mirror of
https://source.quilibrium.com/quilibrium/ceremonyclient.git
synced 2024-11-20 15:15:18 +00:00
2503 lines
62 KiB
Go
2503 lines
62 KiB
Go
package blossomsub
|
|
|
|
import (
|
|
"bytes"
|
|
"context"
|
|
"fmt"
|
|
"io"
|
|
"math/rand"
|
|
"sync"
|
|
"sync/atomic"
|
|
"testing"
|
|
"time"
|
|
|
|
pb "source.quilibrium.com/quilibrium/monorepo/go-libp2p-blossomsub/pb"
|
|
|
|
"github.com/libp2p/go-libp2p/core/host"
|
|
"github.com/libp2p/go-libp2p/core/network"
|
|
"github.com/libp2p/go-libp2p/core/peer"
|
|
"github.com/libp2p/go-libp2p/core/peerstore"
|
|
"github.com/libp2p/go-libp2p/core/record"
|
|
|
|
bhost "github.com/libp2p/go-libp2p/p2p/host/blank"
|
|
swarmt "github.com/libp2p/go-libp2p/p2p/net/swarm/testing"
|
|
|
|
"github.com/libp2p/go-msgio/protoio"
|
|
)
|
|
|
|
func getBlossomSub(ctx context.Context, h host.Host, opts ...Option) *PubSub {
|
|
ps, err := NewBlossomSub(ctx, h, opts...)
|
|
if err != nil {
|
|
panic(err)
|
|
}
|
|
return ps
|
|
}
|
|
|
|
func getBlossomSubs(ctx context.Context, hs []host.Host, opts ...Option) []*PubSub {
|
|
var psubs []*PubSub
|
|
for _, h := range hs {
|
|
psubs = append(psubs, getBlossomSub(ctx, h, opts...))
|
|
}
|
|
return psubs
|
|
}
|
|
|
|
func TestSparseBlossomSub(t *testing.T) {
|
|
ctx, cancel := context.WithCancel(context.Background())
|
|
defer cancel()
|
|
hosts := getNetHosts(t, ctx, 20)
|
|
|
|
psubs := getBlossomSubs(ctx, hosts)
|
|
|
|
var msgs []*Subscription
|
|
for _, ps := range psubs {
|
|
subch, err := ps.Subscribe([]byte{0xf0, 0x0b, 0xa1, 0x20})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
|
|
msgs = append(msgs, subch)
|
|
}
|
|
|
|
sparseConnect(t, hosts)
|
|
|
|
// wait for heartbeats to build mesh
|
|
time.Sleep(time.Second * 2)
|
|
|
|
for i := 0; i < 100; i++ {
|
|
msg := []byte(fmt.Sprintf("%d it's not a floooooood %d", i, i))
|
|
|
|
owner := rand.Intn(len(psubs))
|
|
|
|
psubs[owner].Publish([]byte{0xf0, 0x0b, 0xa1, 0x20}, msg)
|
|
|
|
for _, sub := range msgs {
|
|
got, err := sub.Next(ctx)
|
|
if err != nil {
|
|
t.Fatal(sub.err)
|
|
}
|
|
if !bytes.Equal(msg, got.Data) {
|
|
t.Fatal("got wrong message!")
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
func TestDenseBlossomSub(t *testing.T) {
|
|
ctx, cancel := context.WithCancel(context.Background())
|
|
defer cancel()
|
|
hosts := getNetHosts(t, ctx, 20)
|
|
|
|
psubs := getBlossomSubs(ctx, hosts)
|
|
|
|
var msgs []*Subscription
|
|
for _, ps := range psubs {
|
|
subch, err := ps.Subscribe([]byte{0xf0, 0x0b, 0xa1, 0x20})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
|
|
msgs = append(msgs, subch)
|
|
}
|
|
|
|
denseConnect(t, hosts)
|
|
|
|
// wait for heartbeats to build mesh
|
|
time.Sleep(time.Second * 2)
|
|
|
|
for i := 0; i < 100; i++ {
|
|
msg := []byte(fmt.Sprintf("%d it's not a floooooood %d", i, i))
|
|
|
|
owner := rand.Intn(len(psubs))
|
|
|
|
psubs[owner].Publish([]byte{0xf0, 0x0b, 0xa1, 0x20}, msg)
|
|
|
|
for _, sub := range msgs {
|
|
got, err := sub.Next(ctx)
|
|
if err != nil {
|
|
t.Fatal(sub.err)
|
|
}
|
|
if !bytes.Equal(msg, got.Data) {
|
|
t.Fatal("got wrong message!")
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
func TestBlossomSubFanout(t *testing.T) {
|
|
ctx, cancel := context.WithCancel(context.Background())
|
|
defer cancel()
|
|
hosts := getNetHosts(t, ctx, 20)
|
|
|
|
psubs := getBlossomSubs(ctx, hosts)
|
|
|
|
var msgs []*Subscription
|
|
for _, ps := range psubs[1:] {
|
|
subch, err := ps.Subscribe([]byte{0xf0, 0x0b, 0xa1, 0x20})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
|
|
msgs = append(msgs, subch)
|
|
}
|
|
|
|
denseConnect(t, hosts)
|
|
|
|
// wait for heartbeats to build mesh
|
|
time.Sleep(time.Second * 2)
|
|
|
|
for i := 0; i < 100; i++ {
|
|
msg := []byte(fmt.Sprintf("%d it's not a floooooood %d", i, i))
|
|
|
|
owner := 0
|
|
|
|
psubs[owner].Publish([]byte{0xf0, 0x0b, 0xa1, 0x20}, msg)
|
|
|
|
for _, sub := range msgs {
|
|
got, err := sub.Next(ctx)
|
|
if err != nil {
|
|
t.Fatal(sub.err)
|
|
}
|
|
if !bytes.Equal(msg, got.Data) {
|
|
t.Fatal("got wrong message!")
|
|
}
|
|
}
|
|
}
|
|
|
|
// subscribe the owner
|
|
subch, err := psubs[0].Subscribe([]byte{0xf0, 0x0b, 0xa1, 0x20})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
msgs = append(msgs, subch)
|
|
|
|
// wait for a heartbeat
|
|
time.Sleep(time.Second * 1)
|
|
|
|
for i := 0; i < 100; i++ {
|
|
msg := []byte(fmt.Sprintf("%d it's not a floooooood %d", i, i))
|
|
|
|
owner := 0
|
|
|
|
psubs[owner].Publish([]byte{0xf0, 0x0b, 0xa1, 0x20}, msg)
|
|
|
|
for _, sub := range msgs {
|
|
got, err := sub.Next(ctx)
|
|
if err != nil {
|
|
t.Fatal(sub.err)
|
|
}
|
|
if !bytes.Equal(msg, got.Data) {
|
|
t.Fatal("got wrong message!")
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
func TestBlossomSubFanoutMaintenance(t *testing.T) {
|
|
ctx, cancel := context.WithCancel(context.Background())
|
|
defer cancel()
|
|
hosts := getNetHosts(t, ctx, 20)
|
|
|
|
psubs := getBlossomSubs(ctx, hosts)
|
|
|
|
var msgs []*Subscription
|
|
for _, ps := range psubs[1:] {
|
|
subch, err := ps.Subscribe([]byte{0xf0, 0x0b, 0xa1, 0x20})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
|
|
msgs = append(msgs, subch)
|
|
}
|
|
|
|
denseConnect(t, hosts)
|
|
|
|
// wait for heartbeats to build mesh
|
|
time.Sleep(time.Second * 2)
|
|
|
|
for i := 0; i < 100; i++ {
|
|
msg := []byte(fmt.Sprintf("%d it's not a floooooood %d", i, i))
|
|
|
|
owner := 0
|
|
|
|
psubs[owner].Publish([]byte{0xf0, 0x0b, 0xa1, 0x20}, msg)
|
|
|
|
for _, sub := range msgs {
|
|
got, err := sub.Next(ctx)
|
|
if err != nil {
|
|
t.Fatal(sub.err)
|
|
}
|
|
if !bytes.Equal(msg, got.Data) {
|
|
t.Fatal("got wrong message!")
|
|
}
|
|
}
|
|
}
|
|
|
|
// unsubscribe all peers to exercise fanout maintenance
|
|
for _, sub := range msgs {
|
|
sub.Cancel()
|
|
}
|
|
msgs = nil
|
|
|
|
// wait for heartbeats
|
|
time.Sleep(time.Second * 2)
|
|
|
|
// resubscribe and repeat
|
|
for _, ps := range psubs[1:] {
|
|
subch, err := ps.Subscribe([]byte{0xf0, 0x0b, 0xa1, 0x20})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
|
|
msgs = append(msgs, subch)
|
|
}
|
|
|
|
time.Sleep(time.Second * 2)
|
|
|
|
for i := 0; i < 100; i++ {
|
|
msg := []byte(fmt.Sprintf("%d it's not a floooooood %d", i, i))
|
|
|
|
owner := 0
|
|
|
|
psubs[owner].Publish([]byte{0xf0, 0x0b, 0xa1, 0x20}, msg)
|
|
|
|
for _, sub := range msgs {
|
|
got, err := sub.Next(ctx)
|
|
if err != nil {
|
|
t.Fatal(sub.err)
|
|
}
|
|
if !bytes.Equal(msg, got.Data) {
|
|
t.Fatal("got wrong message!")
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
func TestBlossomSubFanoutExpiry(t *testing.T) {
|
|
BlossomSubFanoutTTL = 1 * time.Second
|
|
defer func() {
|
|
BlossomSubFanoutTTL = 60 * time.Second
|
|
}()
|
|
|
|
ctx, cancel := context.WithCancel(context.Background())
|
|
defer cancel()
|
|
hosts := getNetHosts(t, ctx, 10)
|
|
|
|
psubs := getBlossomSubs(ctx, hosts)
|
|
|
|
var msgs []*Subscription
|
|
for _, ps := range psubs[1:] {
|
|
subch, err := ps.Subscribe([]byte{0xf0, 0x0b, 0xa1, 0x20})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
|
|
msgs = append(msgs, subch)
|
|
}
|
|
|
|
denseConnect(t, hosts)
|
|
|
|
// wait for heartbeats to build mesh
|
|
time.Sleep(time.Second * 2)
|
|
|
|
for i := 0; i < 5; i++ {
|
|
msg := []byte(fmt.Sprintf("%d it's not a floooooood %d", i, i))
|
|
|
|
owner := 0
|
|
|
|
psubs[owner].Publish([]byte{0xf0, 0x0b, 0xa1, 0x20}, msg)
|
|
|
|
for _, sub := range msgs {
|
|
got, err := sub.Next(ctx)
|
|
if err != nil {
|
|
t.Fatal(sub.err)
|
|
}
|
|
if !bytes.Equal(msg, got.Data) {
|
|
t.Fatal("got wrong message!")
|
|
}
|
|
}
|
|
}
|
|
|
|
psubs[0].eval <- func() {
|
|
if len(psubs[0].rt.(*BlossomSubRouter).fanout) == 0 {
|
|
t.Fatal("owner has no fanout")
|
|
}
|
|
}
|
|
|
|
// wait for TTL to expire fanout peers in owner
|
|
time.Sleep(time.Second * 2)
|
|
|
|
psubs[0].eval <- func() {
|
|
if len(psubs[0].rt.(*BlossomSubRouter).fanout) > 0 {
|
|
t.Fatal("fanout hasn't expired")
|
|
}
|
|
}
|
|
|
|
// wait for it to run in the event loop
|
|
time.Sleep(10 * time.Millisecond)
|
|
}
|
|
|
|
func TestBlossomSubGossip(t *testing.T) {
|
|
ctx, cancel := context.WithCancel(context.Background())
|
|
defer cancel()
|
|
hosts := getNetHosts(t, ctx, 20)
|
|
|
|
psubs := getBlossomSubs(ctx, hosts)
|
|
|
|
var msgs []*Subscription
|
|
for _, ps := range psubs {
|
|
subch, err := ps.Subscribe([]byte{0xf0, 0x0b, 0xa1, 0x20})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
|
|
msgs = append(msgs, subch)
|
|
}
|
|
|
|
denseConnect(t, hosts)
|
|
|
|
// wait for heartbeats to build mesh
|
|
time.Sleep(time.Second * 2)
|
|
|
|
for i := 0; i < 100; i++ {
|
|
msg := []byte(fmt.Sprintf("%d it's not a floooooood %d", i, i))
|
|
|
|
owner := rand.Intn(len(psubs))
|
|
|
|
psubs[owner].Publish([]byte{0xf0, 0x0b, 0xa1, 0x20}, msg)
|
|
|
|
for _, sub := range msgs {
|
|
got, err := sub.Next(ctx)
|
|
if err != nil {
|
|
t.Fatal(sub.err)
|
|
}
|
|
if !bytes.Equal(msg, got.Data) {
|
|
t.Fatal("got wrong message!")
