Work with TCP server?

[sce9sc]

Very impressive I tried the example also using it with plain TCP server and it does not work . Will it work only with HyperSwarm??

[sce9sc]

I also tried piping 2 replicating streams . like so

await db2.addPeer(diffFeed.key.toString('hex'))
await db1.addPeer(diffFeed2.key.toString('hex')
let stream1 = await db1.replicate(false, {stream:{},info:{client:false},live: true})
let stream = await db2.replicate(true, {stream:{},info:{client:true},live: true})
stream.pipe(stream1).pipe(stream)

Which actually worked . but only for a few seconds until I got

Error: ETIMEDOUT

I tracked it down to ProtocolStream where the handlers.timeout was undefined

module.exports = class ProtocolStream extends Duplex {
  constructor (initiator, handlers = {}) {
    super()

    if (typeof initiator !== 'boolean') throw new Error('Must specify initiator boolean in replication stream')

    this.initiator = initiator
    this.handlers = handlers
    this.channelizer = new Channelizer(this, {
      encrypted: handlers.encrypted,
      noise: handlers.noise,
      keyPair: handlers.keyPair
    })
    this.state = new SHP(initiator, this.channelizer)
    this.live = !!handlers.live
    this.timeout = null
    this.keepAlive = null
    this.prefinalize = new Nanoguard()
    this.bytesSent = 0
    this.bytesReceived = 0
    this.extensions = StreamExtension.createLocal(this)
    this.remoteExtensions = this.extensions.remote()

    this._utp = null
    this._changes = 0

    this.once('finish', this.push.bind(this, null))
    this.on('pipe', this._onpipe)

         if (handlers.timeout !== false && handlers.timeout !== 0) {
             const timeout = handlers.timeout || 20000
             this.setTimeout(timeout, () => this.destroy(new Error('ETIMEDOUT')))
             this.setKeepAlive(Math.ceil(timeout / 2))
         }

  }

I hope this helps

[pgmemk]

@sce9sc Thank you for your feedback 😊. It is really great to have someone interested in our project. Right now I am working on another urgent project but I'll definitely look into it in a few days.

[pgmemk]

The reason your example does not work is

• You need to pipe the original diff feed into its clone.

It seemed confusing, that's why we changed the API so that you don't have to think about it.

On machine 1:

    let peer2 = await db1.addPeer(key2)
    let dbStream = await db1.replicate(isInitiator, {stream: streamBetweenPeers, live: true})

On machine 2:

    let peer1 = await db2.addPeer(key1)    
    let dbStream = await db2.replicate(isInitiator, {stream: streamBetweenPeers, live: true})

streamBetweenPeers - is either socket or the clone of the same multi-hyperbee's diff hyperbee on the other device. HyperSwarm is one way to establish a stream between peers