Struct ReadonlyTimeoutEstimator 

pub(crate) struct ReadonlyTimeoutEstimator {
    using_estimates: bool,
    latest_timeout: Option<Duration>,
    default_timeout: Duration,
}

A timeout estimator based on reading timeouts that another timeout estimator is computing, in another process.

Fields

using_estimates: bool

Are we using the timeouts?

latest_timeout: Option<Duration>

Latest estimate from the persistent state.

default_timeout: Duration

Timeout to use if we don’t have a computed timeout.

Implementations

impl ReadonlyTimeoutEstimator

pub(crate) fn new() -> Self

Create a new ReadonlyTimeoutEstimator with default settings.

pub(crate) fn from_state(s: &ParetoTimeoutState) -> Self

Create a new ReadonlyTimeoutEstimator, based on persistent state

pub(crate) fn update_from_state(&mut self, s: &ParetoTimeoutState)

Update this estimator based on a newly read state.

Trait Implementations

impl TimeoutEstimator for ReadonlyTimeoutEstimator

fn note_hop_completed(&mut self, _hop: u8, _delay: Duration, _is_last: bool)

Record that a given circuit hop has completed.

fn note_circ_timeout(&mut self, _hop: u8, _delay: Duration)

Record that a circuit failed to complete because it took too long.

fn timeouts(&mut self, action: &Action) -> (Duration, Duration)

Return the current estimation for how long we should wait for a given Action to complete.

fn learning_timeouts(&self) -> bool

Return true if we’re currently trying to learn more timeouts by launching testing circuits.

fn update_params(&mut self, params: &NetParameters)

Replace the network parameters used by this estimator (if any) with ones derived from params.

fn build_state(&mut self) -> Option<ParetoTimeoutState>

Construct a new ParetoTimeoutState to represent the current state of this estimator, if it is possible to store the state to disk.