Calls to cln.Node/Invoice now always include all possible
routehints. Possible in this case refers to channels with peers
that are currently in state CHANNELD_NORMAL, both disconnected
and connected.
June
The JS bindings where clobbering the error messages due to
incorrect context use. Now we return errors as they are emitted.
The scheduler no longer allows creating regtest nodes, since they
are unusable without a faucet to get coins for it.
July
The gl-plugin will now wait for both the initial gossip sync and
the reconnection to the peers to complete, before allowing pay
through. This cuts down on spurious payment failures due to missing
peers or incomplete network view for routing.
An issue concerning reconnecting to peers, if the signer attaches
before the underlying JSON-RPC has become available has been
fixed. This issue would cause peers to remain disconnected despite
a signer being attached. #210 & #204
The gl-client library and the language bindings have keepalive
messages enabled, with a timeout of 90 seconds. This ensures that
clients and signers that have been silently disconnected, e.g., by
suspending the device or losing network connectivity, will notice
and reconnect. #220
The Node Domain has been enabled. This means that ever node now has
a unique URL at which the node can always be reached, without
having to explicitly schedule it first. This allows bypassing of
the scheduler, reducing the time required to start and connect to a
node.
September
Temporarily removed the JS bindings. We will the bindings over to
uniffi, and the JS bindings were outdated and unused. But they'll
be back.
The API has been simplified by removing methods that were both in
greenlight.proto as well as node.proto. The latter is from the
autogenerated cln-grpc which supercedes the greenlight.proto
methods.
November
The methods that used to be in greenlight.proto that have been
superseded with the node.proto version have been removed on the
server side. The proto file now contains only Greenlight-specific
functionality (#317.
Payment optimizations: we are working on getting the success rate
for payments up, and the time to completion down, focusing on
success rate first.