Channel policy enforcement checks

[elnosh]

I was changing the numbers of the json file for a simulated network and found myself confused on how the channel policy checks are enforced in sim_node.rs

after parsing the file, for each node in the channel we have this channel policy struct:

/// Represents one node in the channel's forwarding policy and restrictions. Note that this doesn't directly map to
/// a single concept in the protocol, a few things have been combined for the sake of simplicity. Used to manage the
/// lightning "state machine" and check that HTLCs are added in accordance of the advertised policy.
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct ChannelPolicy {
    pub pubkey: PublicKey,
    #[serde(default)]
    pub alias: String,
    pub max_htlc_count: u64,
    pub max_in_flight_msat: u64,
    pub min_htlc_size_msat: u64,
    pub max_htlc_size_msat: u64,
    pub cltv_expiry_delta: u32,
    pub base_fee: u64,
    pub fee_rate_prop: u64,
}

However the way we enforce the policy does not seem correct. Particularly how we add HTLCs through the route. We do checks on the sending node like such.

    fn check_outgoing_addition(&self, htlc: &Htlc) -> Result<(), ForwardingError> {
        fail_forwarding_inequality!(htlc.amount_msat, >, self.policy.max_htlc_size_msat, MoreThanMaximum);
        fail_forwarding_inequality!(htlc.amount_msat, <, self.policy.min_htlc_size_msat, LessThanMinimum);
        fail_forwarding_inequality!(
            self.in_flight.len() as u64 + 1, >, self.policy.max_htlc_count, ExceedsInFlightCount
        );
        fail_forwarding_inequality!(
            self.in_flight_total() + htlc.amount_msat, >, self.policy.max_in_flight_msat, ExceedsInFlightTotal
        );
        fail_forwarding_inequality!(htlc.amount_msat, >, self.local_balance_msat, InsufficientBalance);
        fail_forwarding_inequality!(htlc.cltv_expiry, >, 500000000, ExpiryInSeconds);

        Ok(())
    }

This translates to the sender checking it's own policy, not the counterparty one i.e the htlc amount should respect the counterparty policy max_htlc_size_msat not self. Unless there is an expectation on how the policies are specified in the simulation json file, this seems incorrect.