During room.disconnect(), LiveKit fires ParticipantDisconnected for
each remote participant BEFORE the final Disconnected event. Because
roomRef was still set, guardedUpdate() called updateParticipants(),
which updated the voiceStore while isLiveKitConnected was still true.
SoundController played user_leave for each departing participant
alongside the disconnect sound.
Fix: set roomRef.current = null before calling destroyRoom(). This
causes guardedUpdate() to return early for all teardown events.
The disconnect function handles cleanup after destroyRoom resolves.
The identity flip during disconnect: updateParticipants resolves
homeUserId → localSnowflake when activeDmCall is set, but reverts
to raw homeUserId when activeDmCall is cleared (before LiveKit
disconnect completes). SoundController sees the snowflake "leave"
and the homeUserId "join" — two phantom events for the same person.
Previous fix only checked homeUserId OR id. Now checks BOTH via
isSelf(id) which matches against a Set of {id, homeUserId}. This
recognizes the user as "self" regardless of which identity format
the participant currently has.
Root cause: SoundController compared LiveKit participant p.userId
(which is homeUserId from the home instance) against currentUser.id
(local snowflake on the current instance). For federated users these
are different IDs, so the controller thought the user's own presence
was a stranger — playing user_join/user_leave for self.
Fix: use homeUserId || id for the self-check. This matches the
LiveKit identity format used in federated calls.
Added justDisconnected guard to the participant sound loop. When
the user hangs up, isLiveKitConnected transitions to false — but
in a separate or same subscription tick, the participants list
also empties. Without the guard, SoundController plays user_leave
for every departed participant AND the disconnect sound simultaneously.
Now: if justDisconnected is true, the entire participant loop is
skipped. Only the disconnect sound plays.
RoomEvent.Disconnected handler set participants=[] and
isLiveKitConnected=false in separate setState calls. SoundController
subscription fired between them — saw empty participants while still
"connected" → played user_leave, then saw disconnected → played
disconnect. Both sounds played simultaneously.
Batching into one setState ensures SoundController sees the final
state atomically: participants gone AND disconnected in one update.
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.
Bug A: handleAccept relied on dm_call_accepted server response to set
activeDmCall. But connectFn's async AudioContext resume yields to the
event loop, dm_call_accepted arrives during the yield, finds
isLiveKitConnected=false (connectFn just reset it), and skips
setActiveDmCall. The acceptor connects to LiveKit but the UI never
shows the call. Fix: set activeDmCall and clear incomingCall
directly in the click handler.
Bug B: ready handler no longer sets activeDmCall for active calls.
On refresh/restart the client has no LiveKit connection — showing
"Connecting..." with no connection is broken. The call exists on
the server but this client session is disconnected.
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.
Root cause: sendToFederatedCallUsers used sendToDmMembers when dmChannelId
was set, which broadcast to ALL DM members including the caller's replicated
stub. The caller's multi-instance WS received dm_call_accepted with the
REMOTE instance's dmChannelId, causing token request for a non-existent
channel (403) and preventing the caller from connecting.
Fix 1: sendToFederatedCallUsers always uses ringedUserIds (exact recipients)
instead of sendToDmMembers (all members including caller stub).
Fix 2: dm_call_accepted handler only sets activeDmCall if the client is the
caller (wasOutgoingCall) or already connected to LiveKit. Other instances of
the same user just clear ringing without entering stuck "Connecting..." state.
callOrigin was set to event.callOrigin (the HOST instance URL), which
routed accept/reject through the multi-instance WS connection. On
mobile hotspot or when the multi-instance WS drops, the accept is
silently lost — the host never knows, the call stays ringing forever.
Now callOrigin = origin (the WS that delivered dm_call_incoming).
This is always connected. The server on that instance finds the
FederatedCallEntry and relays to the host via S2S HTTP, which is
reliable and independent of client WS state.
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.
Three fixes for multi-instance call state consistency:
1. Client dm_call_accepted handler only auto-connects to LiveKit if
the user was the caller (outgoingCall was set). Other instances of
the same user just clear ringing state without connecting.
2. Server processDmCallAcceptEvent remote path skips duplicate
broadcast when FederatedCallEntry is already active (prevents
state conflicts from host fan-out arriving after local accept).
