Files
Warp/server/internal/storage/federation_test.go
T
tajniak81andClaude Opus 5.5 6aaf554336
CI / Test (ubuntu-latest) (push) Canceled after 0s
CI / Test (windows-latest) (push) Canceled after 0s
CI / Storage tests on PostgreSQL (push) Canceled after 0s
CI / Complement (federation off) (push) Canceled after 0s
CI / golangci-lint (push) Canceled after 0s
CI / Smoke test the binary (push) Canceled after 0s
Phase 5: federation
Implements PLAN.md Phase 5: the Server-Server API. Warp joins rooms on
other servers and lets their users join its rooms, invites across servers
with the invitee's server countersigning, rejects remote invites (locally
if the inviter is unreachable), and exchanges events, typing, receipts and
to-device messages with other homeservers — all behind the runtime switch,
which now starts and stops real outbound traffic.

The protocol lives in internal/federation with no knowledge of rooms:
X-Matrix signing (fuzzed parser), the spec's discovery algorithm keeping
the dial address and the TLS name apart, a signed client that refuses
every request while the switch is off or the policy denies the
destination, and a key ring that caches other servers' keys for at most a
week, with optional notaries.

The roomserver takes events from elsewhere through the spec's intake:
signatures and hashes (a hash mismatch keeps the redacted form), the rules
against the event's own auth events, against the state before it — with
/get_missing_events for small gaps and /state_ids for wide ones — and
against the current state, where a failure is a soft failure. Events held
without their state are outliers: a column on roomserver_events, because
the current state of a room joined over federation is almost all
outliers, and excluded from every timeline query. Forks are resolved at
the room's head. Membership across servers is make_join/send_join,
make_leave/send_leave and countersigned invites, in both directions;
m.room.server_acl is enforced on everything another server sends.
History before a remote join is backfilled when /messages pages past it,
stored at negative stream positions.

internal/federationapi is the running component: the inbound endpoints,
and an outbound queue that persists what it owes each destination, sends
it in transactions of 50 PDUs and 100 EDUs, backs off from ten seconds to
six hours, survives restarts, and drops events persisted while federation
is off rather than delivering a backlog in one burst. Its Module is what
the switch starts and stops.

Also: a TLS listener for port 8448, trusted key servers, key validity and
an outbound CA bundle in the config; per-destination health and backoff
reset in the admin API; remote profiles, aliases and directories in the
Client-Server API; a Complement federation profile. See
docs/federation.md, including its known gaps (no partial-state joins, no
key rotation).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-16 17:14:20 +02:00

432 lines
15 KiB
Go

package storage_test
import (
"database/sql"
"os"
"slices"
"testing"
"github.com/dariuszjaniak/warp/server/internal/storage"
)
// migratedFederationDB opens a driver with the room and federation tables
// empty.
func migratedFederationDB(t *testing.T, driver string) storage.Database {
t.Helper()
db := migratedRoomDB(t, driver)
if driver == storage.DriverPostgres {
pool, err := sql.Open("pgx", os.Getenv("WARP_TEST_POSTGRES_DSN"))
if err != nil {
t.Fatalf("open postgres to reset the federation tables: %v", err)
}
defer func() { _ = pool.Close() }()
if _, err := pool.ExecContext(t.Context(), `TRUNCATE
federation_server_keys, federation_queue,
federation_destinations, federation_inbound_txns`); err != nil {
t.Fatalf("reset the federation tables: %v", err)
}
if _, err := pool.ExecContext(t.Context(), `UPDATE federation_sequences SET value = 0`); err != nil {
t.Fatalf("rewind the federation counters: %v", err)
}
}
return db
}
func TestOutliersAreReadableButNotInTheTimeline(t *testing.T) {
for _, driver := range testDrivers {
t.Run(driver, func(t *testing.T) {
ctx := t.Context()
rooms := migratedFederationDB(t, driver).Rooms()
if err := rooms.CreateRoom(ctx, newTestRoom(testRoomID)); err != nil {
t.Fatalf("CreateRoom: %v", err)
}
outliers := []storage.StoredEvent{
newTestEvent(testRoomID, "$create", "m.room.create", stateKey("")),
newTestEvent(testRoomID, "$member", "m.room.member", stateKey("@bob:remote.example")),
}
if err := rooms.WriteOutliers(ctx, testRoomID, outliers); err != nil {
t.Fatalf("WriteOutliers: %v", err)
}
// Writing them again is not an error: an event ID names its
// contents, so the second copy is the first one.
