Phase 4: sync

Implements PLAN.md Phase 4: the read model. /sync, /messages and /context,
the filters that narrow them, and the streams a response carries that are
not room events — typing, receipts, presence and to-device messages. With
this a client can be used against Warp, which is milestone M1.

The one architectural choice is that the syncapi has no copy of the rooms:
it reads them through the roomserver. Dendrite gives its syncapi its own
tables and feeds them from the output stream, which is right for a design
where the two can be separate processes, and Dendrite's could be. Warp is
one process by decision, and there the copy buys nothing and costs the
hardest class of bug a homeserver has — a read model that disagrees with
the room, where a client is shown a message that was rejected and the only
way to find out is to compare two tables. So the roomserver grew a read
model (stream.go, roomsync.go, visibility.go) rather than exporting its
store: reading an event means parsing it under its room's version, and a
second copy of the version registry, the parser and the client projection
would be a second place for them to disagree.

syncapi: stream tokens naming a position in each of five independent
streams, so a client that has seen receipt 7 but only event 4 can say so in
one opaque string. A token with fewer positions than this build writes is
accepted and the rest read as zero, which is what lets a token stored
before a stream existed go on working. The PDU position comes from
MAX(stream_ordering) and not from the counter row the other four use: the
counter is advanced inside the writing transaction, so a token taken from
it would name a position whose event no reader can see yet, and the event
would be skipped for good once that transaction committed.

Long polling goes through a notifier that wakes by room and by user rather
than globally — an event wakes the listeners of its room, a membership
event also wakes the user it names, so an invite reaches a client that is
not in the room yet. A wake-up is only ever a hint: the listener re-reads
the streams and decides for itself what changed, which is what makes it
safe for the notifier to hold no state about what each client has seen. The
listener registers before the read, so an event landing in between leaves
the channel signalled and the sync reads again rather than blocking on a
change it just missed. EventPersisted must not block, because it is called
with the room's write lock still held.

"Something happened" is not "a token moved". A stream can move without this
user being concerned by it, and returning on every such move would turn a
long poll back into a busy loop. The same trap caught presence twice over:
a client says it is still online by syncing, so ?set_presence= makes every
sync a presence write, and a stream that moved each time would wake every
peer, which would sync, which would move it again. SetPresence therefore
reports whether anything a peer would notice actually changed, and a write
that only refreshes the last-active timestamp leaves the stream alone.

Timeline windows are named by their two ends and a direction, Low exclusive
and High inclusive in both directions, rather than by "from" and "to" as
the wire format spells them. Each of those would otherwise be the low bound
in one direction and the high bound in the other, which is exactly the
off-by-one every pagination bug is made of; the handlers translate once and
nothing below them thinks about it again. A sync's timeline is read
backwards and reported forwards, because a client wants the newest events
of the window it missed — a forward read with a limit hands it the oldest
and then calls the rest a gap.

visibility: history visibility is now a span rather than a yes or no, which
is what Phase 3 deferred here. A user who has left a room is served the
room as it was when they left — their state reads, their members and their
timeline — instead of being refused, and is bounded above at their own
membership event so a kicked client cannot go on streaming. Their span is
found from their own member events, so a user who joined, left and rejoined
sees their latest span. ?at= on /members is honoured for the same reason.
What is left is that the span is computed from the room's current
history_visibility where the spec judges each event by the one in force
when it was sent; that is a state read per event and belongs with open
question 1.

Typing is held in memory and nowhere else: notifications expire in seconds,
are worthless across a restart, and arrive one per keystroke-burst per
room, so persisting them would be the busiest table in the database holding
the least valuable data in it. Receipts only ever advance, compared on the
position of the event acknowledged rather than on arrival, because two of a
user's clients read at different speeds and the slower one must not drag
the marker backwards. To-device queues are per device and drained by the
next sync rather than by the one that delivered them, so a client that
crashes in between gets the batch again — the right way round, since a lost
key share leaves a message permanently undecryptable and a duplicate is
ignored. A wildcard addressee is expanded when sent, not stored, because
"every device" means the ones that existed then.

