Struct SortedRuns 

pub struct SortedRuns { /* private fields */ }

Implementations

impl SortedRuns

pub const MAX_FANOUT: usize = 64

The most runs iter will hold open at once.

Without a cap, iter opens EVERY run with a read buffer each, so merge memory is run_count * buffer – and run_count is input / arena, which makes it linear in the input. Law 1 forbids that: it happens to fit at moderate scale and stops fitting above it, which is precisely the shape this project exists to avoid – correct at the test scale, wrong at the target one, invisible until someone’s data grows.

pub fn run_count(&self) -> usize

pub fn discard(self) -> Result<()>

Remove a durable sort checkpoint after its graft and final metadata publication are known durable. Ordinary temporary runs already clean themselves on Drop; accepting both shapes keeps caller cleanup simple.

pub fn iter(&mut self) -> Result<MergeIter>

K-way merge over at most MAX_FANOUT runs. RAM is one buffered reader per open run, a chosen bound rather than an input-shaped one.

Trait Implementations

impl Drop for SortedRuns

fn drop(&mut self)

Executes the destructor for this type.

fn pin_drop(self: Pin<&mut Self>)

🔬This is a nightly-only experimental API. (pin_ergonomics)

Execute the destructor for this type, but different to Drop::drop, it requires self to be pinned.