3. Ready payload handler clears stuck incomingCall when restoring
an already-active call after page refresh.
Enable federated DM calls to route accept/reject/end through the correct
WebSocket connection using callOrigin, and include federatedCallId in all
dm_call payloads for server-side FederatedCallEntry lookup.
When findOrCreateDmChannel creates or finds a local channel for a
federatedId that has an active FederatedCallEntry with null dmChannelId,
update the entry. Prevents stale null references for Path B calls.
The outbox worker's event reconstruction whitelist was missing these two
fields, causing read_state_update and dm_close/dm_reopen relay events to
arrive at remote instances with empty payloads and get rejected.
Add shared isSelfOrigin() helper that normalizes origins before
comparing to window.location.origin. Fixes auto-connect treating
self-referencing replicatedInstances entries as remote connections,
causing duplicate friends/DMs/data. Also hides self-referencing
entries from the Connections panel UI.
populateFromReady() built the federatedId dedup set from ALL existing
DMs, including those belonging to the reconnecting origin. Incoming DMs
then matched their own stale entries and were skipped as "duplicates."
The subsequent origin-removal step deleted the old copies, leaving no
DMs from that origin in state.
Scope the dedup set to DMs from OTHER origins only, so reconnecting
origins replace their DMs cleanly while cross-instance dedup still works.
Three bugs that combined to corrupt DM identities during initial sync:
1. Sync endpoint omitted federatedId for group DMs, causing the receiver
to treat them as 1-on-1 DMs and compute wrong pair hashes — creating
phantom channels that collide with real ones.
2. backfillHomeUserId unconditionally overwrote existing homeUserIds,
so a single wrong match would permanently corrupt a user's identity
and cascade to all subsequent lookups.
3. Migration duplicate-stub Criterion 1 ("shared 1-on-1 DM membership")
incorrectly merged different users from the same domain who were
simply having a conversation, destroying one user's identity.
The relay broadcast loop skipped members whose homeInstance matched the
source instance, assuming they already received the message on their home
server. This broke delivery for federated users (e.g. youruser@nova on
orbit) who are actively connected to the remote instance.
Client-side dedup in addRealtimeMessage already handles double delivery
via sourceMessageId cross-matching, making the server-side skip both
unnecessary and harmful.
Add federation.md section 8b covering dm_close/dm_reopen relay events:
payload, outbound queueing via queueDmCloseRelay, inbound processDmCloseEvent/
processDmReopenEvent handlers (lookup-only identity resolution, silent
no-ops on missing channel/user/membership), and the processCreateEvent
closed-state reopen bug fix.
Update dm-system.md soft-close section with federation behaviour: relay
to peers for both close and reopen, relayed-message reopen in
processCreateEvent, and the federatedId-only guard for legacy DMs.
Queue dm_close relay in DELETE /api/dm/:id after broadcasting locally,
and queue dm_reopen relay inside the myDm.closed === 1 branch of POST
/api/dm so peers are notified only on actual reopen, not new DM creation.
When a federated user first accesses DMs on a remote instance,
all channels appear unread because no read_states rows exist.
Seed missing read states to the latest message during the ready
payload build. The S2S relay keeps things in sync going forward.
Remove home-only restriction on DM events. Add federatedId
dedup check to dm_channel_created to prevent duplicate sidebar
entries for cross-instance DMs.
Replace home-only DM filter with unified DM list. DMs from all
connected instances are merged, deduplicated by federatedId.
First-loaded copy wins; both copies are in sync via S2S relay.
Receive read_state_update events from peers, translate message
coordinates to local IDs via sourceInstance/sourceMessageId
mapping, update read_states with timestamp-only LWW.
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.
Artworks, .claude, .worktrees, .playwright-mcp, docs/internal,
design prototype, and architecture audit are dev-only — no need
on production servers. packages/desktop was already excluded but
stale copies persisted on remotes.
profileUpdatedAt ?? createdAt treated freshly registered users as having
"newer" profiles than users with real edit history, because createdAt is
always NOW at registration time. This broke federation profile sync:
the client correctly pushed home → remote, but the remote server's LWW
guard rejected the write (stored createdAt > incoming profileUpdatedAt).
A null profileUpdatedAt means "never edited" — that's timestamp 0, not
the user's creation time.