Presence is off by default, which answers PLAN.md open question 3. It is
the noisiest thing in Matrix and the one genuinely fan-out notification
here; a private homeserver gains little from it. With it off the endpoints
refuse rather than disappearing, because a client that gets a 404 from
/presence concludes the server is too old to have it and stops asking.

storage: one migration per driver adding the receipt, to-device, presence
and filter tables with a counter row each, and a stream position on the
existing account data table rather than a parallel one — there is exactly
one current value per (user, room, type), and a second table would be a
second thing to keep in step. Account data is reported in an event list but
is not an event, so it gets its own response type: a room event with an
empty sender and no timestamp would make a client decide what that meant.

Deliberately missing, each named in docs/sync.md: notification counts (they
need push rules, Phase 8), device lists (the keyserver, Phase 7), anything
remote (Phase 5), threads (Phase 10). device_lists is sent as two empty
lists because a client reads a missing one as malformed; the notification
counts are left out entirely, because a zero there is a claim and an absent
field is not.

Tested through HTTP rather than against the builder, because what is worth
pinning down is the response a client sees — which section a room lands in,
whether a token handed back can be handed in again, whether paging
terminates and loses nothing. The storage tests run against both SQL
drivers; the notifier has its own tests for the concurrent case. A live
server was driven end to end, which is what caught the two bugs the unit
tests had missed: a backwards /messages page must be in reverse
chronological order, and `start` must name where the page began rather than
position zero when ?from= was omitted.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

