Struct NeighborSyncState 

pub struct NeighborSyncState {
    pub priority_order: VecDeque<PeerId>,
    pub order: Vec<PeerId>,
    pub cursor: usize,
    pub last_sync_times: HashMap<PeerId, Instant>,
    pub bootstrap_claims: HashMap<PeerId, Instant>,
    pub bootstrap_claim_history: HashMap<PeerId, Instant>,
    pub prune_cursor: usize,
}

Neighbor sync cycle state.

Tracks a deterministic walk through the current close-group snapshot, per-peer cooldown times, active bootstrap claims, and peers that have already used their one bootstrap-claim window.

Fields

priority_order: VecDeque<PeerId>

Newly-entered close peers to sync before the normal cycle cursor.

This queue is populated from KClosestPeersChanged routing-table events and is intentionally separate from order: priority syncs do not restart the current round-robin cycle or discard already-scanned peers.

order: Vec<PeerId>

Deterministic ordering of peers for the current cycle (snapshot).

cursor: usize

Current cursor position into order.

last_sync_times: HashMap<PeerId, Instant>

Per-peer last successful sync time (for cooldown).

bootstrap_claims: HashMap<PeerId, Instant>

Active bootstrap claim first-seen timestamps per peer.

Entries are removed when a peer stops claiming bootstrap. The peer remains in bootstrap_claim_history, so a later claim is repeated-claim abuse instead of a fresh grace period.

bootstrap_claim_history: HashMap<PeerId, Instant>

First-ever bootstrap claim timestamp per peer.

This is retained after active claims are cleared so each peer gets at most one bootstrap-claim grace window. Under Sybil attack with many distinct peer IDs claiming bootstrap, this map grows unboundedly. In practice the trust engine limits Sybil impact before this becomes a memory issue.

prune_cursor: usize

Cursor used by post-cycle pruning to rotate through stored records when the per-pass prune-confirmation budget is exhausted.

Implementations

impl NeighborSyncState

pub fn new_cycle(close_neighbors: Vec<PeerId>) -> Self

Create a new cycle from the given close neighbors.

pub fn observe_bootstrap_claim( &mut self, peer: PeerId, now: Instant, grace_period: Duration, ) -> BootstrapClaimObservation

Observe a peer claiming bootstrap status.

A peer receives one grace window from its first observed bootstrap claim. If it later stops claiming bootstrap, callers should clear only the active claim with Self::clear_active_bootstrap_claim. A subsequent claim is then reported as BootstrapClaimObservation::Repeated.

pub fn clear_active_bootstrap_claim(&mut self, peer: &PeerId) -> bool

Clear the active bootstrap claim for a peer, retaining claim history.

pub fn queue_priority_peers<I>(&mut self, peers: I) -> usize

where I: IntoIterator<Item = PeerId>,

Queue newly-entered close peers for priority sync.

Existing queued peers are retained in their original position and duplicates from the same routing-table churn burst are ignored.

pub fn retain_sync_peers(&mut self, close_peers: &HashSet<PeerId>) -> usize

Drop pending sync peers that are no longer in the close-peer set.

Peers still in close_peers keep their relative position. The cursor is adjusted so already-scanned retained peers are not selected again.

pub fn remove_peer(&mut self, peer: &PeerId) -> bool

Remove a peer from any pending neighbor-sync state.

pub fn is_cycle_complete(&self) -> bool

Whether the current cycle is complete.

Trait Implementations

impl Debug for NeighborSyncState

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.