After deleting a federated identity, the server-side user_federation_registry
and users.replicated_instances were not cleaned up, causing "already connected"
errors when trying to re-federate. The deletion endpoint now authoritatively
removes both the registry row and the replicatedInstances entry, and bumps the
LWW timestamp to prevent stale client syncs from re-inserting them.
Also extends the endpoint to accept mode 'leave' (skip S2S, just clean up),
and enables the "Select instances..." scope option in DeleteIdentityDialog.
Add removeRequestsForUser (socialStore), removeUser (discoverStore), and
clearTypingForUser (chatStore) to support isDeleted cleanup in the
user_updated WS handler.
Switch from manual space-ID collection to collectDeletionBroadcastTargets and
add user_updated broadcast so clients patch their caches when a federated user
is deleted via S2S. Force-disconnect moved after broadcasts so other tabs
receive events before the connection is torn down.
- Orphaned DM channels (zero members) are unreachable garbage — clean
them up regardless of purgeContent mode, not just in full/nuke mode.
- Full/nuke mode now also deletes the user's space messages, their
attachments, and embeds. This is the meaningful distinction: "Delete
User" preserves all content as "Deleted User", "Nuke" removes it.
After a prior deletion + re-federation, multiple user records share
the same homeUserId (one deleted, one live). The unfiltered .get()
returned the older deleted record, causing the idempotency check to
short-circuit and miss the live record entirely.
Update the dialog to support three deletion modes (leave/soft/full),
scope selector with disabled "Select instances..." option, loading
state during deletion, and per-instance error handling via toasts.
When a federated user's identity has been tombstoned (isDeleted=1),
findFederatedUser filters them out, causing resolveOrCreateReplicatedUser
to silently create a new stub — a "zombie" resurrection. This guard checks
for a deleted row before creating a stub and returns null instead.
All 11 call sites across federation.ts and dm.ts have been updated with
appropriate null guards: federation relay handlers reject or skip the event
(participant_not_found / accepted no-op), while dm.ts routes convert null
to undefined so the existing 404 path handles it.
The modal read from the global spaceStore.members which only contains
members for the currently active space. Opening the modal via right-click
context menu on a space that hasn't been navigated to yet resulted in an
empty member list. Now fetches members independently via the dedicated
GET /api/spaces/:id/members endpoint with federation-aware API routing.
The modal was manually rendering avatars with a static bg-surface-input
background, bypassing the Avatar component's getAvatarGradient() logic.
Users without profile pictures got blank dark circles instead of their
hash-generated or user-set avatar color gradients.
- Remove tinted description box, use plain text
- Buttons are now flex-1 equal width side by side
- Cancel gets visible bg-interactive-hover background
- Larger padding, rounded-xl, wider max-w
- FriendItem/RequestItem: add plain wrapper div so divide-y border
lands on a non-rounded element (fixes curved separator lines)
- Add border-t to divide-y containers so first row also gets top line
- Remove uppercase from "Direct Messages" in DM sidebar
Add third group for instances the user explicitly disconnected via
the registry. These get a disconnected placeholder with preserved
token for instant reconnect, but no auto-connect on page load.
Aligns the ready payload query with the GET /api/dm approach:
get MAX(created_at) per channel first, then fetch the actual
message rows. Avoids issues with federated relay messages whose
local snowflake IDs don't match chronological order.
Extracts the inline DM list item rendering (~105 lines) from
ChannelSidebar.tsx into a standalone component. Applies cohesive
hover states, 6px border radius, 44px row height, and a selected-
state accent bar matching ChannelItem's pattern.
When the same person registers on multiple federated instances, separate
DM channels are created for each identity, causing duplicate sidebar
entries (e.g., "Nova" appearing twice — once for local, once for
federated account).
Add deduplicateDmChannels() safety net that detects matching counterparts
using two layers: canonicalUserMatch (same ID/homeUserId/username) and
replicatedInstances cross-check (local user has a replicated account on
the federated user's home instance with the same base username). Keeps
the DM with the most recent activity, applied in both setDmChannels and
populateFromReady.
Investigation: queried both live instances (nova.ddns.net, orbit)
and confirmed root cause is same-person-two-accounts, not a server-side
creation bug or identity resolution gap.
Jannis Braun