IPv6).
Applied to hole_punch_single, hole_punch_parallel, hole_punch_with_scanning.

Relay introduce paths switched from is_publicly_routable to is_shareable_addr
so LAN addresses (192.168.x.x) are included for same-WiFi hole punching.
is_shareable_addr made pub(crate).

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-31 17:00:43 -04:00
Scott Reimers
409e44762a Clean up warnings: remove dead code, unused imports
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-22 23:38:18 -04:00
Scott Reimers
7b9edc13da Fix blocking_lock panic inside async runtime, reduce pool to 4 connections
spawn_with_arc used blocking_lock() which panics inside tokio runtime.
Changed to async lock().await. Reduced StoragePool from 8 to 4 connections
for Android compatibility.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-22 22:10:50 -04:00
Scott Reimers
43adbbdf7d v0.4.3: Lock contention overhaul, StoragePool, mobile bottom nav, text scaling
Eliminate all conn_mgr lock holds during network I/O across 14 actor commands
and bi-stream handlers. PostFetch, TcpPunch, PullFromPeer, FetchEngagement,
ResolveAddress, AnchorProbe use brief locks for data gathering only. WormLookup,
ContentSearch, WormQuery use connection snapshots for lock-free cascade fan-out.
RelayIntroduce extracts forwarding data under brief lock, does I/O outside.
BlobRequest, PostFetchRequest, ManifestRefresh use Arc clones instead of conn_mgr
lock. ConnectionActor hoists shared Arcs (storage, blob_store, endpoint) for
lock-free access. ResolveAddress adds 5s per-query timeout (was unbounded).

Initial exchange failure now aborts mesh upgrade (was silently continuing with
broken connection). connect_to_peer/connect_to_anchor use consistent 15s timeout.
Rebalance connects outside the lock via pending_connects pattern.

StoragePool: 8 concurrent SQLite connections in WAL mode replace single
Mutex<Storage>. Reads run fully parallel; writes serialize at SQLite level only.
PRAGMA busy_timeout=5000 for graceful write contention.

Mobile bottom nav bar (<=768px) with icon tabs. Text sizes: XS/S/M/L/XL
(75%/100%/125%/150%/200%), default M. localStorage persistence for instant
restore. Toast repositioned above mobile nav.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-22 21:35:38 -04:00
Scott Reimers
bbaacf9b6c v0.4.0: Protocol v4 — header-driven sync, tiered engagement, multi-upstream
Protocol v4 sync overhaul:
- Slim PullSyncRequest: per-author timestamps (since_ms) replace full post ID lists
  Request size O(follows) instead of O(posts). Backward-compatible via serde default.
- Tiered pull frequency: 60s ticks, only syncs stale authors (4hr default)
  Full pull only on first tick (bootstrap). Most ticks skip — no stale authors.
- Tiered engagement checks: frequency scales with content age
  5min (<72h), 1hr (3-14d), 4hr (14-30d), 24hr (>30d)
  Single SQL query filters posts due for check.
- Header-driven post discovery: ManifestPush triggers PostFetch for missing
  followed-author posts (capped 10 per manifest). CDN tree = notification system.
- Multi-upstream (3 max): composite PK, priority ordering, engagement diffs
  sent to all upstreams, promote/remove on failure.

DB schema:
- follows.last_sync_ms — Self Last Encounter per author
- posts.last_engagement_ms — last reaction/comment timestamp
- posts.last_check_ms — last engagement check timestamp
- post_upstream: single-row → 3-row with priority column

Lock contention fixes:
- get_blob_for_post: 3 locks → 1
- prefetch_blobs_from_peer: lock-free blob checks
- fetch_engagement_from_peer: explicit lock release before I/O
- serve_post: 4 locks → 2 (eliminated redundant queries)
- run_replication_check: 2 locks → 1
- Badge cycle: N+2 IPC calls → 1 (get_badge_counts)