if err := rooms.WriteOutliers(ctx, testRoomID, outliers); err != nil {
t.Fatalf("WriteOutliers twice: %v", err)
}
got, err := rooms.EventsByID(ctx, testRoomID, []string{"$create", "$member"})
if err != nil {
t.Fatalf("EventsByID: %v", err)
}
if len(got) != 2 {
t.Fatalf("EventsByID found %d outliers, want 2", len(got))
}
for _, e := range got {
if !e.Outlier {
t.Errorf("%s came back as a timeline event", e.EventID)
}
}
// A join that cites the outliers as its state.
join := newTestEvent(testRoomID, "$join", "m.room.member", stateKey("@alice:warp.example"))
if _, err := rooms.Write(ctx, storage.RoomWrite{
Event: join,
StateBefore: []storage.StateEntry{
{Type: "m.room.create", EventID: "$create"},
{Type: "m.room.member", StateKey: "@bob:remote.example", EventID: "$member", Membership: "join"},
},
StateDelta: []storage.StateEntry{
{Type: "m.room.create", EventID: "$create"},
{Type: "m.room.member", StateKey: "@bob:remote.example", EventID: "$member", Membership: "join"},
{Type: "m.room.member", StateKey: "@alice:warp.example", EventID: "$join", Membership: "join"},
},
StateDeltaIsWholeState: true,
Extremities: []string{"$join"},
}); err != nil {
t.Fatalf("Write join: %v", err)
}
timeline, err := rooms.Timeline(ctx, storage.TimelineRequest{
RoomID: testRoomID, Limit: 10, Direction: storage.DirectionBackward,
})
if err != nil {
t.Fatalf("Timeline: %v", err)
}
if ids := eventIDsOf(timeline); !slices.Equal(ids, []string{"$join"}) {
t.Errorf("timeline = %v, want only the join", ids)
}
// The current state reads through to the outliers: that is
// the whole point of keeping them in the events table.
state, err := rooms.CurrentStateEvents(ctx, testRoomID)
if err != nil {
t.Fatalf("CurrentStateEvents: %v", err)
}
if len(state) != 3 {
t.Errorf("current state has %d events, want 3", len(state))
}
top, err := rooms.MaxStreamOrdering(ctx)
if err != nil {
t.Fatalf("MaxStreamOrdering: %v", err)
}
at, err := rooms.StateAtStream(ctx, testRoomID, top)
if err != nil {
t.Fatalf("StateAtStream: %v", err)
}
if len(at) != 3 {
t.Errorf("state at the join has %d entries, want 3", len(at))
}
})
}
}
func TestOutlierIsPromotedWhenItArrivesAgainAsAnEvent(t *testing.T) {
for _, driver := range testDrivers {
t.Run(driver, func(t *testing.T) {
ctx := t.Context()
rooms := migratedFederationDB(t, driver).Rooms()
if err := rooms.CreateRoom(ctx, newTestRoom(testRoomID)); err != nil {
t.Fatalf("CreateRoom: %v", err)
}
writeMessages(t, rooms, testRoomID, "$first")
if err := rooms.WriteOutliers(ctx, testRoomID, []storage.StoredEvent{
newTestEvent(testRoomID, "$late", "m.room.message", nil),
}); err != nil {
t.Fatalf("WriteOutliers: %v", err)
}
writeMessages(t, rooms, testRoomID, "$late")
stored, err := rooms.Event(ctx, testRoomID, "$late")
if err != nil {
t.Fatalf("Event: %v", err)
}
if stored.Outlier {
t.Error("the event is still an outlier after being written as one")
}
timeline, err := rooms.Timeline(ctx, storage.TimelineRequest{
RoomID: testRoomID, Limit: 10, Direction: storage.DirectionForward,
})