|
|
}
|
|
}
|
|
|
|
// wait a bit to have some gossip interleaved
|
|
time.Sleep(time.Millisecond * 100)
|
|
}
|
|
|
|
// and wait for some gossip flushing
|
|
time.Sleep(time.Second * 2)
|
|
}
|
|
|
|
func TestBlossomSubGossipPiggyback(t *testing.T) {
|
|
t.Skip("test no longer relevant; gossip propagation has become eager")
|
|
ctx, cancel := context.WithCancel(context.Background())
|
|
defer cancel()
|
|
hosts := getNetHosts(t, ctx, 20)
|
|
|
|
psubs := getBlossomSubs(ctx, hosts)
|
|
|
|
var msgs []*Subscription
|
|
for _, ps := range psubs {
|
|
subch, err := ps.Subscribe([]byte{0xf0, 0x0b, 0xa1, 0x20})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
|
|
msgs = append(msgs, subch)
|
|
}
|
|
|
|
var xmsgs []*Subscription
|
|
for _, ps := range psubs {
|
|
subch, err := ps.Subscribe([]byte{0xba, 0x2c, 0x12, 0x08})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
|
|
xmsgs = append(xmsgs, subch)
|
|
}
|
|
|
|
denseConnect(t, hosts)
|
|
|
|
// wait for heartbeats to build mesh
|
|
time.Sleep(time.Second * 2)
|
|
|
|
for i := 0; i < 100; i++ {
|
|
msg := []byte(fmt.Sprintf("%d it's not a floooooood %d", i, i))
|
|
|
|
owner := rand.Intn(len(psubs))
|
|
|
|
psubs[owner].Publish([]byte{0xf0, 0x0b, 0xa1, 0x20}, msg)
|
|
psubs[owner].Publish([]byte{0xba, 0x2c, 0x12, 0x08}, msg)
|
|
|
|
for _, sub := range msgs {
|
|
got, err := sub.Next(ctx)
|
|
if err != nil {
|
|
t.Fatal(sub.err)
|
|
}
|
|
if !bytes.Equal(msg, got.Data) {
|
|
t.Fatal("got wrong message!")
|
|
}
|
|
}
|
|
|
|
for _, sub := range xmsgs {
|
|
got, err := sub.Next(ctx)
|
|
if err != nil {
|
|
t.Fatal(sub.err)
|
|
}
|
|
if !bytes.Equal(msg, got.Data) {
|
|
t.Fatal("got wrong message!")
|
|
}
|
|
}
|
|
|
|
// wait a bit to have some gossip interleaved
|
|
time.Sleep(time.Millisecond * 100)
|
|
}
|
|
|
|
// and wait for some gossip flushing
|
|
time.Sleep(time.Second * 2)
|
|
}
|
|
|
|
func TestBlossomSubGossipPropagation(t *testing.T) {
|
|
ctx, cancel := context.WithCancel(context.Background())
|
|
defer cancel()
|
|
|
|
hosts := getNetHosts(t, ctx, 20)
|
|
psubs := getBlossomSubs(ctx, hosts)
|
|
|
|
hosts1 := hosts[:BlossomSubD+1]
|
|
hosts2 := append(hosts[BlossomSubD+1:], hosts[0])
|
|
|
|
denseConnect(t, hosts1)
|
|
denseConnect(t, hosts2)
|
|
|
|
var msgs1 []*Subscription
|
|
for _, ps := range psubs[1 : BlossomSubD+1] {
|
|
subch, err := ps.Subscribe([]byte{0xf0, 0x0b, 0xa1, 0x20})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
|
|
msgs1 = append(msgs1, subch)
|
|
}
|
|
|
|
time.Sleep(time.Second * 1)
|
|
|
|
for i := 0; i < 10; i++ {
|
|
msg := []byte(fmt.Sprintf("%d it's not a floooooood %d", i, i))
|
|
|
|
owner := 0
|
|
|
|
psubs[owner].Publish([]byte{0xf0, 0x0b, 0xa1, 0x20}, msg)
|
|
|
|
for _, sub := range msgs1 {
|
|
got, err := sub.Next(ctx)
|
|
if err != nil {
|
|
t.Fatal(sub.err)
|
|
}
|
|
if !bytes.Equal(msg, got.Data) {
|
|
t.Fatal("got wrong message!")
|
|
}
|
|
}
|
|
}
|
|
|
|
time.Sleep(time.Millisecond * 100)
|
|
|
|
var msgs2 []*Subscription
|
|
for _, ps := range psubs[BlossomSubD+1:] {
|
|
subch, err := ps.Subscribe([]byte{0xf0, 0x0b, 0xa1, 0x20})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
|
|
msgs2 = append(msgs2, subch)
|
|
}
|
|
|
|
var collect [][]byte
|
|
for i := 0; i < 10; i++ {
|
|
for _, sub := range msgs2 {
|
|
got, err := sub.Next(ctx)
|
|
if err != nil {
|
|
t.Fatal(sub.err)
|
|
}
|
|
collect = append(collect, got.Data)
|
|
}
|
|
}
|
|
|
|
for i := 0; i < 10; i++ {
|
|
msg := []byte(fmt.Sprintf("%d it's not a floooooood %d", i, i))
|
|
gotit := false
|
|
for j := 0; j < len(collect); j++ {
|
|
if bytes.Equal(msg, collect[j]) {
|
|
gotit = true
|
|
break
|
|
}
|
|
}
|
|
if !gotit {
|
|
t.Fatalf("Didn't get message %s", string(msg))
|
|
}
|
|
}
|
|
}
|
|
|
|
func TestBlossomSubPrune(t *testing.T) {
|
|
ctx, cancel := context.WithCancel(context.Background())
|
|
defer cancel()
|
|
hosts := getNetHosts(t, ctx, 20)
|
|
|
|
psubs := getBlossomSubs(ctx, hosts)
|
|
|
|
var msgs []*Subscription
|
|
for _, ps := range psubs {
|
|
subch, err := ps.Subscribe([]byte{0xf0, 0x0b, 0xa1, 0x20})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
|
|
msgs = append(msgs, subch)
|
|
}
|
|
|
|
denseConnect(t, hosts)
|
|
|
|
// wait for heartbeats to build mesh
|
|
time.Sleep(time.Second * 2)
|
|
|
|
// disconnect some peers from the mesh to get some PRUNEs
|
|
for _, sub := range msgs[:5] {
|
|
sub.Cancel()
|
|
}
|
|
|
|
// wait a bit to take effect
|
|
time.Sleep(time.Millisecond * 100)
|
|
|
|
for i := 0; i < 10; i++ {
|
|
msg := []byte(fmt.Sprintf("%d it's not a floooooood %d", i, i))
|
|
|
|
owner := rand.Intn(len(psubs))
|
|
|
|
psubs[owner].Publish([]byte{0xf0, 0x0b, 0xa1, 0x20}, msg)
|
|
|
|
for _, sub := range msgs[5:] {
|
|
got, err := sub.Next(ctx)
|
|
if err != nil {
|
|
t.Fatal(sub.err)
|
|
}
|
|
if !bytes.Equal(msg, got.Data) {
|
|
t.Fatal("got wrong message!")
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
func TestBlossomSubPruneBackoffTime(t *testing.T) {
|
|
ctx, cancel := context.WithCancel(context.Background())
|
|
defer cancel()
|
|
hosts := getNetHosts(t, ctx, 10)
|
|
|
|
// App specific score that we'll change later.
|
|
currentScoreForHost0 := int32(0)
|
|
|
|
params := DefaultBlossomSubParams()
|
|
params.HeartbeatInitialDelay = time.Millisecond * 10
|
|
params.HeartbeatInterval = time.Millisecond * 100
|
|
|
|
psubs := getBlossomSubs(ctx, hosts, WithBlossomSubParams(params), WithPeerScore(
|
|
&PeerScoreParams{
|
|
AppSpecificScore: func(p peer.ID) float64 {
|
|
if p == hosts[0].ID() {
|
|
return float64(atomic.LoadInt32(¤tScoreForHost0))
|
|
} else {
|
|
return 0
|
|
}
|
|
},
|
|
AppSpecificWeight: 1,
|
|
DecayInterval: time.Second,
|
|
DecayToZero: 0.01,
|
|
},
|
|
&PeerScoreThresholds{
|
|
GossipThreshold: -1,
|
|
PublishThreshold: -1,
|
|
GraylistThreshold: -1,
|
|
}))
|
|
|
|
var msgs []*Subscription
|
|
for _, ps := range psubs {
|
|
subch, err := ps.Subscribe([]byte{0xf0, 0x0b, 0xa1, 0x20})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
|
|
msgs = append(msgs, subch)
|
|
}
|
|
|
|
connectAll(t, hosts)
|
|
|
|
// wait for heartbeats to build mesh
|
|
time.Sleep(time.Second)
|
|
|
|
pruneTime := time.Now()
|
|
// Flip the score. Host 0 should be pruned from everyone
|
|
atomic.StoreInt32(¤tScoreForHost0, -1000)
|
|
|
|
// wait for heartbeats to run and prune
|
|
time.Sleep(time.Second)
|
|
|
|
wg := sync.WaitGroup{}
|
|
var missingBackoffs uint32 = 0
|
|
for i := 1; i < 10; i++ {
|
|
wg.Add(1)
|
|
// Copy i so this func keeps the correct value in the closure.
|
|
var idx = i
|
|
// Run this check in the eval thunk so that we don't step over the heartbeat goroutine and trigger a race.
|
|
psubs[idx].rt.(*BlossomSubRouter).p.eval <- func() {
|
|
defer wg.Done()
|
|
backoff, ok := psubs[idx].rt.(*BlossomSubRouter).backoff[string([]byte{0xf0, 0x0b, 0xa1, 0x20})][hosts[0].ID()]
|
|
if !ok {
|
|
atomic.AddUint32(&missingBackoffs, 1)
|
|
}
|
|
if ok && backoff.Sub(pruneTime)-params.PruneBackoff > time.Second {
|
|
t.Errorf("backoff time should be equal to prune backoff (with some slack) was %v", backoff.Sub(pruneTime)-params.PruneBackoff)
|
|
}
|
|
}
|
|
}
|
|
wg.Wait()
|
|
|
|
// Sometimes not all the peers will have updated their backoffs by this point. If the majority haven't we'll fail this test.
|
|
if missingBackoffs >= 5 {
|
|
t.Errorf("missing too many backoffs: %v", missingBackoffs)
|
|
}
|
|
|
|
for i := 0; i < 10; i++ {
|
|
msg := []byte(fmt.Sprintf("%d it's not a floooooood %d", i, i))
|
|
|
|
// Don't publish from host 0, since everyone should have pruned it.
|
|
owner := rand.Intn(len(psubs)-1) + 1
|
|
|
|
psubs[owner].Publish([]byte{0xf0, 0x0b, 0xa1, 0x20}, msg)
|
|
|
|
for _, sub := range msgs[1:] {
|
|
got, err := sub.Next(ctx)
|
|
if err != nil {
|
|
t.Fatal(sub.err)
|
|
}
|
|
if !bytes.Equal(msg, got.Data) {
|
|
t.Fatal("got wrong message!")
