Struct EventCountCheckpoint 

pub struct EventCountCheckpoint<E: Context> { /* private fields */ }

A durable “last known committed event count” for one conversation partition, stored independently of that partition’s own journal files.

See the module docs for why this exists and the guarantee it provides.

Implementations

impl<E: Context> EventCountCheckpoint<E>

pub async fn open(context: E, partition: &str) -> Result<Self, EventLogError>

Open (or create) the checkpoint store tracking partition.

Errors

Returns EventLogError::Checkpoint if the underlying metadata store fails to initialize.

pub fn expected_count(&self) -> u64

The last durably recorded event count, or 0 if none was ever recorded — a fresh partition, or one whose checkpoint was cleared by Self::clear.

pub async fn record(&mut self, count: u64) -> Result<(), EventLogError>

Durably record count as the new expected minimum.

Callers must only pass a count already observed durably committed (see the module docs) — this store does not itself validate that.

Errors

Returns EventLogError::Checkpoint if the durable sync fails.

pub async fn clear(&mut self) -> Result<(), EventLogError>

Clear the checkpoint because the partition it tracks was destroyed or erased, so a subsequent Self::open of a reused partition name starts fresh at 0 rather than remembering content the partition no longer has.

Errors

Returns EventLogError::Checkpoint if the durable sync fails.