sendFederatedCallStart now collects per-targeted-peer results and emits
a single dm_call_undeliverable event to the caller when any targeted
peer relay fails. Destroys the local ring room when no plausible
recipient remains (no targeted success + no connected local ringee).
LiveKit pre-flight also emits via this path with reason
'livekit_unavailable' instead of a silent console.warn, closing the
60s hang for unconfigured instances.
Guards against phantom toasts when the caller cancels mid-race by
checking getRoom() before emitting.
sendToFederatedCallUsers sent dm_call_ended/rejected back to the user
who initiated the action. They already disconnected in their click
handler — the redundant event triggered disconnectFn() again, causing
connect and disconnect sounds to play simultaneously.
Added excludeUserId parameter to sendToFederatedCallUsers, used in
handleDmCallEnd and handleDmCallReject Path 2.
sendFederatedCallStart was sending `https://${domain}/livekit` as the
LiveKit URL. The LiveKit SDK requires `wss://` for WebSocket connections.
The caller (local) worked because it gets the URL from config.livekit.url
(wss://). The federated acceptor failed because it used the relay URL
(https://) — the SDK can't connect over HTTPS.
Now uses config.livekit.url directly, falling back to wss:// if unset.
Four fixes addressing the full state management problem:
1. Passive ready handler: no longer auto-connects to LiveKit on
page refresh. Prevents identity conflicts when the same user
has multiple sessions fighting for one LiveKit identity slot.
The user must re-accept to join; state is shown but not acted on.
2. SoundController sync guard: incomingCallLoading/outgoingCallLoading
refs prevent multiple playSound calls during async audio load.
If call is cancelled while sound loads, stops it immediately on
completion. Eliminates the "5 ringtones at once" bug.
3. Host dm_call_accepted broadcasts now include federatedCallId so
all clients (including remote instances) can match the event.
4. Removed all diagnostic console.log statements.
1. Accept/reject/end from remote instance now resolves federatedId
to local dmChannelId via DB lookup, so the host can find its
VoiceRoom when the event arrives with only a federatedCallId.
Previously silently failed with "No active call" error.
2. Batch all dm_call_incoming state updates into a single
useVoiceStore.setState() call. Prevents SoundController from
starting multiple ringtone instances (async playSound guard
race when 4 separate set() calls each triggered the subscription).
3. Always overwrite callOrigin/federatedCallId (with null if absent)
on dm_call_incoming. Prevents stale values from a previous
federated call routing local accepts to the wrong instance.
Queue read_state_update events when users ack DM messages on
channels with a federatedId. Translates local message IDs to
federation coordinates using sourceInstance/sourceMessageId.
sendTypingRelay() mirrors sendCallRelay() — direct POST to peers,
no outbox, no retry. Uses federatedId for cross-instance channel
identification. Wired into handleDmTypingStart() for typing_start
and broadcastDmMessage() for typing_stop.
Broadcasts dm_typing_stop to DM members before dm_message_created,
so the typing indicator clears immediately when a message arrives
instead of lingering for up to 3 seconds after delivery.
- Resolve homeUserId from DB in sendFederatedCallStart/End (not raw userId)
- Clear existing timeout in createFederatedCall before overwriting
- Clear federatedCallToken/Url in leaveVoice and handleForceDisconnect
- Remove unnecessary `as any` cast in relay processor
Prevent malicious peers from forging events attributed to users on other
instances. Every relay event processor now verifies the acting user's
homeInstance (from payload) matches X-Federation-Origin (from HMAC-verified
header) via verifyAttribution(), normalized to bare domain.
- Add verifyAttribution() helper using extractDomain normalization
- Guard all 13 event processors before any user resolution or DB writes
- Add homeInstance to FederationRelayReaction type + outbound payloads
- Replace unnormalized string equality in friend handlers
- Log mismatched values on rejection for debugging
Resolves canonical messageId (sourceMessageId for replicated messages,
local ID for originated messages) and includes messageHomeInstance so
the receiver can deterministically find the target message regardless
of which instance created it.
Reaction outbox events used reactionId (add) or a composite dedup key
(remove) as the event messageId. The receiver looked up dm_messages by
sourceMessageId = event.messageId, which never matched because it was
searching for a reactionId, not the actual message ID.
Fix: include the actual DM messageId in the reaction payload JSON.
The receiver now uses event.reaction.messageId for the lookup, with
fallback to event.messageId for backward compatibility.
The relay was failing because processCreateEvent relied on the friends
table to discover the DM recipient, but friendships aren't federated
across instances. Also, resolveLocalUser matched deleted replicated
users before active ones.
- Add participants[] to FederationRelayEvent with homeUserId/homeInstance
for all DM channel members
- Add getDmParticipants() helper to look up member identities
- Include participants in outbox payloads (create/update) and sync events
- Rewrite processCreateEvent to resolve participants directly, compute
canonicalDmPairId, and findOrCreateDmChannel — removing the entire
friends-list fallback (60+ lines)
- Fix resolveLocalUser to filter out deleted users (is_deleted = 0)
and prefer the replicated user match when multiple candidates exist
Adds device-switch guardrail to handleVoiceJoin: if the user already
has a voice session on a different WebSocket, that old socket receives
a voice_disconnected/displaced event before the new session takes over.
Also binds voiceWs on the caller socket in handleDmCallStart and on
the acceptor socket in handleDmCallAccept.