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
func TestBlossomSubGraft(t *testing.T) {
|
|
ctx, cancel := context.WithCancel(context.Background())
|
|
defer cancel()
|
|
hosts := getNetHosts(t, ctx, 20)
|
|
|
|
psubs := getBlossomSubs(ctx, hosts)
|
|
|
|
sparseConnect(t, hosts)
|
|
|
|
time.Sleep(time.Second * 1)
|
|
|
|
var msgs []*Subscription
|
|
for _, ps := range psubs {
|
|
subch, err := ps.Subscribe([]byte{0xf0, 0x0b, 0xa1, 0x20})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
|
|
msgs = append(msgs, subch)
|
|
|
|
// wait for announce to propagate
|
|
time.Sleep(time.Millisecond * 100)
|
|
}
|
|
|
|
time.Sleep(time.Second * 1)
|
|
|
|
for i := 0; i < 100; i++ {
|
|
msg := []byte(fmt.Sprintf("%d it's not a floooooood %d", i, i))
|
|
|
|
owner := rand.Intn(len(psubs))
|
|
|
|
psubs[owner].Publish([]byte{0xf0, 0x0b, 0xa1, 0x20}, msg)
|
|
|
|
for _, sub := range msgs {
|
|
got, err := sub.Next(ctx)
|
|
if err != nil {
|
|
t.Fatal(sub.err)
|
|
}
|
|
if !bytes.Equal(msg, got.Data) {
|
|
t.Fatal("got wrong message!")
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
func TestBlossomSubRemovePeer(t *testing.T) {
|
|
ctx, cancel := context.WithCancel(context.Background())
|
|
defer cancel()
|
|
hosts := getNetHosts(t, ctx, 20)
|
|
|
|
psubs := getBlossomSubs(ctx, hosts)
|
|
|
|
var msgs []*Subscription
|
|
for _, ps := range psubs {
|
|
subch, err := ps.Subscribe([]byte{0xf0, 0x0b, 0xa1, 0x20})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
|
|
msgs = append(msgs, subch)
|
|
}
|
|
|
|
denseConnect(t, hosts)
|
|
|
|
// wait for heartbeats to build mesh
|
|
time.Sleep(time.Second * 2)
|
|
|
|
// disconnect some peers to exercise RemovePeer paths
|
|
for _, host := range hosts[:5] {
|
|
host.Close()
|
|
}
|
|
|
|
// wait a heartbeat
|
|
time.Sleep(time.Second * 1)
|
|
|
|
for i := 0; i < 10; i++ {
|
|
msg := []byte(fmt.Sprintf("%d it's not a floooooood %d", i, i))
|
|
|
|
owner := 5 + rand.Intn(len(psubs)-5)
|
|
|
|
psubs[owner].Publish([]byte{0xf0, 0x0b, 0xa1, 0x20}, msg)
|
|
|
|
for _, sub := range msgs[5:] {
|
|
got, err := sub.Next(ctx)
|
|
if err != nil {
|
|
t.Fatal(sub.err)
|
|
}
|
|
if !bytes.Equal(msg, got.Data) {
|
|
t.Fatal("got wrong message!")
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
func TestBlossomSubGraftPruneRetry(t *testing.T) {
|
|
ctx, cancel := context.WithCancel(context.Background())
|
|
defer cancel()
|
|
|
|
hosts := getNetHosts(t, ctx, 10)
|
|
psubs := getBlossomSubs(ctx, hosts)
|
|
denseConnect(t, hosts)
|
|
|
|
var bitmasks [][]byte
|
|
var msgs [][]*Subscription
|
|
for i := 0; i < 35; i++ {
|
|
bitmask := []byte{byte(i)}
|
|
bitmasks = append(bitmasks, bitmask)
|
|
|
|
var subs []*Subscription
|
|
for _, ps := range psubs {
|
|
subch, err := ps.Subscribe(bitmask)
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
|
|
subs = append(subs, subch)
|
|
}
|
|
msgs = append(msgs, subs)
|
|
}
|
|
|
|
// wait for heartbeats to build meshes
|
|
time.Sleep(time.Second * 5)
|
|
|
|
for i, bitmask := range bitmasks {
|
|
msg := []byte(fmt.Sprintf("%d it's not a floooooood %d", i, i))
|
|
|
|
owner := rand.Intn(len(psubs))
|
|
|
|
psubs[owner].Publish(bitmask, msg)
|
|
|
|
for _, sub := range msgs[i] {
|
|
got, err := sub.Next(ctx)
|
|
if err != nil {
|
|
t.Fatal(sub.err)
|
|
}
|
|
if !bytes.Equal(msg, got.Data) {
|
|
t.Fatal("got wrong message!")
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
func TestBlossomSubControlPiggyback(t *testing.T) {
|
|
t.Skip("travis regularly fails on this test")
|
|
|
|
ctx, cancel := context.WithCancel(context.Background())
|
|
defer cancel()
|
|
|
|
hosts := getNetHosts(t, ctx, 10)
|
|
psubs := getBlossomSubs(ctx, hosts)
|
|
denseConnect(t, hosts)
|
|
|
|
for _, ps := range psubs {
|
|
subch, err := ps.Subscribe([]byte{0xff, 0xff, 0xff, 0xff})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
go func(sub *Subscription) {
|
|
for {
|
|
_, err := sub.Next(ctx)
|
|
if err != nil {
|
|
break
|
|
}
|
|
}
|
|
}(subch)
|
|
}
|
|
|
|
time.Sleep(time.Second * 1)
|
|
|
|
// create a background flood of messages that overloads the queues
|
|
done := make(chan struct{})
|
|
go func() {
|
|
owner := rand.Intn(len(psubs))
|
|
for i := 0; i < 10000; i++ {
|
|
msg := []byte("background flooooood")
|
|
psubs[owner].Publish([]byte{0xff, 0xff, 0xff, 0xff}, msg)
|
|
}
|
|
done <- struct{}{}
|
|
}()
|
|
|
|
time.Sleep(time.Millisecond * 20)
|
|
|
|
// and subscribe to a bunch of bitmasks in the meantime -- this should
|
|
// result in some dropped control messages, with subsequent piggybacking
|
|
// in the background flood
|
|
var bitmasks [][]byte
|
|
var msgs [][]*Subscription
|
|
for i := 0; i < 5; i++ {
|
|
bitmask := []byte{byte(i)}
|
|
bitmasks = append(bitmasks, bitmask)
|
|
|
|
var subs []*Subscription
|
|
for _, ps := range psubs {
|
|
subch, err := ps.Subscribe(bitmask)
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
|
|
subs = append(subs, subch)
|
|
}
|
|
msgs = append(msgs, subs)
|
|
}
|
|
|
|
// wait for the flood to stop
|
|
<-done
|
|
|
|
// and test that we have functional overlays
|
|
for i, bitmask := range bitmasks {
|
|
msg := []byte(fmt.Sprintf("%d it's not a floooooood %d", i, i))
|
|
|
|
owner := rand.Intn(len(psubs))
|
|
|
|
psubs[owner].Publish(bitmask, msg)
|
|
|
|
for _, sub := range msgs[i] {
|
|
got, err := sub.Next(ctx)
|
|
if err != nil {
|
|
t.Fatal(sub.err)
|
|
}
|
|
if !bytes.Equal(msg, got.Data) {
|
|
t.Fatal("got wrong message!")
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
func TestMixedBlossomSub(t *testing.T) {
|
|
ctx, cancel := context.WithCancel(context.Background())
|
|
defer cancel()
|
|
hosts := getNetHosts(t, ctx, 30)
|
|
|
|
gsubs := getBlossomSubs(ctx, hosts[:20])
|
|
fsubs := getPubsubs(ctx, hosts[20:])
|
|
psubs := append(gsubs, fsubs...)
|
|
|
|
var msgs []*Subscription
|
|
for _, ps := range psubs {
|
|
subch, err := ps.Subscribe([]byte{0xf0, 0x0b, 0xa1, 0x2b})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
|
|
msgs = append(msgs, subch)
|
|
}
|
|
|
|
sparseConnect(t, hosts)
|
|
|
|
// wait for heartbeats to build mesh
|
|
time.Sleep(time.Second * 2)
|
|
|
|
for i := 0; i < 100; i++ {
|
|
msg := []byte(fmt.Sprintf("%d it's not a floooooood %d", i, i))
|
|
|
|
owner := rand.Intn(len(psubs))
|
|
|
|
psubs[owner].Publish([]byte{0xf0, 0x0b, 0xa1, 0x2b}, msg)
|
|
|
|
for _, sub := range msgs {
|
|
got, err := sub.Next(ctx)
|
|
if err != nil {
|
|
t.Fatal(sub.err)
|
|
}
|
|
if !bytes.Equal(msg, got.Data) {
|
|
t.Fatal("got wrong message!")
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
func TestBlossomSubMultihops(t *testing.T) {
|
|
ctx, cancel := context.WithCancel(context.Background())
|
|
defer cancel()
|
|
|
|
hosts := getNetHosts(t, ctx, 6)
|
|
|
|
psubs := getBlossomSubs(ctx, hosts)
|
|
|
|
connect(t, hosts[0], hosts[1])
|
|
connect(t, hosts[1], hosts[2])
|
|
connect(t, hosts[2], hosts[3])
|
|
connect(t, hosts[3], hosts[4])
|
|
connect(t, hosts[4], hosts[5])
|
|
|
|
var subs []*Subscription
|
|
for i := 1; i < 6; i++ {
|
|
ch, err := psubs[i].Subscribe([]byte{0xf0, 0x0b, 0xa1, 0x2b})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
subs = append(subs, ch)
|
|
}
|
|
|
|
// wait for heartbeats to build mesh
|
|
time.Sleep(time.Second * 2)
|
|
|
|
msg := []byte("i like cats")
|
|
err := psubs[0].Publish([]byte{0xf0, 0x0b, 0xa1, 0x2b}, msg)
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
|
|
// last node in the chain should get the message
|
|
select {
|
|
case out := <-subs[4].ch:
|
|
if !bytes.Equal(out.GetData(), msg) {
|
|
t.Fatal("got wrong data")
|
|
}
|
|
case <-time.After(time.Second * 5):
|
|
t.Fatal("timed out waiting for message")
|
|
}
|
|
}
|
|
|
|
func TestBlossomSubTreeTopology(t *testing.T) {
|
|
ctx, cancel := context.WithCancel(context.Background())
|
|
defer cancel()
|
|
|
|
hosts := getNetHosts(t, ctx, 10)
|
|
psubs := getBlossomSubs(ctx, hosts)
|
|
|
|
connect(t, hosts[0], hosts[1])
|
|
connect(t, hosts[1], hosts[2])
|
|
connect(t, hosts[1], hosts[4])
|
|
connect(t, hosts[2], hosts[3])
|
|
connect(t, hosts[0], hosts[5])
|
|
connect(t, hosts[5], hosts[6])
|
|
connect(t, hosts[5], hosts[8])
|
|
connect(t, hosts[6], hosts[7])
|
|
connect(t, hosts[8], hosts[9])
|
|
|
|
/*
|
|
[0] -> [1] -> [2] -> [3]
|
|
| L->[4]
|
|
v
|
|
[5] -> [6] -> [7]
|
|
|
|
|
v
|
|
[8] -> [9]
|
|
*/
|
|
|
|
var chs []*Subscription
|
|
for _, ps := range psubs {
|
|
ch, err := ps.Subscribe([]byte{0xf1, 0x22, 0xb0, 0x22})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
|
|
chs = append(chs, ch)
|
|
}
|
|
|
|
// wait for heartbeats to build mesh
|
|
time.Sleep(time.Second * 2)
|
|
|
|
assertPeerLists(t, hosts, psubs[0], 1, 5)
|
|
assertPeerLists(t, hosts, psubs[1], 0, 2, 4)
|
|
assertPeerLists(t, hosts, psubs[2], 1, 3)
|
|
|
|
checkMessageRouting(t, []byte{0xf1, 0x22, 0xb0, 0x22}, []*PubSub{psubs[9], psubs[3]}, chs)
|
|
}
|
|
|
|
// this tests overlay bootstrapping through px in BlossomSub v1.2
|
|
// we start with a star topology and rely on px through prune to build the mesh
|
|
func TestBlossomSubStarTopology(t *testing.T) {
|
|
originalBlossomSubD := BlossomSubD
|
|
BlossomSubD = 4
|
|
originalBlossomSubDhi := BlossomSubDhi
|
|
BlossomSubDhi = BlossomSubD + 1
|
|
originalBlossomSubDlo := BlossomSubDlo
|
|
BlossomSubDlo = BlossomSubD - 1
|
|
originalBlossomSubDscore := BlossomSubDscore
|
|
BlossomSubDscore = BlossomSubDlo
|
|
defer func() {
|
|
BlossomSubD = originalBlossomSubD
|
|
BlossomSubDhi = originalBlossomSubDhi
|
|
BlossomSubDlo = originalBlossomSubDlo
|
|
BlossomSubDscore = originalBlossomSubDscore
|
|
}()
|
|
|
|
ctx, cancel := context.WithCancel(context.Background())
|
|
defer cancel()
|
|
|
|
hosts := getNetHosts(t, ctx, 20)
|
|
psubs := getBlossomSubs(ctx, hosts, WithPeerExchange(true), WithFloodPublish(true))
|
|
|
|
// configure the center of the star with a very low D
|
|
psubs[0].eval <- func() {
|
|
gs := psubs[0].rt.(*BlossomSubRouter)
|
|
gs.params.D = 0
|
|
gs.params.Dlo = 0
|
|
gs.params.Dhi = 0
|
|
gs.params.Dscore = 0
|
|
}
|
|
|
|
// add all peer addresses to the peerstores
|
|
// this is necessary because we can't have signed address records witout identify
|
|
// pushing them
|
|
for i := range hosts {
|
|
for j := range hosts {
|
|
if i == j {
|
|
continue
|
|
}
|
|
hosts[i].Peerstore().AddAddrs(hosts[j].ID(), hosts[j].Addrs(), peerstore.PermanentAddrTTL)
|
|
}
|
|
}
|
|
|
|
// build the star
|
|
for i := 1; i < 20; i++ {
|
|
connect(t, hosts[0], hosts[i])
|
|
}
|
|
|
|
time.Sleep(time.Second)
|
|
|
|
// build the mesh
|
|
var subs []*Subscription
|
|
for _, ps := range psubs {
|
|
sub, err := ps.Subscribe([]byte{0xf1, 0x22, 0xb0, 0x22})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
subs = append(subs, sub)
|
|
}
|
|
|
|
// wait a bit for the mesh to build
|
|
time.Sleep(10 * time.Second)
|
|
|
|
// check that all peers have > 1 connection
|
|
for i, h := range hosts {
|
|
if len(h.Network().Conns()) == 1 {
|
|
t.Errorf("peer %d has ony a single connection", i)
|
|
}
|
|
}
|
|
|
|
// send a message from each peer and assert it was propagated
|
|
for i := 0; i < 20; i++ {
|
|
msg := []byte(fmt.Sprintf("message %d", i))
|
|
psubs[i].Publish([]byte{0xf1, 0x22, 0xb0, 0x22}, msg)
|
|
|
|
for _, sub := range subs {
|
|
assertReceive(t, sub, msg)
|
|
}
|
|
}
|
|
}
|
|
|
|
// this tests overlay bootstrapping through px in BlossomSub v1.2, with addresses
|
|
// exchanged in signed peer records.
