Struct SyncEngine 

pub struct SyncEngine { /* private fields */ }

The primary entry-point for the sync layer.

Call SyncEngine::new once per database, then use record_mutation after every write and apply_remote_ops when you receive a batch from a peer.

Implementations

impl SyncEngine

pub fn new(db: &AgentDB, node_id: &str) -> Result<Self>

Open the sync layer on top of an existing AgentDB connection.

Creates _adb_sync_log and _adb_sync_peers if they don’t exist yet.

pub fn record_mutation( &mut self, table: &str, record_id: &str, op: OpType, payload: Option<Value>, ) -> Result<String>

Record a mutation in the sync log and return its op_id.

pub fn get_ops_since(&self, since_hlc: i64) -> Result<Vec<SyncOp>>

Return all ops with hlc_ts > since_hlc, ordered by timestamp.

pub fn apply_remote_ops(&mut self, ops: Vec<SyncOp>) -> Result<SyncResult>

Apply a batch of ops received from a remote peer.

For each op:

• If no existing op exists for the same (table_name, record_id) → insert it (counts as applied).
• If a conflict exists → apply the configured ConflictStrategy: the winning op is inserted/updated; conflicts is always incremented; the losing op increments skipped.

pub fn add_peer(&self, peer_id: &str, endpoint: Option<&str>) -> Result<()>

Register a peer node, optionally with a network endpoint.

pub fn sync_status(&self) -> Result<Vec<PeerStatus>>

Return the sync status of every registered peer.

pub fn set_strategy(&mut self, strategy: ConflictStrategy)

Change the conflict resolution strategy (default: LastWriterWins).