- Introduced the concept of a Cancellation error so that services can choose to watch for a cancellation flag.
- Services can then send this error back and it will be forwarded to the node.
- The node will then know to call shutdown appropriately.
- Added the concept of loadServices on the node so that the node can conditionally call stop on loadingServices
- This serves the case where services might be loading versus fully loaded (which is not always the cases for heavy services like bitcoind)
- A Node will shutdown if there is an error starting a service as it would lead to cascading errors.
- `node.start()` needs to be called, and nolonger is called automatically when the instance is created.
- A service will only be added to node.services after it's started
- Stopping services that are not started will gracefully continue.
- Logging sync status of db will only apply if the service is started.
- Debug log about a service without a route will always include the service name
- added bindings to listen for tip updates from the daemon
- update the height during syncing for continuity
- call sync when there is a new tip
- added ability to rewind to a ancestor when there is a fork