|
|
// we start with a star topology and rely on px through prune to build the mesh
|
|
func TestBlossomSubStarTopologyWithSignedPeerRecords(t *testing.T) {
|
|
originalBlossomSubD := BlossomSubD
|
|
BlossomSubD = 4
|
|
originalBlossomSubDhi := BlossomSubDhi
|
|
BlossomSubDhi = BlossomSubD + 1
|
|
originalBlossomSubDlo := BlossomSubDlo
|
|
BlossomSubDlo = BlossomSubD - 1
|
|
originalBlossomSubDscore := BlossomSubDscore
|
|
BlossomSubDscore = BlossomSubDlo
|
|
defer func() {
|
|
BlossomSubD = originalBlossomSubD
|
|
BlossomSubDhi = originalBlossomSubDhi
|
|
BlossomSubDlo = originalBlossomSubDlo
|
|
BlossomSubDscore = originalBlossomSubDscore
|
|
}()
|
|
|
|
ctx, cancel := context.WithCancel(context.Background())
|
|
defer cancel()
|
|
|
|
hosts := getNetHosts(t, ctx, 20)
|
|
psubs := getBlossomSubs(ctx, hosts, WithPeerExchange(true), WithFloodPublish(true))
|
|
|
|
// configure the center of the star with a very low D
|
|
psubs[0].eval <- func() {
|
|
gs := psubs[0].rt.(*BlossomSubRouter)
|
|
gs.params.D = 0
|
|
gs.params.Dlo = 0
|
|
gs.params.Dhi = 0
|
|
gs.params.Dscore = 0
|
|
}
|
|
|
|
// manually create signed peer records for each host and add them to the
|
|
// peerstore of the center of the star, which is doing the bootstrapping
|
|
for i := range hosts[1:] {
|
|
privKey := hosts[i].Peerstore().PrivKey(hosts[i].ID())
|
|
if privKey == nil {
|
|
t.Fatalf("unable to get private key for host %s", hosts[i].ID().Pretty())
|
|
}
|
|
ai := host.InfoFromHost(hosts[i])
|
|
rec := peer.PeerRecordFromAddrInfo(*ai)
|
|
signedRec, err := record.Seal(rec, privKey)
|
|
if err != nil {
|
|
t.Fatalf("error creating signed peer record: %s", err)
|
|
}
|
|
|
|
cab, ok := peerstore.GetCertifiedAddrBook(hosts[0].Peerstore())
|
|
if !ok {
|
|
t.Fatal("peerstore does not implement CertifiedAddrBook")
|
|
}
|
|
_, err = cab.ConsumePeerRecord(signedRec, peerstore.PermanentAddrTTL)
|
|
if err != nil {
|
|
t.Fatalf("error adding signed peer record: %s", err)
|
|
}
|
|
}
|
|
|
|
// build the star
|
|
for i := 1; i < 20; i++ {
|
|
connect(t, hosts[0], hosts[i])
|
|
}
|
|
|
|
time.Sleep(time.Second)
|
|
|
|
// build the mesh
|
|
var subs []*Subscription
|
|
for _, ps := range psubs {
|
|
sub, err := ps.Subscribe([]byte{0xf1, 0x22, 0xb0, 0x22})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
subs = append(subs, sub)
|
|
}
|
|
|
|
// wait a bit for the mesh to build
|
|
time.Sleep(10 * time.Second)
|
|
|
|
// check that all peers have > 1 connection
|
|
for i, h := range hosts {
|
|
if len(h.Network().Conns()) == 1 {
|
|
t.Errorf("peer %d has ony a single connection", i)
|
|
}
|
|
}
|
|
|
|
// send a message from each peer and assert it was propagated
|
|
for i := 0; i < 20; i++ {
|
|
msg := []byte(fmt.Sprintf("message %d", i))
|
|
psubs[i].Publish([]byte{0xf1, 0x22, 0xb0, 0x22}, msg)
|
|
|
|
for _, sub := range subs {
|
|
assertReceive(t, sub, msg)
|
|
}
|
|
}
|
|
}
|
|
|
|
func TestBlossomSubDirectPeers(t *testing.T) {
|
|
ctx, cancel := context.WithCancel(context.Background())
|
|
defer cancel()
|
|
|
|
h := getNetHosts(t, ctx, 3)
|
|
psubs := []*PubSub{
|
|
getBlossomSub(ctx, h[0], WithDirectConnectTicks(2)),
|
|
getBlossomSub(ctx, h[1], WithDirectPeers([]peer.AddrInfo{{ID: h[2].ID(), Addrs: h[2].Addrs()}}), WithDirectConnectTicks(2)),
|
|
getBlossomSub(ctx, h[2], WithDirectPeers([]peer.AddrInfo{{ID: h[1].ID(), Addrs: h[1].Addrs()}}), WithDirectConnectTicks(2)),
|
|
}
|
|
|
|
connect(t, h[0], h[1])
|
|
connect(t, h[0], h[2])
|
|
|
|
// verify that the direct peers connected
|
|
time.Sleep(2 * time.Second)
|
|
if len(h[1].Network().ConnsToPeer(h[2].ID())) == 0 {
|
|
t.Fatal("expected a connection between direct peers")
|
|
}
|
|
|
|
// build the mesh
|
|
var subs []*Subscription
|
|
for _, ps := range psubs {
|
|
sub, err := ps.Subscribe([]byte{0xf1, 0x22, 0xb0, 0x22})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
subs = append(subs, sub)
|
|
}
|
|
|
|
time.Sleep(time.Second)
|
|
|
|
// publish some messages
|
|
for i := 0; i < 3; i++ {
|
|
msg := []byte(fmt.Sprintf("message %d", i))
|
|
psubs[i].Publish([]byte{0xf1, 0x22, 0xb0, 0x22}, msg)
|
|
|
|
for _, sub := range subs {
|
|
assertReceive(t, sub, msg)
|
|
}
|
|
}
|
|
|
|
// disconnect the direct peers to test reconnection
|
|
for _, c := range h[1].Network().ConnsToPeer(h[2].ID()) {
|
|
c.Close()
|
|
}
|
|
|
|
time.Sleep(5 * time.Second)
|
|
|
|
if len(h[1].Network().ConnsToPeer(h[2].ID())) == 0 {
|
|
t.Fatal("expected a connection between direct peers")
|
|
}
|
|
|
|
// publish some messages
|
|
for i := 0; i < 3; i++ {
|
|
msg := []byte(fmt.Sprintf("message %d", i))
|
|
psubs[i].Publish([]byte{0xf1, 0x22, 0xb0, 0x22}, msg)
|
|
|
|
for _, sub := range subs {
|
|
assertReceive(t, sub, msg)
|
|
}
|
|
}
|
|
}
|
|
|
|
func TestBlossomSubPeerFilter(t *testing.T) {
|
|
ctx, cancel := context.WithCancel(context.Background())
|
|
defer cancel()
|
|
|
|
h := getNetHosts(t, ctx, 3)
|
|
psubs := []*PubSub{
|
|
getBlossomSub(ctx, h[0], WithPeerFilter(func(pid peer.ID, bitmask []byte) bool {
|
|
return pid == h[1].ID()
|
|
})),
|
|
getBlossomSub(ctx, h[1], WithPeerFilter(func(pid peer.ID, bitmask []byte) bool {
|
|
return pid == h[0].ID()
|
|
})),
|
|
getBlossomSub(ctx, h[2]),
|
|
}
|
|
|
|
connect(t, h[0], h[1])
|
|
connect(t, h[0], h[2])
|
|
|
|
// Join all peers
|
|
var subs []*Subscription
|
|
for _, ps := range psubs {
|
|
sub, err := ps.Subscribe([]byte{0xf1, 0x22, 0xb0, 0x22})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
subs = append(subs, sub)
|
|
}
|
|
|
|
time.Sleep(time.Second)
|
|
|
|
msg := []byte("message")
|
|
|
|
psubs[0].Publish([]byte{0xf1, 0x22, 0xb0, 0x22}, msg)
|
|
assertReceive(t, subs[1], msg)
|
|
assertNeverReceives(t, subs[2], time.Second)
|
|
|
|
psubs[1].Publish([]byte{0xf1, 0x22, 0xb0, 0x22}, msg)
|
|
assertReceive(t, subs[0], msg)
|
|
assertNeverReceives(t, subs[2], time.Second)
|
|
}
|
|
|
|
func TestBlossomSubDirectPeersFanout(t *testing.T) {
|
|
// regression test for #371
|
|
ctx, cancel := context.WithCancel(context.Background())
|
|
defer cancel()
|
|
|
|
h := getNetHosts(t, ctx, 3)
|
|
psubs := []*PubSub{
|
|
getBlossomSub(ctx, h[0]),
|
|
getBlossomSub(ctx, h[1], WithDirectPeers([]peer.AddrInfo{{ID: h[2].ID(), Addrs: h[2].Addrs()}})),
|
|
getBlossomSub(ctx, h[2], WithDirectPeers([]peer.AddrInfo{{ID: h[1].ID(), Addrs: h[1].Addrs()}})),
|
|
}
|
|
|
|
connect(t, h[0], h[1])
|
|
connect(t, h[0], h[2])
|
|
|
|
// Join all peers except h2
|
|
var subs []*Subscription
|
|
for _, ps := range psubs[:2] {
|
|
sub, err := ps.Subscribe([]byte{0xf1, 0x22, 0xb0, 0x22})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
subs = append(subs, sub)
|
|
}
|
|
|
|
time.Sleep(time.Second)
|
|
|
|
// h2 publishes some messages to build a fanout
|
|
for i := 0; i < 3; i++ {
|
|
msg := []byte(fmt.Sprintf("message %d", i))
|
|
psubs[2].Publish([]byte{0xf1, 0x22, 0xb0, 0x22}, msg)
|
|
|
|
for _, sub := range subs {
|
|
assertReceive(t, sub, msg)
|
|
}
|
|
}
|
|
|
|
// verify that h0 is in the fanout of h2, but not h1 who is a direct peer
|
|
result := make(chan bool, 2)
|
|
psubs[2].eval <- func() {
|
|
rt := psubs[2].rt.(*BlossomSubRouter)
|
|
fanout := rt.fanout[string([]byte{0xf1, 0x22, 0xb0, 0x22})]
|
|
_, ok := fanout[h[0].ID()]
|
|
result <- ok
|
|
_, ok = fanout[h[1].ID()]
|
|
result <- ok
|
|
}
|
|
|
|
inFanout := <-result
|
|
if !inFanout {
|
|
t.Fatal("expected peer 0 to be in fanout")
|
|
}
|
|
|
|
inFanout = <-result
|
|
if inFanout {
|
|
t.Fatal("expected peer 1 to not be in fanout")
|
|
}
|
|
|
|
// now subscribe h2 too and verify tht h0 is in the mesh but not h1
|
|
_, err := psubs[2].Subscribe([]byte{0xf1, 0x22, 0xb0, 0x22})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
|
|
time.Sleep(2 * time.Second)
|
|
|
|
psubs[2].eval <- func() {
|
|
rt := psubs[2].rt.(*BlossomSubRouter)
|
|
mesh := rt.mesh[string([]byte{0xf1, 0x22, 0xb0, 0x22})]
|
|
_, ok := mesh[h[0].ID()]
|
|
result <- ok
|
|
_, ok = mesh[h[1].ID()]
|
|
result <- ok
|
|
}
|
|
|
|
inMesh := <-result
|
|
if !inMesh {
|
|
t.Fatal("expected peer 0 to be in mesh")
|
|
}
|
|
|
|
inMesh = <-result
|
|
if inMesh {
|
|
t.Fatal("expected peer 1 to not be in mesh")
|
|
}
|
|
}
|
|
|
|
func TestBlossomSubFloodPublish(t *testing.T) {
|
|