if err != nil {
t.Fatalf("Timeline: %v", err)
}
if ids := eventIDsOf(timeline); !slices.Equal(ids, []string{"$first", "$late"}) {
t.Errorf("timeline = %v, want [$first $late]", ids)
}
})
}
}
func TestInviteStateLastsUntilTheNextMembershipChange(t *testing.T) {
for _, driver := range testDrivers {
t.Run(driver, func(t *testing.T) {
ctx := t.Context()
rooms := migratedFederationDB(t, driver).Rooms()
if err := rooms.CreateRoom(ctx, newTestRoom(testRoomID)); err != nil {
t.Fatalf("CreateRoom: %v", err)
}
const invitee = "@alice:warp.example"
stripped := []byte(`[{"type":"m.room.name","state_key":"","sender":"@bob:remote.example","content":{"name":"Hi"}}]`)
invite := newTestEvent(testRoomID, "$invite", "m.room.member", stateKey(invitee))
if _, err := rooms.Write(ctx, storage.RoomWrite{
Event: invite,
Outlier: true,
StateDelta: []storage.StateEntry{{Type: "m.room.member", StateKey: invitee, EventID: "$invite", Membership: "invite"}},
InviteState: stripped,
}); err != nil {
t.Fatalf("Write invite: %v", err)
}
got, err := rooms.InviteState(ctx, testRoomID, invitee)
if err != nil {
t.Fatalf("InviteState: %v", err)
}
if string(got) != string(stripped) {
t.Errorf("InviteState = %s, want %s", got, stripped)
}
memberships, err := rooms.MembershipsForUser(ctx, invitee, []string{"invite"})
if err != nil {
t.Fatalf("MembershipsForUser: %v", err)
}
if len(memberships) != 1 || memberships[0].StreamOrdering == 0 {
t.Errorf("an outlier invite must still have a stream position for /sync, got %+v", memberships)
}
leave := newTestEvent(testRoomID, "$leave", "m.room.member", stateKey(invitee))
if _, err := rooms.Write(ctx, storage.RoomWrite{
Event: leave,
Outlier: true,
StateDelta: []storage.StateEntry{{Type: "m.room.member", StateKey: invitee, EventID: "$leave", Membership: "leave"}},
}); err != nil {
t.Fatalf("Write leave: %v", err)
}
if got, err := rooms.InviteState(ctx, testRoomID, invitee); err != nil || got != nil {
t.Errorf("InviteState after a rejection = %s, %v; want nothing", got, err)
}
})
}
}
func TestServerKeysAreUpsertedAndFiltered(t *testing.T) {
for _, driver := range testDrivers {
t.Run(driver, func(t *testing.T) {
ctx := t.Context()
fed := migratedFederationDB(t, driver).Federation()
if err := fed.StoreServerKeys(ctx, []storage.ServerKey{
{ServerName: "remote.example", KeyID: "ed25519:a", PublicKey: "AAA", ValidUntilTS: 10, FetchedAtMS: 1},
{ServerName: "remote.example", KeyID: "ed25519:b", PublicKey: "BBB", ValidUntilTS: 20, ExpiredTS: 15, FetchedAtMS: 1},
{ServerName: "other.example", KeyID: "ed25519:a", PublicKey: "CCC", ValidUntilTS: 30, FetchedAtMS: 1},
}); err != nil {
t.Fatalf("StoreServerKeys: %v", err)
}
if err := fed.StoreServerKeys(ctx, []storage.ServerKey{
{ServerName: "remote.example", KeyID: "ed25519:a", PublicKey: "AAA", ValidUntilTS: 99, FetchedAtMS: 2},
}); err != nil {
t.Fatalf("StoreServerKeys again: %v", err)
}
all, err := fed.ServerKeys(ctx, "remote.example", nil)
if err != nil {
t.Fatalf("ServerKeys: %v", err)
}
if len(all) != 2 {
t.Fatalf("ServerKeys found %d keys, want 2", len(all))