server/internal/syncapi/notifier.go

@@ -0,0 +1,207 @@
+package syncapi
+
+import (
+	"context"
+	"sync"
+
+	"github.com/dariuszjaniak/warp/server/internal/matrix/event"
+	"github.com/dariuszjaniak/warp/server/internal/roomserver"
+)
+
+// Notifier wakes the long-polling syncs that something concerns.
+//
+// It is the other half of the roomserver's emit stage: the input pipeline
+// is the only thing that knows an event has landed, and a client blocked in
+// /sync for thirty seconds has to hear about it in milliseconds rather than
+// when its timeout expires. Polling the database instead would mean every
+// connected client querying every second, which is the cost this avoids.
+//
+// Waking is by room and by user rather than globally. A sync registers the
+// rooms it just read plus its own user ID; an event wakes the listeners of
+// its room, and a membership event also wakes the listener of the user it
+// names — which is how an invite reaches a client that is not in the room
+// yet and so is not listening for it. An event in a room nobody is polling
+// wakes nobody.
+//
+// A wake-up is only ever a hint. The listener re-reads the streams from the
+// database and works out for itself what changed, so a spurious wake-up
+// costs one query and a missed one costs at most the sync timeout. That is
+// what makes it safe for this to hold no state about what each client has
+// seen.
+type Notifier struct {
+	mu    sync.Mutex
+	rooms map[string]map[*Listener]struct{}
+	users map[string]map[*Listener]struct{}
+}
+
+// NewNotifier returns a notifier with no listeners.
+func NewNotifier() *Notifier {
+	return &Notifier{
+		rooms: map[string]map[*Listener]struct{}{},
+		users: map[string]map[*Listener]struct{}{},
+	}
+}
+
+// Listener is one blocked sync's wake-up channel.
+type Listener struct {
+	notifier *Notifier
+	wake     chan struct{}
+
+	userID string
+	rooms  []string
+}
+
+// Listen registers interest in a user and a set of rooms.
+//
+// The caller must Close the listener. Registering before reading the
+// streams is what closes the race: an event that lands between the
+// registration and the read leaves the channel already signalled, so the
+// following Wait returns at once and the sync reads again rather than
+// blocking on a change it just missed.
+func (n *Notifier) Listen(userID string, roomIDs []string) *Listener {
+	l := &Listener{
+		notifier: n,
+		// One slot, so a wake-up is never lost and never queued: the
+		// listener re-reads everything anyway, so two wake-ups and one are
+		// the same instruction.
+		wake:   make(chan struct{}, 1),
+		userID: userID,
+		rooms:  roomIDs,
+	}
+
+	n.mu.Lock()
+	defer n.mu.Unlock()
+	if userID != "" {
+		add(n.users, userID, l)
+	}
+	for _, roomID := range roomIDs {
+		add(n.rooms, roomID, l)
+	}
+	return l
+}
+
+// Wait returns the channel a wake-up arrives on.
+func (l *Listener) Wait() <-chan struct{} { return l.wake }
+
+// Close deregisters the listener. It is safe to call more than once.
+func (l *Listener) Close() {
+	if l == nil || l.notifier == nil {
+		return
+	}
+	n := l.notifier
+	l.notifier = nil
+
+	n.mu.Lock()
+	defer n.mu.Unlock()
+	if l.userID != "" {
+		remove(n.users, l.userID, l)
+	}
+	for _, roomID := range l.rooms {
+		remove(n.rooms, roomID, l)
+	}
+}
+
+// EventPersisted implements roomserver.Notifier: it is called once per
+// event the input pipeline has committed, in the order they were committed.
+//
+// It must not block. The call happens with the room's write lock still
+// held, so anything slow here would serialise the room's writers behind the
+// slowest reader of it — which is why waking is a non-blocking send onto a
+// one-slot channel and nothing else.
+func (n *Notifier) EventPersisted(_ context.Context, notification roomserver.Notification) {
+	n.NotifyRoom(notification.RoomID)
+
+	// A membership event also concerns the user it names, who for an
+	// invite is not in the room and so is not listening for it.
+	e := notification.Event
+	if e == nil || e.Type() != event.TypeMember {
+		return
+	}
+	if stateKey, ok := e.StateKey(); ok && stateKey != "" {
+		n.NotifyUser(stateKey)
+	}
+}
+
+// NotifyRoom wakes the listeners of one room. It is what the ephemeral
+// streams — typing and receipts — use, since those are about a room but are
+// not events.
+func (n *Notifier) NotifyRoom(roomID string) {
+	n.mu.Lock()
+	defer n.mu.Unlock()
+	wakeAll(n.rooms[roomID])
+}
+
+// NotifyUser wakes the listeners of one user: their to-device messages,
+// their account data and the invites addressed to them.
+func (n *Notifier) NotifyUser(userID string) {
+	n.mu.Lock()
+	defer n.mu.Unlock()
+	wakeAll(n.users[userID])
+}
+
+// NotifyUsers wakes several users at once, under one lock.
+func (n *Notifier) NotifyUsers(userIDs []string) {
+	n.mu.Lock()
+	defer n.mu.Unlock()
+	for _, userID := range userIDs {
+		wakeAll(n.users[userID])
+	}
+}
+
+// Listeners reports how many syncs are currently blocked, for metrics and
+// for the tests that would otherwise have to guess.
+func (n *Notifier) Listeners() int {
+	n.mu.Lock()
+	defer n.mu.Unlock()
+
+	seen := map[*Listener]struct{}{}
+	for _, set := range n.rooms {
+		for l := range set {
+			seen[l] = struct{}{}
+		}
+	}
+	for _, set := range n.users {
+		for l := range set {
+			seen[l] = struct{}{}
+		}
+	}
+	return len(seen)
+}
+
+// wakeAll signals every listener in a set without blocking on any of them.
+func wakeAll(set map[*Listener]struct{}) {
+	for l := range set {
+		select {
+		case l.wake <- struct{}{}:
+		default:
+			// Already signalled and not yet read: the listener will see
+			// this change when it acts on the signal it has.
+		}
+	}
+}
+
+func add(index map[string]map[*Listener]struct{}, key string, l *Listener) {
+	set, ok := index[key]
+	if !ok {
+		set = map[*Listener]struct{}{}
+		index[key] = set
+	}
+	set[l] = struct{}{}
+}
+
+func remove(index map[string]map[*Listener]struct{}, key string, l *Listener) {
+	set, ok := index[key]
+	if !ok {
+		return
+	}
+	delete(set, l)
+	if len(set) == 0 {
+		// An empty set left behind would make the index grow with every
+		// room any client has ever polled.
+		delete(index, key)
+	}
+}
+
+// Compile-time proof that the notifier is what the roomserver's emit stage
+// expects.
+var _ roomserver.Notifier = (*Notifier)(nil)