// uses a star topology without PX and publishes from the star to verify that all
|
|
// messages get received
|
|
ctx, cancel := context.WithCancel(context.Background())
|
|
defer cancel()
|
|
|
|
hosts := getNetHosts(t, ctx, 20)
|
|
psubs := getBlossomSubs(ctx, hosts, WithFloodPublish(true))
|
|
|
|
// build the star
|
|
for i := 1; i < 20; i++ {
|
|
connect(t, hosts[0], hosts[i])
|
|
}
|
|
|
|
// build the (partial, unstable) mesh
|
|
var subs []*Subscription
|
|
for _, ps := range psubs {
|
|
sub, err := ps.Subscribe([]byte{0xf1, 0x22, 0xb0, 0x22})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
subs = append(subs, sub)
|
|
}
|
|
|
|
time.Sleep(time.Second)
|
|
|
|
// send a message from the star and assert it was received
|
|
for i := 0; i < 20; i++ {
|
|
msg := []byte(fmt.Sprintf("message %d", i))
|
|
psubs[0].Publish([]byte{0xf1, 0x22, 0xb0, 0x22}, msg)
|
|
|
|
for _, sub := range subs {
|
|
assertReceive(t, sub, msg)
|
|
}
|
|
}
|
|
}
|
|
|
|
func TestBlossomSubEnoughPeers(t *testing.T) {
|
|
ctx, cancel := context.WithCancel(context.Background())
|
|
defer cancel()
|
|
|
|
hosts := getNetHosts(t, ctx, 20)
|
|
psubs := getBlossomSubs(ctx, hosts)
|
|
|
|
for _, ps := range psubs {
|
|
_, err := ps.Subscribe([]byte{0xf1, 0x22, 0xb0, 0x22})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
}
|
|
|
|
// at this point we have no connections and no mesh, so EnoughPeers should return false
|
|
res := make(chan bool, 1)
|
|
psubs[0].eval <- func() {
|
|
res <- psubs[0].rt.EnoughPeers([]byte{0xf1, 0x22, 0xb0, 0x22}, 0)
|
|
}
|
|
enough := <-res
|
|
if enough {
|
|
t.Fatal("should not have enough peers")
|
|
}
|
|
|
|
// connect them densly to build up the mesh
|
|
denseConnect(t, hosts)
|
|
|
|
time.Sleep(3 * time.Second)
|
|
|
|
psubs[0].eval <- func() {
|
|
res <- psubs[0].rt.EnoughPeers([]byte{0xf1, 0x22, 0xb0, 0x22}, 0)
|
|
}
|
|
enough = <-res
|
|
if !enough {
|
|
t.Fatal("should have enough peers")
|
|
}
|
|
}
|
|
|
|
func TestBlossomSubCustomParams(t *testing.T) {
|
|
// in this test we score sinkhole a peer to exercise code paths relative to negative scores
|
|
ctx, cancel := context.WithCancel(context.Background())
|
|
defer cancel()
|
|
|
|
params := DefaultBlossomSubParams()
|
|
|
|
wantedFollowTime := 1 * time.Second
|
|
params.IWantFollowupTime = wantedFollowTime
|
|
|
|
customGossipFactor := 0.12
|
|
params.GossipFactor = customGossipFactor
|
|
|
|
wantedMaxPendingConns := 23
|
|
params.MaxPendingConnections = wantedMaxPendingConns
|
|
hosts := getNetHosts(t, ctx, 1)
|
|
psubs := getBlossomSubs(ctx, hosts,
|
|
WithBlossomSubParams(params))
|
|
|
|
if len(psubs) != 1 {
|
|
t.Fatalf("incorrect number of pusbub objects received: wanted %d but got %d", 1, len(psubs))
|
|
}
|
|
|
|
rt, ok := psubs[0].rt.(*BlossomSubRouter)
|
|
if !ok {
|
|
t.Fatal("Did not get gossip sub router from pub sub object")
|
|
}
|
|
|
|
if rt.params.IWantFollowupTime != wantedFollowTime {
|
|
t.Errorf("Wanted %d of param BlossomSubIWantFollowupTime but got %d", wantedFollowTime, rt.params.IWantFollowupTime)
|
|
}
|
|
if rt.params.GossipFactor != customGossipFactor {
|
|
t.Errorf("Wanted %f of param BlossomSubGossipFactor but got %f", customGossipFactor, rt.params.GossipFactor)
|
|
}
|
|
if rt.params.MaxPendingConnections != wantedMaxPendingConns {
|
|
t.Errorf("Wanted %d of param BlossomSubMaxPendingConnections but got %d", wantedMaxPendingConns, rt.params.MaxPendingConnections)
|
|
}
|
|
}
|
|
|
|
func TestBlossomSubNegativeScore(t *testing.T) {
|
|
// in this test we score sinkhole a peer to exercise code paths relative to negative scores
|
|
ctx, cancel := context.WithCancel(context.Background())
|
|
defer cancel()
|
|
|
|
hosts := getNetHosts(t, ctx, 20)
|
|
psubs := getBlossomSubs(ctx, hosts,
|
|
WithPeerScore(
|
|
&PeerScoreParams{
|
|
AppSpecificScore: func(p peer.ID) float64 {
|
|
if p == hosts[0].ID() {
|
|
return -1000
|
|
} else {
|
|
return 0
|
|
}
|
|
},
|
|
AppSpecificWeight: 1,
|
|
DecayInterval: time.Second,
|
|
DecayToZero: 0.01,
|
|
},
|
|
&PeerScoreThresholds{
|
|
GossipThreshold: -10,
|
|
PublishThreshold: -100,
|
|
GraylistThreshold: -10000,
|
|
}))
|
|
|
|
denseConnect(t, hosts)
|
|
|
|
var subs []*Subscription
|
|
for _, ps := range psubs {
|
|
sub, err := ps.Subscribe([]byte{0xf1, 0x22, 0xb0, 0x22})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
subs = append(subs, sub)
|
|
}
|
|
|
|
time.Sleep(3 * time.Second)
|
|
|
|
for i := 0; i < 20; i++ {
|
|
msg := []byte(fmt.Sprintf("message %d", i))
|
|
psubs[i%20].Publish([]byte{0xf1, 0x22, 0xb0, 0x22}, msg)
|
|
time.Sleep(20 * time.Millisecond)
|
|
}
|
|
|
|
// let the sinkholed peer try to emit gossip as well
|
|
time.Sleep(2 * time.Second)
|
|
|
|
// checks:
|
|
// 1. peer 0 should only receive its own message
|
|
// 2. peers 1-20 should not receive a message from peer 0, because it's not part of the mesh
|
|
// and its gossip is rejected
|
|
collectAll := func(sub *Subscription) []*Message {
|
|
var res []*Message
|
|
ctx, cancel := context.WithTimeout(context.Background(), 100*time.Millisecond)
|
|
defer cancel()
|
|
|
|
for {
|
|
msg, err := sub.Next(ctx)
|
|
if err != nil {
|
|
break
|
|
}
|
|
|
|
res = append(res, msg)
|
|
}
|
|
|
|
return res
|
|
}
|
|
|
|
count := len(collectAll(subs[0]))
|
|
if count != 1 {
|
|
t.Fatalf("expected 1 message but got %d instead", count)
|
|
}
|
|
|
|
for _, sub := range subs[1:] {
|
|
all := collectAll(sub)
|
|
for _, m := range all {
|
|
if m.ReceivedFrom == hosts[0].ID() {
|
|
t.Fatal("received message from sinkholed peer")
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
func TestBlossomSubScoreValidatorEx(t *testing.T) {
|
|
// this is a test that of the two message drop responses from a validator
|
|
ctx, cancel := context.WithCancel(context.Background())
|
|
defer cancel()
|
|
|
|
hosts := getNetHosts(t, ctx, 3)
|
|
psubs := getBlossomSubs(ctx, hosts,
|
|
WithPeerScore(
|
|
&PeerScoreParams{
|
|
AppSpecificScore: func(p peer.ID) float64 { return 0 },
|
|
DecayInterval: time.Second,
|
|
DecayToZero: 0.01,
|
|
Bitmasks: map[string]*BitmaskScoreParams{
|
|
string([]byte{0xf1, 0x22, 0xb0, 0x22}): {
|
|
BitmaskWeight: 1,
|
|
TimeInMeshQuantum: time.Second,
|
|
InvalidMessageDeliveriesWeight: -1,
|
|
InvalidMessageDeliveriesDecay: 0.9999,
|
|
},
|
|
},
|
|
},
|
|
&PeerScoreThresholds{
|
|
GossipThreshold: -10,
|
|
PublishThreshold: -100,
|
|
GraylistThreshold: -10000,
|
|
}))
|
|
|
|
connectAll(t, hosts)
|
|
|
|
err := psubs[0].RegisterBitmaskValidator([]byte{0xf1, 0x22, 0xb0, 0x22}, func(ctx context.Context, p peer.ID, msg *Message) ValidationResult {
|
|
// we ignore host1 and reject host2
|
|
if p == hosts[1].ID() {
|
|
return ValidationIgnore
|
|
}
|
|
if p == hosts[2].ID() {
|
|
return ValidationReject
|
|
}
|
|
|
|
return ValidationAccept
|
|
})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
|
|
sub, err := psubs[0].Subscribe([]byte{0xf1, 0x22, 0xb0, 0x22})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
|
|
time.Sleep(100 * time.Millisecond)
|
|
|
|
expectNoMessage := func(sub *Subscription) {
|
|
ctx, cancel := context.WithTimeout(ctx, 100*time.Millisecond)
|
|
defer cancel()
|
|
|
|
m, err := sub.Next(ctx)
|
|
if err == nil {
|
|
t.Fatal("expected no message, but got ", string(m.Data))
|
|
}
|
|
}
|
|
|
|
psubs[1].Publish([]byte{0xf1, 0x22, 0xb0, 0x22}, []byte("i am not a walrus"))
|
|
psubs[2].Publish([]byte{0xf1, 0x22, 0xb0, 0x22}, []byte("i am not a walrus either"))
|
|
|
|
// assert no messages
|
|
expectNoMessage(sub)
|
|
|
|
// assert that peer1's score is still 0 (its message was ignored) while peer2 should have
|
|
// a negative score (its message got rejected)
|
|
res := make(chan float64, 1)
|
|
psubs[0].eval <- func() {
|
|
res <- psubs[0].rt.(*BlossomSubRouter).score.Score(hosts[1].ID())
|
|
}
|
|
score := <-res
|
|
if score != 0 {
|
|
t.Fatalf("expected 0 score for peer1, but got %f", score)
|
|
}
|
|
|
|
psubs[0].eval <- func() {
|
|
res <- psubs[0].rt.(*BlossomSubRouter).score.Score(hosts[2].ID())
|
|
}
|
|
score = <-res
|
|
if score >= 0 {
|
|
t.Fatalf("expected negative score for peer2, but got %f", score)
|
|
}
|
|
}
|
|
|
|
func TestBlossomSubPiggybackControl(t *testing.T) {
|
|
// this is a direct test of the piggybackControl function as we can't reliably