}
if all[0].KeyID != "ed25519:a" || all[0].ValidUntilTS != 99 {
t.Errorf("the refetched key was not replaced: %+v", all[0])
}
if all[1].ExpiredTS != 15 {
t.Errorf("an old key lost its expiry: %+v", all[1])
}
one, err := fed.ServerKeys(ctx, "remote.example", []string{"ed25519:b"})
if err != nil {
t.Fatalf("ServerKeys(b): %v", err)
}
if len(one) != 1 || one[0].PublicKey != "BBB" {
t.Errorf("ServerKeys(b) = %+v", one)
}
})
}
}
func TestQueueIsPerDestinationAndInOrder(t *testing.T) {
for _, driver := range testDrivers {
t.Run(driver, func(t *testing.T) {
ctx := t.Context()
fed := migratedFederationDB(t, driver).Federation()
queued, err := fed.Enqueue(ctx, []storage.QueuedItem{
{Destination: "a.example", Kind: storage.QueueKindPDU, EventID: "$1", Payload: []byte(`{"n":1}`)},
{Destination: "b.example", Kind: storage.QueueKindPDU, EventID: "$1", Payload: []byte(`{"n":1}`)},
{Destination: "a.example", Kind: storage.QueueKindEDU, Payload: []byte(`{"edu":true}`)},
{Destination: "a.example", Kind: storage.QueueKindPDU, EventID: "$2", Payload: []byte(`{"n":2}`)},
})
if err != nil {
t.Fatalf("Enqueue: %v", err)
}
for i := 1; i < len(queued); i++ {
if queued[i].ID <= queued[i-1].ID {
t.Fatalf("queue IDs are not increasing: %d then %d", queued[i-1].ID, queued[i].ID)
}
}
pdus, err := fed.Queued(ctx, "a.example", storage.QueueKindPDU, 10)
if err != nil {
t.Fatalf("Queued: %v", err)
}
if len(pdus) != 2 || pdus[0].EventID != "$1" || pdus[1].EventID != "$2" {
t.Fatalf("Queued(a, pdu) = %+v", pdus)
}
if string(pdus[1].Payload) != `{"n":2}` {
t.Errorf("payload = %s", pdus[1].Payload)
}
destinations, err := fed.QueuedDestinations(ctx)
if err != nil {
t.Fatalf("QueuedDestinations: %v", err)
}
if !slices.Equal(destinations, []string{"a.example", "b.example"}) {
t.Errorf("QueuedDestinations = %v", destinations)
}
if err := fed.Dequeue(ctx, []int64{pdus[0].ID}); err != nil {
t.Fatalf("Dequeue: %v", err)
}
left, err := fed.Queued(ctx, "a.example", storage.QueueKindPDU, 10)
if err != nil {
t.Fatalf("Queued after Dequeue: %v", err)
}
if len(left) != 1 || left[0].EventID != "$2" {
t.Errorf("after dequeue = %+v", left)
}
})
}
}
func TestDestinationHealth(t *testing.T) {
for _, driver := range testDrivers {
t.Run(driver, func(t *testing.T) {
ctx := t.Context()
fed := migratedFederationDB(t, driver).Federation()
d, err := fed.Destination(ctx, "never.example")
if err != nil {
t.Fatalf("Destination: %v", err)
}
if d.ServerName != "never.example" || d.Failures != 0 {
t.Errorf("an unknown destination = %+v, want a healthy zero", d)
}
if err := fed.SetDestination(ctx, storage.Destination{
ServerName: "down.example", Failures: 3, RetryAtMS: 500, LastError: "connection refused",
}); err != nil {
t.Fatalf("SetDestination: %v", err)
}
if err := fed.SetDestination(ctx, storage.Destination{
ServerName: "down.example", Failures: 4, RetryAtMS: 900, LastError: "timeout",
}); err != nil {
t.Fatalf("SetDestination again: %v", err)
}
if _, err := fed.Enqueue(ctx, []storage.QueuedItem{
