Teo Gebhard
Teo Gebhard
**TODO Why some changes needed for Storage.ts, killing-dead-connections.test.ts, SubscriptionSession.ts. Without the changes we get eslint/build errors. Most likely not related to this PR, and maybe these fail in `main`, too?...
Based on https://github.com/streamr-dev/network/pull/1034 Remove `networkId` command line option from broker bin script. It is most likely not needed, because it has been broken since https://github.com/streamr-dev/network/pull/309. It was originally added in...
## Functional changes TODO: update changelog: that resend subscriptions can be unsubscribed TODO: remove Subscription#unsubscribe test cases from unsubscribe.test.ts (maybe replace with unit test which ensures that Subscription#unsubscribe is called)...
## Summary Added test case which demonstrates how subscription is automatically unsubscribed if user iterates it and breaks out of loop. It is not clear whether this is a good...
Extracted `ConnectionsView` interface from `ITransport`. Now `ITransport` handles just the transport logic. Connection management and queries are in separate classes: in `ConnectionLocker` and in the new `ConnectionsView`. ## TODO Extract...
Moved into `webrtc` directory as the class is WebRTC-specific.
When we sort the oldContacts, we should sort all contacts, not just a subset? In practice the oldContacts array length is already limited by the `numberOfNodesPerKBucket` (see `PeerManager.ts:82`), so there...
Based on https://github.com/streamr-dev/network/pull/2507, merge that first. See comment in a previous PR: https://github.com/streamr-dev/network/pull/2507#issuecomment-2069273895
Renamed `PeerManager#closestContacts` -> `nearbyContacts`. This way we can separate the concepts of: - nearby contacts (all currently known nodes having short Kademlia distance to this node) - closest contacts/neighbors/etc. (a...
Improve `WebsocketServer` cleanup. When we explicitly remove `httpServer` listeners, the instance doesn't leak memory immediately after `WebsocketServer#stop` is called. There wasn't a permanent memory leak: e.g. `await wait(100)` before the...