Struct Dispatcher 

pub struct Dispatcher { /* private fields */ }

Fans record ops across N replica stores with round-robin + failover.

Implementations

impl Dispatcher

pub fn new(replicas: Vec<Arc<dyn Store>>) -> Self

Build a dispatcher over one or more replica stores.

pub fn len(&self) -> usize

Number of replica handles (live or not — deadness is discovered per op).

pub fn is_empty(&self) -> bool

Always false; provided to satisfy clippy’s len-without-is_empty lint.

pub fn replicas(&self) -> &[Arc<dyn Store>]

Every replica handle, in construction order. Used to read each replica directly (no round-robin, no failover) once the run has settled.

pub async fn get(&self, key: &RecordKey) -> Result<Option<Record>>

get with round-robin start + failover across all replicas.

pub async fn put( &self, record: Record, expected: Option<Revision>, ) -> Result<PutResult>

put with round-robin start + failover across all replicas. The record is cloned per attempt so a failover can re-issue it to another replica. A Conflict is a valid Ok outcome and returns immediately (not a failover).

pub async fn delete( &self, key: &RecordKey, expected: Option<Revision>, ) -> Result<DeleteResult>

delete with round-robin start + failover across all replicas. A Conflict is a valid Ok outcome and returns immediately (not a failover). A failover after a delete that did commit on the dead replica is safe: the retry sees a tombstone and returns Deleted without writing.

pub async fn get_raw(&self, key: &RecordKey) -> Result<Option<Record>>

get_raw (tombstones visible) with round-robin start + failover.