Struct SpilledBody 

pub struct SpilledBody<D: Columnar, T> {
    records: usize,
    fences: D::Container,
    time_lower: Antichain<T>,
    time_upper: SmallVec<[T; 1]>,
    compressed: bool,
    handle: ChunkHandle,
}

A spilled chunk body: the serialized column in the pool, plus the resident metadata every Chunk must answer without fetching. That metadata is the record count, the first and last data items (the fence entries UnloadChunk::locate consults), and the time bounds extract consults to pass frontier-disjoint chunks through without loading them.

Fields

records: usize

Number of updates in the body.

fences: D::Container

The first and last data items, as a two-element container. One container rather than two singletons, so the leaf allocations are not duplicated per fence.

time_lower: Antichain<T>

The minimal times in the body: a lower bound antichain every contained time is greater-or-equal to. Folded into extract’s residual frontier when the chunk is kept whole.

time_upper: SmallVec<[T; 1]>

The maximal times in the body. Some contained time is greater-or-equal to a frontier exactly when some maximal time is, which is extract’s ship-whole test. A single element for totally ordered times, hence the inline capacity.

compressed: bool

Whether the body was inserted under the compressing codec. The pool stores the codec itself and reads decode through it, so this is the only handle chunk code has on what a body is stored as, and it is what survive_merge consults to decide a body wants migrating. Deriving that from depth instead would tie it to a single transition and miss every path that skips it.

handle: ChunkHandle

The pool chunk holding the serialized column.