config example shows unable to interpret strict_checkpoint_age, load_external_fallback, and data_dir from ~/.helios/helios.toml

[ltfschoen]

If i create an example config file and cp helios.example.toml ~/.helios/helios.toml, where its contents are:

[mainnet]
# The consensus rpc to use. This should be a trusted rpc endpoint. Defaults to "https://www.lightclientdata.org".
consensus_rpc = "https://www.lightclientdata.org"
# [REQUIRED] The execution rpc to use. This should be a trusted rpc endpoint.
execution_rpc = "https://eth-mainnet.g.alchemy.com/v2"
# The port to run the JSON-RPC server on. By default, Helios will use port 8545.
rpc_port = 8545
# The latest checkpoint. This should be a trusted checkpoint that is no greater than ~2 weeks old.
# If you are unsure what checkpoint to use, you can skip this option and set either `load_external_fallback` or `fallback` values (described below) to fetch a checkpoint. Though this is not recommended and less secure.
checkpoint = "0x0ed5b225c09f70c1adcdd96f6355f6ca7bf8b8dc7ab7deda9acad97c4cf6224f"
# The directory to store the checkpoint database in. If not provided, Helios will use "~/.helios/data/mainnet", where `mainnet` is the network.
# It is recommended to set this directory to a persistent location mapped to a fast storage device.
data_dir = "/root/.helios/mainnet"
# The maximum age of a checkpoint in seconds. If the checkpoint is older than this, Helios will attempt to fetch a new checkpoint.
max_checkpoint_age = 86400
# A checkpoint fallback is used if no checkpoint is provided or the given checkpoint is too old.
# This is expected to be a trusted checkpoint sync api (like provided in https://github.com/ethpandaops/checkpoint-sync-health-checks/blob/master/_data/endpoints.yaml).
fallback = "https://sync-mainnet.beaconcha.in"
# If no checkpoint is provided, or the checkpoint is too old, Helios will attempt to dynamically fetch a checkpoint from a maintained list of checkpoint sync apis.
# NOTE: This is an insecure feature and not recommended for production use. Checkpoint manipulation is possible.
load_external_fallback = true

[goerli]
# The consensus rpc to use. This should be a trusted rpc endpoint. Defaults to Nimbus testnet.
consensus_rpc = "http://testing.prater.beacon-api.nimbus.team"
# [REQUIRED] The execution rpc to use. This should be a trusted rpc endpoint.
execution_rpc = "https://eth-goerli.g.alchemy.com/v2"
# The port to run the JSON-RPC server on. By default, Helios will use port 8545.
rpc_port = 8545
# The latest checkpoint. This should be a trusted checkpoint that is no greater than ~2 weeks old.
# If you are unsure what checkpoint to use, you can skip this option and set either `load_external_fallback` or `fallback` values (described below) to fetch a checkpoint. Though this is not recommended and less secure.
checkpoint = "0x7beab8f82587b1e9f2079beddebde49c2ed5c0da4ce86ea22de6a6b2dc7aa86b"
# The directory to store the checkpoint database in. If not provided, Helios will use "~/.helios/data/goerli", where `goerli` is the network.
# It is recommended to set this directory to a persistent location mapped to a fast storage device.
data_dir = "/root/.helios/goerli"
# The maximum age of a checkpoint in seconds. If the checkpoint is older than this, Helios will attempt to fetch a new checkpoint.
max_checkpoint_age = 86400
# A checkpoint fallback is used if no checkpoint is provided or the given checkpoint is too old.
# This is expected to be a trusted checkpoint sync api (like provided in https://github.com/ethpandaops/checkpoint-sync-health-checks/blob/master/_data/endpoints.yaml).
fallback = "https://sync-goerli.beaconcha.in"
# If no checkpoint is provided, or the checkpoint is too old, Helios will attempt to dynamically fetch a checkpoint from a maintained list of checkpoint sync apis.
# NOTE: This is an insecure feature and not recommended for production use. Checkpoint manipulation is possible.
load_external_fallback = true

and then run cargo run -p helios --example config using "mainnet" in examples/config.rs

the following arguments values aren't being read from the file strict_checkpoint_age and load_external_fallback and data_dir

same issue if run cargo run -p helios --example config after switching to "goerli" in examples/config.rs and run

why isn't data_dir set in the base configuration config/src/config.rs?

[ncitron]

Thanks for pointing this out! This is a bug.

I'm actually planning on doing a refactor soon with a bunch of breaking changes which will include simplifying the configuration significantly. Part of the reason there are bugs here is that the configuration code is just too complicated.