{Destination: "busy.example", Kind: storage.QueueKindPDU, Payload: []byte(`{}`)},
{Destination: "busy.example", Kind: storage.QueueKindEDU, Payload: []byte(`{}`)},
{Destination: "down.example", Kind: storage.QueueKindPDU, Payload: []byte(`{}`)},
}); err != nil {
t.Fatalf("Enqueue: %v", err)
}
all, err := fed.Destinations(ctx)
if err != nil {
t.Fatalf("Destinations: %v", err)
}
if len(all) != 2 {
t.Fatalf("Destinations = %+v, want busy and down", all)
}
busy, down := all[0], all[1]
if busy.ServerName != "busy.example" || busy.QueuedPDUs != 1 || busy.QueuedEDUs != 1 || busy.Failures != 0 {
t.Errorf("busy = %+v", busy)
}
if down.ServerName != "down.example" || down.Failures != 4 || down.LastError != "timeout" || down.QueuedPDUs != 1 {
t.Errorf("down = %+v", down)
}
})
}
}
func TestInboundTransactionsAreRememberedOnce(t *testing.T) {
for _, driver := range testDrivers {
t.Run(driver, func(t *testing.T) {
ctx := t.Context()
fed := migratedFederationDB(t, driver).Federation()
if _, ok, err := fed.InboundTransaction(ctx, "remote.example", "t1"); err != nil || ok {
t.Fatalf("InboundTransaction before = %v, %v", ok, err)
}
if err := fed.RecordInboundTransaction(ctx, "remote.example", "t1", []byte(`{"pdus":{}}`), 1); err != nil {
t.Fatalf("RecordInboundTransaction: %v", err)
}
if err := fed.RecordInboundTransaction(ctx, "remote.example", "t1", []byte(`{"other":1}`), 2); err != nil {
t.Fatalf("RecordInboundTransaction again: %v", err)
}
got, ok, err := fed.InboundTransaction(ctx, "remote.example", "t1")
if err != nil || !ok || string(got) != `{"pdus":{}}` {
t.Errorf("InboundTransaction = %s, %v, %v; want the first response", got, ok, err)
}
})
}
}
func eventIDsOf(events []storage.StoredEvent) []string {
out := make([]string, 0, len(events))
for _, e := range events {
out = append(out, e.EventID)
}
return out
}
func TestHasJoinedMemberOn(t *testing.T) {
for _, driver := range testDrivers {
t.Run(driver, func(t *testing.T) {
ctx := t.Context()
rooms := migratedFederationDB(t, driver).Rooms()
if err := rooms.CreateRoom(ctx, newTestRoom(testRoomID)); err != nil {
t.Fatal(err)
}
for _, m := range []struct{ user, membership string }{
{"@a:one.example", "join"},
{"@b:two.example", "leave"},
{"@c:x_y.example", "join"},
} {
e := newTestEvent(testRoomID, "$"+m.user, "m.room.member", stateKey(m.user))
if _, err := rooms.Write(ctx, storage.RoomWrite{
Event: e,
StateDelta: []storage.StateEntry{{Type: "m.room.member", StateKey: m.user, EventID: e.EventID, Membership: m.membership}},
Extremities: []string{e.EventID},
}); err != nil {
t.Fatal(err)
}
}
cases := map[string]bool{
"one.example": true,
"two.example": false, // left
"x_y.example": true,
"xay.example": false, // '_' is not a LIKE wildcard here
"e.example": false, // a suffix of a name is not the name
"none.example": false,
}
for server, want := range cases {
got, err := rooms.HasJoinedMemberOn(ctx, testRoomID, server)
if err != nil {
t.Fatal(err)
}
if got != want {
t.Errorf("HasJoinedMemberOn(%s) = %v, want %v", server, got, want)
}
}
})
}
}