|
|
// trigger it on travis
|
|
ctx, cancel := context.WithCancel(context.Background())
|
|
defer cancel()
|
|
|
|
h := bhost.NewBlankHost(swarmt.GenSwarm(t))
|
|
defer h.Close()
|
|
ps := getBlossomSub(ctx, h)
|
|
|
|
blah := peer.ID("bogotr0n")
|
|
|
|
res := make(chan *RPC, 1)
|
|
ps.eval <- func() {
|
|
gs := ps.rt.(*BlossomSubRouter)
|
|
test1 := []byte{0xff, 0x00, 0x00, 0x00}
|
|
test2 := []byte{0x00, 0xff, 0x00, 0x00}
|
|
test3 := []byte{0x00, 0x00, 0xff, 0x00}
|
|
gs.mesh[string(test1)] = make(map[peer.ID]struct{})
|
|
gs.mesh[string(test2)] = make(map[peer.ID]struct{})
|
|
gs.mesh[string(test1)][blah] = struct{}{}
|
|
|
|
rpc := &RPC{RPC: pb.RPC{}}
|
|
gs.piggybackControl(blah, rpc, &pb.ControlMessage{
|
|
Graft: []*pb.ControlGraft{{Bitmask: test1}, {Bitmask: test2}, {Bitmask: test3}},
|
|
Prune: []*pb.ControlPrune{{Bitmask: test1}, {Bitmask: test2}, {Bitmask: test3}},
|
|
})
|
|
res <- rpc
|
|
}
|
|
|
|
rpc := <-res
|
|
if rpc.Control == nil {
|
|
t.Fatal("expected non-nil control message")
|
|
}
|
|
if len(rpc.Control.Graft) != 1 {
|
|
t.Fatal("expected 1 GRAFT")
|
|
}
|
|
if !bytes.Equal(rpc.Control.Graft[0].GetBitmask(), []byte{0xff, 0x00, 0x00, 0x00}) {
|
|
t.Fatal("expected test1 as graft bitmask ID")
|
|
}
|
|
if len(rpc.Control.Prune) != 2 {
|
|
t.Fatal("expected 2 PRUNEs")
|
|
}
|
|
if !bytes.Equal(rpc.Control.Prune[0].GetBitmask(), []byte{0x00, 0xff, 0x00, 0x00}) {
|
|
t.Fatal("expected test2 as prune bitmask ID")
|
|
}
|
|
if !bytes.Equal(rpc.Control.Prune[1].GetBitmask(), []byte{0x00, 0x00, 0xff, 0x00}) {
|
|
t.Fatal("expected test3 as prune bitmask ID")
|
|
}
|
|
}
|
|
|
|
func TestBlossomSubMultipleGraftBitmasks(t *testing.T) {
|
|
ctx, cancel := context.WithCancel(context.Background())
|
|
defer cancel()
|
|
|
|
hosts := getNetHosts(t, ctx, 2)
|
|
psubs := getBlossomSubs(ctx, hosts)
|
|
sparseConnect(t, hosts)
|
|
|
|
time.Sleep(time.Second * 1)
|
|
|
|
firstBitmask := []byte{0xff, 0x00, 0x00, 0x00}
|
|
secondBitmask := []byte{0x00, 0xff, 0x00, 0x00}
|
|
thirdBitmask := []byte{0x00, 0x00, 0xff, 0x00}
|
|
|
|
firstPeer := hosts[0].ID()
|
|
secondPeer := hosts[1].ID()
|
|
|
|
p2Sub := psubs[1]
|
|
p1Router := psubs[0].rt.(*BlossomSubRouter)
|
|
p2Router := psubs[1].rt.(*BlossomSubRouter)
|
|
|
|
finChan := make(chan struct{})
|
|
|
|
p2Sub.eval <- func() {
|
|
// Add bitmasks to second peer
|
|
p2Router.mesh[string(firstBitmask)] = map[peer.ID]struct{}{}
|
|
p2Router.mesh[string(secondBitmask)] = map[peer.ID]struct{}{}
|
|
p2Router.mesh[string(thirdBitmask)] = map[peer.ID]struct{}{}
|
|
|
|
finChan <- struct{}{}
|
|
}
|
|
<-finChan
|
|
|
|
// Send multiple GRAFT messages to second peer from
|
|
// 1st peer
|
|
p1Router.sendGraftPrune(map[peer.ID][][]byte{
|
|
secondPeer: {firstBitmask, secondBitmask, thirdBitmask},
|
|
}, map[peer.ID][][]byte{}, map[peer.ID]bool{})
|
|
|
|
time.Sleep(time.Second * 1)
|
|
|
|
p2Sub.eval <- func() {
|
|
if _, ok := p2Router.mesh[string(firstBitmask)][firstPeer]; !ok {
|
|
t.Errorf("First peer wasnt added to mesh of the second peer for the bitmask %s", firstBitmask)
|
|
}
|
|
if _, ok := p2Router.mesh[string(secondBitmask)][firstPeer]; !ok {
|
|
t.Errorf("First peer wasnt added to mesh of the second peer for the bitmask %s", secondBitmask)
|
|
}
|
|
if _, ok := p2Router.mesh[string(thirdBitmask)][firstPeer]; !ok {
|
|
t.Errorf("First peer wasnt added to mesh of the second peer for the bitmask %s", thirdBitmask)
|
|
}
|
|
finChan <- struct{}{}
|
|
}
|
|
<-finChan
|
|
}
|
|
|
|
func TestBlossomSubOpportunisticGrafting(t *testing.T) {
|
|
originalBlossomSubPruneBackoff := BlossomSubPruneBackoff
|
|
BlossomSubPruneBackoff = 500 * time.Millisecond
|
|
originalBlossomSubGraftFloodThreshold := BlossomSubGraftFloodThreshold
|
|
BlossomSubGraftFloodThreshold = 100 * time.Millisecond
|
|
originalBlossomSubOpportunisticGraftTicks := BlossomSubOpportunisticGraftTicks
|
|
BlossomSubOpportunisticGraftTicks = 2
|
|
defer func() {
|
|
BlossomSubPruneBackoff = originalBlossomSubPruneBackoff
|
|
BlossomSubGraftFloodThreshold = originalBlossomSubGraftFloodThreshold
|
|
BlossomSubOpportunisticGraftTicks = originalBlossomSubOpportunisticGraftTicks
|
|
}()
|
|
|
|
ctx, cancel := context.WithCancel(context.Background())
|
|
defer cancel()
|
|
|
|
hosts := getNetHosts(t, ctx, 50)
|
|
// pubsubs for the first 10 hosts
|
|
psubs := getBlossomSubs(ctx, hosts[:10],
|
|
WithFloodPublish(true),
|
|
WithPeerScore(
|
|
&PeerScoreParams{
|
|
AppSpecificScore: func(peer.ID) float64 { return 0 },
|
|
AppSpecificWeight: 0,
|
|
DecayInterval: time.Second,
|
|
DecayToZero: 0.01,
|
|
Bitmasks: map[string]*BitmaskScoreParams{
|
|
string([]byte{0xf1, 0x22, 0xb0, 0x22}): {
|
|
BitmaskWeight: 1,
|
|
TimeInMeshWeight: 0.0002777,
|
|
TimeInMeshQuantum: time.Second,
|
|
TimeInMeshCap: 3600,
|
|
FirstMessageDeliveriesWeight: 1,
|
|
FirstMessageDeliveriesDecay: 0.9997,
|
|
FirstMessageDeliveriesCap: 100,
|
|
InvalidMessageDeliveriesDecay: 0.99997,
|
|
},
|
|
},
|
|
},
|
|
&PeerScoreThresholds{
|
|
GossipThreshold: -10,
|
|
PublishThreshold: -100,
|
|
GraylistThreshold: -10000,
|
|
OpportunisticGraftThreshold: 1,
|
|
}))
|
|
|
|
// connect the real hosts with degree 5
|
|
connectSome(t, hosts[:10], 5)
|
|
|
|
// sybil squatters for the remaining 40 hosts
|
|
for _, h := range hosts[10:] {
|
|
squatter := &sybilSquatter{h: h}
|
|
h.SetStreamHandler(BlossomSubID_v12, squatter.handleStream)
|
|
}
|
|
|
|
// connect all squatters to every real host
|
|
for _, squatter := range hosts[10:] {
|
|
for _, real := range hosts[:10] {
|
|
connect(t, squatter, real)
|
|
}
|
|
}
|
|
|
|
// wait a bit for the connections to propagate events to the pubsubs
|
|
time.Sleep(time.Second)
|
|
|
|
// ask the real pubsus to join the bitmask
|
|
for _, ps := range psubs {
|
|
sub, err := ps.Subscribe([]byte{0xf1, 0x22, 0xb0, 0x22})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
// consume the messages
|
|
go func(sub *Subscription) {
|
|
for {
|
|
_, err := sub.Next(ctx)
|
|
if err != nil {
|
|
return
|
|
}
|
|
}
|
|
}(sub)
|
|
}
|
|
|
|
// publish a bunch of messages from the real hosts
|
|
for i := 0; i < 1000; i++ {
|
|
msg := []byte(fmt.Sprintf("message %d", i))
|
|
psubs[i%10].Publish([]byte{0xf1, 0x22, 0xb0, 0x22}, msg)
|
|
time.Sleep(20 * time.Millisecond)
|
|
}
|
|
|
|
// now wait a few of oppgraft cycles
|
|
time.Sleep(7 * time.Second)
|
|
|
|
// check the honest peer meshes, they should have at least 3 honest peers each
|
|
res := make(chan int, 1)
|
|
for _, ps := range psubs {
|
|
ps.eval <- func() {
|
|
gs := ps.rt.(*BlossomSubRouter)
|
|
count := 0
|
|
for _, h := range hosts[:10] {
|
|
_, ok := gs.mesh[string([]byte{0xf1, 0x22, 0xb0, 0x22})][h.ID()]
|
|
if ok {
|
|
count++
|
|
}
|
|
}
|
|
res <- count
|
|
}
|
|
|
|
count := <-res
|
|
if count < 3 {
|
|
t.Fatalf("expected at least 3 honest peers, got %d", count)
|
|
}
|
|
}
|
|
}
|
|
func TestBlossomSubLeaveBitmask(t *testing.T) {
|
|
ctx, cancel := context.WithCancel(context.Background())
|
|
defer cancel()
|
|
|
|
h := getNetHosts(t, ctx, 2)
|
|
psubs := []*PubSub{
|
|
getBlossomSub(ctx, h[0]),
|
|
getBlossomSub(ctx, h[1]),
|
|
}
|
|
|
|
connect(t, h[0], h[1])
|
|
|
|
// Join all peers
|
|
var subs []*Subscription
|
|
for _, ps := range psubs {
|
|
sub, err := ps.Subscribe([]byte{0xf1, 0x22, 0xb0, 0x22})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
subs = append(subs, sub)
|
|
}
|
|
|
|
time.Sleep(time.Second)
|
|
|
|
leaveTime := time.Now()
|
|
done := make(chan struct{})
|
|
|
|
psubs[0].rt.(*BlossomSubRouter).p.eval <- func() {
|
|
defer close(done)
|
|
psubs[0].rt.Leave([]byte{0xf1, 0x22, 0xb0, 0x22})
|
|
time.Sleep(time.Second)
|
|
peerMap := psubs[0].rt.(*BlossomSubRouter).backoff[string([]byte{0xf1, 0x22, 0xb0, 0x22})]
|
|
if len(peerMap) != 1 {
|
|
t.Fatalf("No peer is populated in the backoff map for peer 0")
|
|
}
|
|
_, ok := peerMap[h[1].ID()]
|
|
if !ok {
|
|
t.Errorf("Expected peer does not exist in the backoff map")
|
|
}
|
|
|
|
backoffTime := peerMap[h[1].ID()].Sub(leaveTime)
|
|
// Check that the backoff time is roughly the unsubscribebackoff time (with a slack of 1s)
|
|
if backoffTime-BlossomSubUnsubscribeBackoff > time.Second {
|
|
t.Error("Backoff time should be set to BlossomSubUnsubscribeBackoff.")
