Struct SchedulerScope 

pub struct SchedulerScope<'scope, C, const BLOCKING_QUEUE_CAPACITY: usize = 256, const SPIN_LIMIT: usize = 8192>
where
    C: WorkClass,
{ /* private fields */ }

Borrowing scope for scheduler jobs that must complete before the scope exits.

Implementations

impl<'scope, C, const BLOCKING_QUEUE_CAPACITY: usize, const SPIN_LIMIT: usize> SchedulerScope<'scope, C, BLOCKING_QUEUE_CAPACITY, SPIN_LIMIT>

where C: WorkClass,

pub fn spawn<F>(&self, task: F) -> Result<(), ExecutorError>

where F: FnOnce(usize) + Send + 'scope,

Spawn a job into this scope.

The job may borrow values that outlive the scope call. Scoped jobs are coalesced into worker-sized scheduler batches and complete before ThreadScheduler::scope returns. Jobs are not guaranteed to start while the scope body is still registering work.

The usize the job receives identifies the lane running it. Worker lanes are 0..worker_count(); a job the admission queue turned away runs on the calling lane, identified as worker_count(). It is a lane identity, not an index into the worker set.

pub fn flush(&self) -> Result<(), ExecutorError>

Schedule all jobs currently buffered in this scope.

ThreadScheduler::scope calls this before waiting, so most callers do not need to invoke it directly. It is exposed for two-lane fork/join shapes where one branch should enter the scheduler before the caller executes the second branch locally. The scope still waits for every flushed job before returning, so borrowed data cannot escape.