Add script to setup two gl nodes with one local CLN node

[ShahanaFarooqui]

How to Test:

• Run make gltestserver in first terminal
• Run cd examples/javascript && npm install
• Run cd ../.. && . setup-nodes.sh after installation is finished.

[cdecker]

Very good 👍

I have a bit more scaffolding in mind to make controlling the mock LN network a bit more comfortable from outside the docker image. I have the following graph in mind:

flowchart LR
    classDef planned fill:#fcc
    classDef mock fill:#AAA
    subgraph Host
    direction LR
      cli
      IDE
      client-app
      browser
    end


    subgraph Docker
      S["grpc-interface"]
      subgraph gl-testserver
      direction LR
      S --> cln-nodes:::planned
      
        mock-scheduler:::mock
        bitcoind
        grpc-web-proxy:::mock
        S --> mock-scheduler
      end
    client-app & cli & IDE --> S 
    client-app & cli & IDE--> mock-scheduler
    cli --> bitcoind
    browser --> grpc-web-proxy --> cln1 & cln2 & cln3
    browser --> grpc-web-proxy --> mock-scheduler
    subgraph "Virtual Network"
        direction LR
        cln-nodes -- JSON-RPC --> cln1
        cln-nodes -- JSON-RPC --> cln2
        cln-nodes -- JSON-RPC --> cln3
        cln-nodes -- JSON-RPC --> lsp1:::planned
      end
      subgraph gl-nodes
        direction LR
        gl1
        gl2
        gl3 
      end
      mock-scheduler --> gl1 & gl2 & gl3
    end

Loading

The two grey boxes being the mock services that emulate GL, and the red highlights are the parts that I hope we can build out. It essentially exposes the CLN JSON-RPC over a web address, and multiplexes multiple CLN nodes into one server. This way you'll be able to say things like cln-cli my-lsp-node exec getinfo to connect to the multiplexer, which locates the my-lsp-node, connects to it, executes the RPC command, and ships the result back. This way we have a unified API to setup, control and tear down networks at will, even automatically in CI testing for example.