|
|
}
|
|
}
|
|
<-done
|
|
|
|
done = make(chan struct{})
|
|
// Ensure that remote peer 1 also applies the backoff appropriately
|
|
// for peer 0.
|
|
psubs[1].rt.(*BlossomSubRouter).p.eval <- func() {
|
|
defer close(done)
|
|
peerMap2 := psubs[1].rt.(*BlossomSubRouter).backoff[string([]byte{0xf1, 0x22, 0xb0, 0x22})]
|
|
if len(peerMap2) != 1 {
|
|
t.Fatalf("No peer is populated in the backoff map for peer 1")
|
|
}
|
|
_, ok := peerMap2[h[0].ID()]
|
|
if !ok {
|
|
t.Errorf("Expected peer does not exist in the backoff map")
|
|
}
|
|
|
|
backoffTime := peerMap2[h[0].ID()].Sub(leaveTime)
|
|
// Check that the backoff time is roughly the unsubscribebackoff time (with a slack of 1s)
|
|
if backoffTime-BlossomSubUnsubscribeBackoff > time.Second {
|
|
t.Error("Backoff time should be set to BlossomSubUnsubscribeBackoff.")
|
|
}
|
|
}
|
|
<-done
|
|
}
|
|
|
|
func TestBlossomSubJoinBitmask(t *testing.T) {
|
|
ctx, cancel := context.WithCancel(context.Background())
|
|
defer cancel()
|
|
|
|
h := getNetHosts(t, ctx, 3)
|
|
psubs := []*PubSub{
|
|
getBlossomSub(ctx, h[0]),
|
|
getBlossomSub(ctx, h[1]),
|
|
getBlossomSub(ctx, h[2]),
|
|
}
|
|
|
|
connect(t, h[0], h[1])
|
|
connect(t, h[0], h[2])
|
|
|
|
router0 := psubs[0].rt.(*BlossomSubRouter)
|
|
|
|
// Add in backoff for peer.
|
|
peerMap := make(map[peer.ID]time.Time)
|
|
peerMap[h[1].ID()] = time.Now().Add(router0.params.UnsubscribeBackoff)
|
|
|
|
router0.backoff[string([]byte{0xf1, 0x22, 0xb0, 0x22})] = peerMap
|
|
|
|
// Join all peers
|
|
var subs []*Subscription
|
|
for _, ps := range psubs {
|
|
sub, err := ps.Subscribe([]byte{0xf1, 0x22, 0xb0, 0x22})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
subs = append(subs, sub)
|
|
}
|
|
|
|
time.Sleep(time.Second)
|
|
|
|
meshMap := router0.mesh[string([]byte{0xf1, 0x22, 0xb0, 0x22})]
|
|
if len(meshMap) != 1 {
|
|
t.Fatalf("Unexpect peer included in the mesh")
|
|
}
|
|
|
|
_, ok := meshMap[h[1].ID()]
|
|
if ok {
|
|
t.Fatalf("Peer that was to be backed off is included in the mesh")
|
|
}
|
|
}
|
|
|
|
type sybilSquatter struct {
|
|
h host.Host
|
|
}
|
|
|
|
func (sq *sybilSquatter) handleStream(s network.Stream) {
|
|
defer s.Close()
|
|
|
|
os, err := sq.h.NewStream(context.Background(), s.Conn().RemotePeer(), BlossomSubID_v12)
|
|
if err != nil {
|
|
panic(err)
|
|
}
|
|
|
|
// send a subscription for test in the output stream to become candidate for GRAFT
|
|
// and then just read and ignore the incoming RPCs
|
|
r := protoio.NewDelimitedReader(s, 1<<20)
|
|
w := protoio.NewDelimitedWriter(os)
|
|
truth := true
|
|
bitmask := []byte{0xf1, 0x22, 0xb0, 0x22}
|
|
err = w.WriteMsg(&pb.RPC{Subscriptions: []*pb.RPC_SubOpts{{Subscribe: truth, Bitmask: bitmask}}})
|
|
if err != nil {
|
|
panic(err)
|
|
}
|
|
|
|
var rpc pb.RPC
|
|
for {
|
|
rpc.Reset()
|
|
err = r.ReadMsg(&rpc)
|
|
if err != nil {
|
|
if err != io.EOF {
|
|
s.Reset()
|
|
}
|
|
return
|
|
}
|
|
}
|
|
}
|
|
|
|
func TestBlossomSubPeerScoreInspect(t *testing.T) {
|
|
// this test exercises the code path sof peer score inspection
|
|
ctx, cancel := context.WithCancel(context.Background())
|
|
defer cancel()
|
|
|
|
hosts := getNetHosts(t, ctx, 2)
|
|
|
|
inspector := &mockPeerScoreInspector{}
|
|
psub1 := getBlossomSub(ctx, hosts[0],
|
|
WithPeerScore(
|
|
&PeerScoreParams{
|
|
Bitmasks: map[string]*BitmaskScoreParams{
|
|
string([]byte{0xf1, 0x22, 0xb0, 0x22}): {
|
|
BitmaskWeight: 1,
|
|
TimeInMeshQuantum: time.Second,
|
|
FirstMessageDeliveriesWeight: 1,
|
|
FirstMessageDeliveriesDecay: 0.999,
|
|
FirstMessageDeliveriesCap: 100,
|
|
InvalidMessageDeliveriesWeight: -1,
|
|
InvalidMessageDeliveriesDecay: 0.9999,
|
|
},
|
|
},
|
|
AppSpecificScore: func(peer.ID) float64 { return 0 },
|
|
DecayInterval: time.Second,
|
|
DecayToZero: 0.01,
|
|
},
|
|
&PeerScoreThresholds{
|
|
GossipThreshold: -1,
|
|
PublishThreshold: -10,
|
|
GraylistThreshold: -1000,
|
|
}),
|
|
WithPeerScoreInspect(inspector.inspect, time.Second))
|
|
psub2 := getBlossomSub(ctx, hosts[1])
|
|
psubs := []*PubSub{psub1, psub2}
|
|
|
|
connect(t, hosts[0], hosts[1])
|
|
|
|
for _, ps := range psubs {
|
|
_, err := ps.Subscribe([]byte{0xf1, 0x22, 0xb0, 0x22})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
}
|
|
|
|
time.Sleep(time.Second)
|
|
|
|
for i := 0; i < 20; i++ {
|
|
msg := []byte(fmt.Sprintf("message %d", i))
|
|
psubs[i%2].Publish([]byte{0xf1, 0x22, 0xb0, 0x22}, msg)
|
|
time.Sleep(20 * time.Millisecond)
|
|
}
|
|
|
|
time.Sleep(time.Second + 200*time.Millisecond)
|
|
|
|
score2 := inspector.score(hosts[1].ID())
|
|
if score2 < 9 {
|
|
t.Fatalf("expected score to be at least 9, instead got %f", score2)
|
|
}
|
|
}
|
|
|
|
func TestBlossomSubPeerScoreResetBitmaskParams(t *testing.T) {
|
|
// this test exercises the code path sof peer score inspection
|
|
ctx, cancel := context.WithCancel(context.Background())
|
|
defer cancel()
|
|
|
|
hosts := getNetHosts(t, ctx, 1)
|
|
|
|
ps := getBlossomSub(ctx, hosts[0],
|
|
WithPeerScore(
|
|
&PeerScoreParams{
|
|
Bitmasks: map[string]*BitmaskScoreParams{
|
|
string([]byte{0xf1, 0x22, 0xb0, 0x22}): {
|
|
BitmaskWeight: 1,
|
|
TimeInMeshQuantum: time.Second,
|
|
FirstMessageDeliveriesWeight: 1,
|
|
FirstMessageDeliveriesDecay: 0.999,
|
|
FirstMessageDeliveriesCap: 100,
|
|
InvalidMessageDeliveriesWeight: -1,
|
|
InvalidMessageDeliveriesDecay: 0.9999,
|
|
},
|
|
},
|
|
AppSpecificScore: func(peer.ID) float64 { return 0 },
|
|
DecayInterval: time.Second,
|
|
DecayToZero: 0.01,
|
|
},
|
|
&PeerScoreThresholds{
|
|
GossipThreshold: -1,
|
|
PublishThreshold: -10,
|
|
GraylistThreshold: -1000,
|
|
}))
|
|
|
|
bitmask, err := ps.Join([]byte{0xf1, 0x22, 0xb0, 0x22})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
|
|
err = bitmask.SetScoreParams(
|
|
&BitmaskScoreParams{
|
|
BitmaskWeight: 1,
|
|
TimeInMeshQuantum: time.Second,
|
|
FirstMessageDeliveriesWeight: 1,
|
|
FirstMessageDeliveriesDecay: 0.999,
|
|
FirstMessageDeliveriesCap: 200,
|
|
InvalidMessageDeliveriesWeight: -1,
|
|
InvalidMessageDeliveriesDecay: 0.9999,
|
|
})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
}
|
|
|
|
type mockPeerScoreInspector struct {
|
|
mx sync.Mutex
|
|
scores map[peer.ID]float64
|
|
}
|
|
|
|
func (ps *mockPeerScoreInspector) inspect(scores map[peer.ID]float64) {
|
|
ps.mx.Lock()
|
|
defer ps.mx.Unlock()
|
|
ps.scores = scores
|
|
}
|
|
|
|
func (ps *mockPeerScoreInspector) score(p peer.ID) float64 {
|
|
ps.mx.Lock()
|
|
defer ps.mx.Unlock()
|
|
return ps.scores[p]
|
|
}
|
|
|
|
func TestBlossomSubRPCFragmentation(t *testing.T) {
|
|
ctx, cancel := context.WithCancel(context.Background())
|
|
defer cancel()
|
|
|
|
hosts := getNetHosts(t, ctx, 2)
|
|
ps := getBlossomSub(ctx, hosts[0])
|
|
|
|
// make a fake peer that requests everything through IWANT gossip
|
|
iwe := iwantEverything{h: hosts[1]}
|
|
iwe.h.SetStreamHandler(BlossomSubID_v12, iwe.handleStream)
|
|
|
|
connect(t, hosts[0], hosts[1])
|
|
|
|
// have the real pubsub join the test bitmask
|
|
_, err := ps.Subscribe([]byte{0xf1, 0x22, 0xb0, 0x22})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
|
|
// wait for the real pubsub to connect and try to graft to the faker
|
|
time.Sleep(time.Second)
|
|
|
|
// publish a bunch of fairly large messages from the real host
|
|
nMessages := 1000
|
|
msgSize := 20000
|
|
for i := 0; i < nMessages; i++ {
|
|
msg := make([]byte, msgSize)
|
|
rand.Read(msg)
|
|
ps.Publish([]byte{0xf1, 0x22, 0xb0, 0x22}, msg)
|
|
time.Sleep(20 * time.Millisecond)
|
|
}
|
|
|
|
// wait a bit for them to be received via gossip by the fake peer
|
|
time.Sleep(5 * time.Second)
|
|
iwe.lk.Lock()
|
|
defer iwe.lk.Unlock()
|
|
|
|
// we should have received all the messages
|
|
if iwe.msgsReceived != nMessages {
|
|
t.Fatalf("expected fake BlossomSub peer to receive all messages, got %d / %d", iwe.msgsReceived, nMessages)
|
|
}
|
|
|
|
// and we should have seen an IHAVE message for each of them
|
|
if iwe.ihavesReceived != nMessages {
|
|
t.Fatalf("expected to get IHAVEs for every message, got %d / %d", iwe.ihavesReceived, nMessages)
|
|
}
|
|
|
|
// If everything were fragmented with maximum efficiency, we would expect to get
|
|
// (nMessages * msgSize) / ps.maxMessageSize total RPCs containing the messages we sent IWANTs for.
