c59b91bce5
The node binary is most often run as a transport-only service/fleet node, so `logosdeliverynode` no longer mounts the messaging client and reliable channel manager unless asked. Protocol flags are unchanged in the default case: the individual CLI defaults already match what `applyMode(Core)` was setting. Only the CLI default moves. `LogosDelivery.new(entryLayer = ...)` and `parseLogosDeliveryConf` keep defaulting to `channels`, so the Nim and C library entry points are unaffected. `MessagingClientConf.toWakuNodeConf` pins `entryLayer = channels` for the same reason it already pins `mode`: a conf derived from a messaging config is by construction not kernel-only. `LogosDeliveryConf.init` now honours `entryLayer == kernel` instead of always mounting the messaging layer, so the layer it records and the layers it mounts agree. Doc comments follow the code: `mode`/`preset` shape the kernel conf for every layer, `kernel` included, and `init(kernelConf)` stays the raw entry. Co-authored-by: NagyZoltanPeter <113987313+NagyZoltanPeter@users.noreply.github.com> Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
Examples
Compile
Make all examples.
make example2
Waku API
Uses the simplified Waku API to create and start a node, you need an RPC endpoint for Linea Sepolia for RLN:
./build/waku_api --ethRpcEndpoint=https://linea-sepolia.infura.io/v3/<your key>
If you can't be bothered but still want to see some action, just run the binary and it will use a non-RLN network:
./build/waku_api
## publisher/subscriber
Within examples/ you can find a publisher and a subscriber. The first one publishes messages to the default pubsub topic on a given content topic, and the second one runs forever listening to that pubsub topic and printing the content it receives.
Some notes:
- These examples are meant to work even if you are behind a firewall and you can't be discovered by discv5.
- You only need to provide a reachable bootstrap peer (see our fleets)
- The examples are meant to work out of the box.
- Note that both services wait for some time until a given minimum amount of connections are reached. This is to ensure messages are gossiped.
Run:
Wait until the subscriber is ready.
./build/subscriber
And run a publisher
./build/publisher
See how the subscriber received the messages published by the publisher. Feel free to experiment from different machines in different locations.
resource-restricted publisher/subscriber (lightpush/filter)
To illustrate publishing and receiving messages on a resource-restricted client,
examples/v2 also provides a lightpush_publisher and a filter_subscriber.
The lightpush_publisher continually publishes messages via a lightpush service node
to the default pubsub topic on a given content topic.
The filter_subscriber subscribes via a filter service node
to the same pubsub and content topic.
It runs forever, maintaining this subscription
and printing the content it receives.
Run Start the filter subscriber.
./build/filter_subscriber
And run a lightpush publisher
./build/lightpush_publisher
See how the filter subscriber receives messages published by the lightpush publisher.
Neither the publisher nor the subscriber participates in relay,
but instead make use of service nodes to save resources.
Feel free to experiment from different machines in different locations.