- Add validateActivities() with full field validation against ACTIVITY_LIMITS
- Add handleActivityUpdate() with showActivity gate and rate limiting
- Enhance handlePresenceUpdate() to include activities in broadcast payload
- Update status cache via connectionManager.setUserStatus() on presence change
- Wire activity_update case into event router switch statement
- Fix ?? to || in metadataFetcher.ts to handle empty strings from Cheerio
- Fix stale embeds on message edit: delete old embeds synchronously before
broadcast, then resolve new ones async (all 4 edit paths: REST+WS, msg+DM)
- Revert unrelated MessageList.tsx scroll threshold change (5000 not 150)
- Remove duplicate embed indexes from migrateAddIndexes (kept standalone ones)
Integrate embed infrastructure into the complete message flow:
- messages.ts: batch-fetch embeds in GET, resolve on POST, re-resolve on PATCH
- dm.ts: same pattern for DM messages with isDm=true
- search.ts: include embeds in all 4 search/around endpoints
- events.ts: embed resolution in WS message create/edit for both space and DM
- Fix embedClassifier.ts type errors (regex match undefined → null)
- Add embeds: [] to all inline MessageWithUser/DmMessageWithUser constructions
- Add embeds: [] to chatStore optimistic message
When a user joins a voice channel while already connected from another
tab, the server now sends a 'displaced' reason on voice_disconnected.
The client tears down LiveKit and shows an informational toast.
Root cause: own messages echoed by the server marked channels unread when
the user had already navigated away. Seven related bugs compounded the
problem — stale read states, missing cleanup on space/DM removal, REST
broadcast ignoring VIEW_CHANNEL, and no validation on channel_ack writes.
Frontend:
- Skip markChannelUnread for the user's own messages (federation-aware)
- Walk backward past temp_ IDs in ackChannel instead of bailing
- Re-fire ack timer when temp message is replaced by server-confirmed ID
- Add removeChannelStates to clean up unread/read/message caches
- Clean up chatStore on removeSpace, removeDmChannel, removeInstanceSpaces
Server:
- Use sendToChannel instead of sendToSpace for REST message creation
- Clean up read_states on space deletion, member kick/leave, and ban
- Validate channel membership before accepting channel_ack writes
- Clean up read_states on DM leave and DM channel deletion
- SSRF protection: DNS resolution + private IP blocking on metadata fetcher
- Upload security: CSP/X-Frame-Options headers, SVG forced download, nosniff
- Auth hardening: JWT secret min length, password min 8 chars, token revocation via password_changed_at
- Attachment ownership verification before linking to messages
- Message length limit (4000 chars) enforced on client and server
- Asset URL validation on avatar/banner updates
- Federation instance validation (domain regex, origin scheme, length limits)
- DB indexes on all FK columns for query performance
- Migrations: nullable moderator columns, dm_messages reply_to FK constraint
- File cleanup on avatar/banner replacement and space deletion
- Fastify trustProxy, AbortController on fetches, typing map size cap
Avatars/banners now resize to max 512px/1920px and convert to WebP before
upload (zero server cost). Chat image uploads generate an 800px-wide WebP
thumbnail via Sharp; the feed shows the thumbnail, click opens the full-res
original. Adds lazy loading to avatars. Federation-compatible: remote
instances without this feature fall back gracefully.
Voice channels already support video/screen share, so the separate video
type was redundant. Adds migration to convert existing video channels.
Also adds border-border-soft to CreateChannel input fields for visibility.
Permission changes now take effect immediately without requiring
disconnect/reconnect. Modeled as "permission mute" parallel to
server mute — server recomputes SPEAK for all voice participants
on role/override changes and broadcasts state via WebSocket.
Includes amber UI indicators and mic toggle blocking.
- Add DISCONNECT_MEMBERS permission (bit 27) to disconnect users from voice
- Implement voice_disconnect WebSocket handler with permission checks
- Add disconnect button to voice user context menu
- Grant instance admins full permissions across all spaces
- Fix voice_disconnected handler to use federation-aware identity resolution
- Update CLAUDE.md with new event types and permission docs
- Replaces global `userId` tracking with `spaceId:userId` composite keys across both backend and frontend, fixing the issue where server-muting a user in one space bled into others.
- Modifies client-side `ready` event hydration to merge voice states per-origin instead of completely overwriting the store, preventing federated connections from wiping out home instance mutes.
- Excludes server voice restrictions from `zustand/persist` so stale client caches don't override the server's authority on reload.
- Fixes React component reactivity by using reactive store selections for `spaceId` instead of imperative `getState()` calls, ensuring UI lockdown indicators accurately reflect the initial websocket handshake.
- Client ready handler now builds restriction Sets atomically in a single
setState call, eliminating intermediate empty-Set state that caused amber
icons to flash rose on reload
- buildReadyPayload queries all voice restrictions from DB across the
user's spaces instead of relying on in-memory state (which is lost on
disconnect timeout)
- voice_join early-return path now re-broadcasts restrictions to the
reconnecting user, covering fast reload where the user is still in the room
Server-side: add DB persistence for voice restrictions (schema, migration,
ready payload, cleanup on leave). Client-side: fix four bugs that wiped or
bypassed server restriction state — leaveVoice() no longer clears global
restriction Sets, voice_state_update leave no longer drops amber icons,
toggleMic/toggleDeafen now guard against server restrictions, and force-mute/
deafen uses direct setState instead of fragile toggle calls.
Jannis Braun