|
|
// The actual number will probably be larger, since there's some overhead for the RPC itself, and
|
|
// we probably aren't packing each RPC to it's maximum size
|
|
minExpectedRPCS := (nMessages * msgSize) / ps.maxMessageSize
|
|
if iwe.rpcsWithMessages < minExpectedRPCS {
|
|
t.Fatalf("expected to receive at least %d RPCs containing messages, got %d", minExpectedRPCS, iwe.rpcsWithMessages)
|
|
}
|
|
}
|
|
|
|
// iwantEverything is a simple BlossomSub client that never grafts onto a mesh,
|
|
// instead requesting everything through IWANT gossip messages. It is used to
|
|
// test that large responses to IWANT requests are fragmented into multiple RPCs.
|
|
type iwantEverything struct {
|
|
h host.Host
|
|
lk sync.Mutex
|
|
rpcsWithMessages int
|
|
msgsReceived int
|
|
ihavesReceived int
|
|
}
|
|
|
|
func (iwe *iwantEverything) handleStream(s network.Stream) {
|
|
defer s.Close()
|
|
|
|
os, err := iwe.h.NewStream(context.Background(), s.Conn().RemotePeer(), BlossomSubID_v12)
|
|
if err != nil {
|
|
panic(err)
|
|
}
|
|
|
|
msgIdsReceived := make(map[string]struct{})
|
|
gossipMsgIdsReceived := make(map[string]struct{})
|
|
|
|
// send a subscription for test in the output stream to become candidate for gossip
|
|
r := protoio.NewDelimitedReader(s, 1<<20)
|
|
w := protoio.NewDelimitedWriter(os)
|
|
truth := true
|
|
bitmask := []byte{0xf1, 0x22, 0xb0, 0x22}
|
|
err = w.WriteMsg(&pb.RPC{Subscriptions: []*pb.RPC_SubOpts{{Subscribe: truth, Bitmask: bitmask}}})
|
|
if err != nil {
|
|
panic(err)
|
|
}
|
|
|
|
var rpc pb.RPC
|
|
for {
|
|
rpc.Reset()
|
|
err = r.ReadMsg(&rpc)
|
|
if err != nil {
|
|
if err != io.EOF {
|
|
s.Reset()
|
|
}
|
|
return
|
|
}
|
|
|
|
iwe.lk.Lock()
|
|
if len(rpc.Publish) != 0 {
|
|
iwe.rpcsWithMessages++
|
|
}
|
|
// keep track of unique message ids received
|
|
for _, msg := range rpc.Publish {
|
|
id := string(msg.Seqno)
|
|
if _, seen := msgIdsReceived[id]; !seen {
|
|
iwe.msgsReceived++
|
|
}
|
|
msgIdsReceived[id] = struct{}{}
|
|
}
|
|
|
|
if rpc.Control != nil {
|
|
// send a PRUNE for all grafts, so we don't get direct message deliveries
|
|
var prunes []*pb.ControlPrune
|
|
for _, graft := range rpc.Control.Graft {
|
|
prunes = append(prunes, &pb.ControlPrune{Bitmask: graft.Bitmask})
|
|
}
|
|
|
|
var iwants []*pb.ControlIWant
|
|
for _, ihave := range rpc.Control.Ihave {
|
|
iwants = append(iwants, &pb.ControlIWant{MessageIDs: ihave.MessageIDs})
|
|
for _, msgId := range ihave.MessageIDs {
|
|
if _, seen := gossipMsgIdsReceived[msgId]; !seen {
|
|
iwe.ihavesReceived++
|
|
}
|
|
gossipMsgIdsReceived[msgId] = struct{}{}
|
|
}
|
|
}
|
|
|
|
out := rpcWithControl(nil, nil, iwants, nil, prunes)
|
|
err = w.WriteMsg(out)
|
|
if err != nil {
|
|
panic(err)
|
|
}
|
|
}
|
|
iwe.lk.Unlock()
|
|
}
|
|
}
|
|
|
|
func TestFragmentRPCFunction(t *testing.T) {
|
|
p := peer.ID("some-peer")
|
|
bitmask := []byte{0xf1, 0x22, 0xb0, 0x22}
|
|
rpc := &RPC{from: p}
|
|
limit := 1024
|
|
|
|
mkMsg := func(size int) *pb.Message {
|
|
msg := &pb.Message{}
|
|
msg.Data = make([]byte, size-4) // subtract the protobuf overhead, so msg.Size() returns requested size
|
|
rand.Read(msg.Data)
|
|
return msg
|
|
}
|
|
|
|
ensureBelowLimit := func(rpcs []*RPC) {
|
|
for _, r := range rpcs {
|
|
if r.Size() > limit {
|
|
t.Fatalf("expected fragmented RPC to be below %d bytes, was %d", limit, r.Size())
|
|
}
|
|
}
|
|
}
|
|
|
|
// it should not fragment if everything fits in one RPC
|
|
rpc.Publish = []*pb.Message{}
|
|
rpc.Publish = []*pb.Message{mkMsg(10), mkMsg(10)}
|
|
results, err := fragmentRPC(rpc, limit)
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
if len(results) != 1 {
|
|
t.Fatalf("expected single RPC if input is < limit, got %d", len(results))
|
|
}
|
|
|
|
// if there's a message larger than the limit, we should fail
|
|
rpc.Publish = []*pb.Message{mkMsg(10), mkMsg(limit * 2)}
|
|
results, err = fragmentRPC(rpc, limit)
|
|
if err == nil {
|
|
t.Fatalf("expected an error if a message exceeds limit, got %d RPCs instead", len(results))
|
|
}
|
|
|
|
// if the individual messages are below the limit, but the RPC as a whole is larger, we should fragment
|
|
nMessages := 100
|
|
msgSize := 200
|
|
truth := true
|
|
rpc.Subscriptions = []*pb.RPC_SubOpts{
|
|
{
|
|
Subscribe: truth,
|
|
Bitmask: bitmask,
|
|
},
|
|
}
|
|
rpc.Publish = make([]*pb.Message, nMessages)
|
|
for i := 0; i < nMessages; i++ {
|
|
rpc.Publish[i] = mkMsg(msgSize)
|
|
}
|
|
results, err = fragmentRPC(rpc, limit)
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
ensureBelowLimit(results)
|
|
msgsPerRPC := limit / msgSize
|
|
expectedRPCs := nMessages / msgsPerRPC
|
|
if len(results) != expectedRPCs {
|
|
t.Fatalf("expected %d RPC messages in output, got %d", expectedRPCs, len(results))
|
|
}
|
|
var nMessagesFragmented int
|
|
var nSubscriptions int
|
|
for _, r := range results {
|
|
nMessagesFragmented += len(r.Publish)
|
|
nSubscriptions += len(r.Subscriptions)
|
|
}
|
|
if nMessagesFragmented != nMessages {
|
|
t.Fatalf("expected fragemented RPCs to contain same number of messages as input, got %d / %d", nMessagesFragmented, nMessages)
|
|
}
|
|
if nSubscriptions != 1 {
|
|
t.Fatal("expected subscription to be present in one of the fragmented messages, but not found")
|
|
}
|
|
|
|
// if we're fragmenting, and the input RPC has control messages,
|
|
// the control messages should be in a separate RPC at the end
|
|
// reuse RPC from prev test, but add a control message
|
|
rpc.Control = &pb.ControlMessage{
|
|
Graft: []*pb.ControlGraft{{Bitmask: bitmask}},
|
|
Prune: []*pb.ControlPrune{{Bitmask: bitmask}},
|
|
Ihave: []*pb.ControlIHave{{MessageIDs: []string{"foo"}}},
|
|
Iwant: []*pb.ControlIWant{{MessageIDs: []string{"bar"}}},
|
|
}
|
|
results, err = fragmentRPC(rpc, limit)
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
ensureBelowLimit(results)
|
|
// we expect one more RPC than last time, with the final one containing the control messages
|
|
expectedCtrl := 1
|
|
expectedRPCs = (nMessages / msgsPerRPC) + expectedCtrl
|
|
if len(results) != expectedRPCs {
|
|
t.Fatalf("expected %d RPC messages in output, got %d", expectedRPCs, len(results))
|
|
}
|
|
ctl := results[len(results)-1].Control
|
|
if ctl == nil {
|
|
t.Fatal("expected final fragmented RPC to contain control messages, but .Control was nil")
|
|
}
|
|
// since it was not altered, the original control message should be identical to the output control message
|
|
originalBytes, err := rpc.Control.Marshal()
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
receivedBytes, err := ctl.Marshal()
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
if !bytes.Equal(originalBytes, receivedBytes) {
|
|
t.Fatal("expected control message to be unaltered if it fits within one RPC message")
|
|
}
|
|
|
|
// if the control message is too large to fit into a single RPC, it should be split into multiple RPCs
|
|
nBitmasks := 5 // pretend we're subscribed to multiple bitmasks and sending IHAVE / IWANTs for each
|
|
messageIdSize := 32
|
|
msgsPerBitmask := 100 // enough that a single IHAVE or IWANT will exceed the limit
|
|
rpc.Control.Ihave = make([]*pb.ControlIHave, nBitmasks)
|
|
rpc.Control.Iwant = make([]*pb.ControlIWant, nBitmasks)
|
|
for i := 0; i < nBitmasks; i++ {
|
|
messageIds := make([]string, msgsPerBitmask)
|
|
for m := 0; m < msgsPerBitmask; m++ {
|
|
mid := make([]byte, messageIdSize)
|
|
rand.Read(mid)
|
|
messageIds[m] = string(mid)
|
|
}
|
|
rpc.Control.Ihave[i] = &pb.ControlIHave{MessageIDs: messageIds}
|
|
rpc.Control.Iwant[i] = &pb.ControlIWant{MessageIDs: messageIds}
|
|
}
|
|
results, err = fragmentRPC(rpc, limit)
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
ensureBelowLimit(results)
|
|
minExpectedCtl := rpc.Control.Size() / limit
|
|
minExpectedRPCs := (nMessages / msgsPerRPC) + minExpectedCtl
|
|
if len(results) < minExpectedRPCs {
|
|
t.Fatalf("expected at least %d total RPCs (at least %d with control messages), got %d total", expectedRPCs, expectedCtrl, len(results))
|
|
}
|
|
|
|
// Test the pathological case where a single gossip message ID exceeds the limit.
|
|
// It should not be present in the fragmented messages, but smaller IDs should be
|
|
rpc.Reset()
|
|
giantIdBytes := make([]byte, limit*2)
|
|
rand.Read(giantIdBytes)
|
|
rpc.Control = &pb.ControlMessage{
|
|
Iwant: []*pb.ControlIWant{
|
|
{MessageIDs: []string{"hello", string(giantIdBytes)}},
|
|
},
|
|
}
|
|
results, err = fragmentRPC(rpc, limit)
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
if len(results) != 1 {
|
|
t.Fatalf("expected 1 RPC, got %d", len(results))
|
|
}
|
|
if len(results[0].Control.Iwant) != 1 {
|
|
t.Fatalf("expected 1 IWANT, got %d", len(results[0].Control.Iwant))
|
|
}
|
|
if results[0].Control.Iwant[0].MessageIDs[0] != "hello" {
|
|
t.Fatalf("expected small message ID to be included unaltered, got %s instead",
|
|
results[0].Control.Iwant[0].MessageIDs[0])
|
|
}
|
|
}
|