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-21 16:13:45 -04:00
Scott Reimers
24b78a8d41 v0.3.6: Network indicator, tab badges, message read tracking, UI cleanup
UI:
- Network indicator dot (black/red/yellow/green) + capability labels (Public, Server)
- Tab badges: Feed (new posts), My Posts (new engagement), People (online), Messages (unread)
- Stats bar removed — contextual counts in tab labels
- Message thread popup variable scoping fix

Message read tracking:
- mark_conversation_read on popover open, close, and message send
- Prevents re-notification of already-seen messages after app restart

Network:
- Added has_public_v6(), has_upnp() getters on Network
- NetworkSummaryDto includes hasPublicV6, hasPublicV4, hasUpnp

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-20 23:09:57 -04:00
Scott Reimers
a7e632de88 v0.3.6: Active CDN replication, device roles, budgets, tombstones, engagement fix, DOS hardening
Active CDN replication:
- All devices proactively replicate recent posts (<72h, <2 replicas) to peers
- Target priority: desktops (300) > anchors (200) > phones (100) + cache_pressure
- ReplicationRequest/Response (0xE1/0xE2) wire messages
- 10-min cycle, 2-min initial delay, cap 20 posts per request
- Graceful with small networks (1 peer = 1 replica, 0 peers = silent skip)

Device roles & budgets:
- Intermittent (phone), Available (desktop), Persistent (anchor)
- Advertised in InitialExchange, stored per-peer
- Replication budget: phones 100MB/hr, desktops/anchors 200MB/hr
- Delivery budget: phones 1GB/hr, desktops 2GB/hr, anchors 1GB/hr
- Hourly auto-reset, enforcement on blob serving

Cache management:
- 1GB default cache limit, configurable in settings UI
- Eviction cycle activated (was implemented but never started)
- Share-link priority boost (+100 for 3+ downstream)
- Cache pressure score (0-255) for replication targeting

Engagement distribution fix:
- BlobHeader JSON rebuilt after BlobHeaderDiff ops
- Previously reactions/comments stored in tables but header stayed stale

Tombstone system:
- deleted_at column on reactions and comments
- Tombstones propagate through pull sync (additive merge respects timestamps)
- UI queries filter WHERE deleted_at IS NULL

Persistent notifications:
- seen_engagement and seen_messages tables replace in-memory Sets
- Only notify on genuinely unseen content, survives restarts

DOS hardening:
- BlobHeaderDiff fan-out: single batched task, max 10 concurrent via JoinSet
- Blob prefetch: cap 20 per cycle, newest first
- PostDownstreamRegister: cap 50 per sync
- Delivery budget enforcement on BlobRequest handler
- Pull preference: non-anchors first to preserve anchor delivery budget

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-20 21:00:28 -04:00
Scott Reimers
8fad30cf95 v0.3.3: Rate limiting, IPv6 fix, schema versioning, video preload, engagement propagation
Security & stability:
- Incoming auth-fail rate limiting per source IP (3 attempts, then exponential backoff)
- Schema versioning via PRAGMA user_version with migration framework

Networking:
- IPv6 http_addr fix: advertise actual public IPv6 instead of 0.0.0.0
- N2/N3 TTL reduced from 7 days to 5 hours
- Full N1/N2 state re-broadcast every 4 hours
- Bootstrap isolation recovery: 24h check with sticky N1 advertising
- Bidirectional engagement propagation (upstream + downstream)
- Auto downstream registration on pull sync and push notification
- post_upstream table for CDN tree traversal

Media & UI:
- Video preload="auto" for share links and in-app blob URLs
- Following: Online/Offline split with last-seen timestamps
- DMs filtered from My Posts tab
- Image lightbox, audio player, file attachments with download prompt
- Share link unroutable address filtering

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-16 18:37:24 -04:00
Scott Reimers
800388cda4 ItsGoin v0.3.2 — Decentralized social media network
No central server, user-owned data, reverse-chronological feed.
Rust core + Tauri desktop + Android app + plain HTML/CSS/JS frontend.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-15 20:23:09 -04:00