twoeths

Unstable IWANT messages we don't have

1

Investigate why the "IWANT received we don't have" metrics are not as stable as version 0.14.0 - Version 0.14.0 - Version 4.0.0

Add protons back

4

**Motivation** In #310 we have to switch back to protobufjs because protons is 20x slower than protobufjs, we should add it back once the performance of protons is comparable or...

feat: switch back to protons

1

**Motivation** - We should switch back to protons when it's performance is comparable or better than protobufjs - Prepare a branch to track the performance of different versions of protons...

New parameter: pruneBackoff and unsubcribeBackoff

2

**Motivation** + Right now we have backoff time as a constant of 1 minute ```typescript export const GossipsubPruneBackoff = minute ``` + Lodestar wants to specifies its own `backoff` time...

Deduplicate iwant list

1

**Motivation** Right now if n peers sending the same message that we don't have, we'll ask for the same message to these n peers - If we receive N `IHAVE`...

Consider synced flow when validating gossip messages

2

**Motivation** Right now lodestar has "StrictNoSign" signature policy, and a synced `msgIdFn`, running synced version of "handleReceivedMessage" and "validateReceivedMessage" (hard-coded in `node_modules`) gives us some improvement in a contabo machine...

Fix failed test: test mixed gossipsub

1

This test sometimes failed, even in local environment: ``` 1) go-libp2p-pubsub gossipsub tests test mixed gossipsub: Error: Message never received, sender 6, receiver 19, index 67 at Timeout._onTimeout (file:///Users/tuyennguyen/Projects/workshop/gossipsub-js/dist/test/e2e/go-gossipsub.spec.js:28:16) at...

Performance of gauge.inc()

5

As found in the below profile, the total `gauge.inc()` calls takes > 2% of cpu time + maybe switch to counter.inc() instead? + write performance test to make sure [0522_contabo_20_subscribe_all_subnets.cpuprofile.zip](https://github.com/ChainSafe/js-libp2p-gossipsub/files/8877932/0522_contabo_20_subscribe_all_subnets.cpuprofile.zip)

perf: improve metric performance

2

**Motivation** - `gauge.inc()` takes 2.45% cpu time which is > 2 / 3 `runHeartbeat()` cpu time **Description** - `gauge.inc()` will hash the label objects and push to Map every time...

Start network before sync

1

**Motivation** Sometimes the sync is too fast and it tries to subscribe to core topics while network is not ready **Description** Start network before sync Closes #4543