Struct SnapshotChunkDecoder 

pub struct SnapshotChunkDecoder { /* private fields */ }

Incrementally reconstruct frames and check the snapshot’s end marker. Call Self::next_record repeatedly on each opened chunk, updating the oracle after each applied-write record before asking for the next record.

Implementations

impl SnapshotChunkDecoder

pub fn new(prefix: SnapshotObjectPrefix) -> Self

Supply the authenticated prefix, including the snapshot’s coverage.

pub fn header(&self) -> Option<&SnapshotHeader>

Header metadata becomes available as soon as its frame is complete.

pub fn next_record<O: WriteSlots>( &mut self, bytes: &mut &[u8], oracle: &O, ) -> Result<Option<SnapshotRecord>, O::Error>

where O::Error: From<Error>,

Consume bytes through the next complete record, or all available bytes if its frame is unfinished. None means this input has no further record. Discard the decoder after an error; commit staged rows only after finish.

pub fn next_history( &mut self, bytes: &mut &[u8], ) -> Result<Option<SnapshotRecord>, Error>

Like Self::next_record, before the store’s historical signing keys are known: only original slots and last claims are decoded, and every other record is counted and passed over (SnapshotDecoder::history_frame). A loader still decodes every record with Self::next_record.

pub fn frame(&self) -> &[u8]

The exact bytes of the frame the last returned record came from, for a loader that stages the record and reads it back with SnapshotRecord::decode_checked.

pub fn finish(&self) -> Result<(), Error>

Refuse EOF without the complete header, declared records and end marker.