For AVS node operator, there are two commands for registration steps,
create-bls-signature and register-bls.
Unfortunately, usages of those two have quite different.
For example, register-bls command requires .env file to set RPC and Private Key but create-bls-signature does not.
This types of issue could be thought to be ignorable but absolutely make huge confusions to users.
Providing unified usage always better and for that we could two simple choices.
Option 1: provide config file like config.yaml
Exactly, eigenlayer cli uses this method.
All required variables should defined in the config file and the cli application loads those variable from file and using it.
It could be easy to execute command because there are less command line arguments and we can provide a config file templates, which can reduce confusions.
But we should change all "flag handle" parts for using config file,
and also the update can change experience for previous users(only 4 teams 😆).
Option 2: Use command line arguments(flags)
Retrieve all required variables only form CLI flag. This is totally opposite way of option 1.
Option 2 is very similar to current implementation but .env relating parts should be removed.
For AVS node operator, there are two commands for registration steps,
create-bls-signature
andregister-bls
.Unfortunately, usages of those two have quite different. For example,
register-bls
command requires.env
file to setRPC
andPrivate Key
butcreate-bls-signature
does not.This types of issue could be thought to be ignorable but absolutely make huge confusions to users. Providing unified usage always better and for that we could two simple choices.
Option 1: provide config file like
config.yaml
Exactly,eigenlayer
cli uses this method. All required variables should defined in the config file and the cli application loads those variable from file and using it.It could be easy to execute command because there are less command line arguments and we can provide a config file templates, which can reduce confusions.
But we should change all "flag handle" parts for using config file, and also the update can change experience for previous users(only 4 teams 😆).
Option 2: Use command line arguments(flags) Retrieve all required variables only form CLI flag. This is totally opposite way of option 1.
Option 2 is very similar to current implementation but
.env
relating parts should be removed.