TaskPool

1 min read

Import: import system.task;

Read the usage guide for examples and common tasks. This reference lists the public declarations for this part of the library. See Reading API signatures for parameter and result notation.

TaskPool

Reuses a fixed number of parallel workers; local async tasks do not consume pool workers. Each worker holds at most two jobs. Further dispatches wait in submission order and go to the least-loaded worker once one finishes, so a long job holds up at most one job behind it.

public class TaskPool

constructor

Creates a fixed worker set; nonpositive sizes produce InvalidSize on operations.

public constructor(size: int)

size

Number of reusable parallel workers in this pool.

public get size(): int

dispatch

Dispatches a CPU job; capture and return values follow the shared/move worker rules.

public fun dispatch<TOut: shared>(
    body: fun(): TOut,
    token: Option<CancellationToken> = Option.None
): Future<Result<TOut, TaskPoolError>>

dispatch_async

Dispatches an async job on a pool worker and waits for its operation to settle.

public fun dispatch_async<TOut: shared>(
    body: fun(): Future<TOut>,
    token: Option<CancellationToken> = Option.None
): Future<Result<TOut, TaskPoolError>>

map

Maps CPU work with at most size active dispatches and preserves input order.

public fun map<T: shared, U: shared>(
    items: T[],
    body: fun(T): U,
    token: Option<CancellationToken> = Option.None
): Future<Result<U[], TaskFailure<TaskPoolError>>>

map_async

Maps async worker jobs with bounded dispatch rather than creating one worker per input.

public fun map_async<T: shared, U: shared>(
    items: T[],
    body: fun(T): Future<U>,
    token: Option<CancellationToken> = Option.None
): Future<Result<U[], TaskFailure<TaskPoolError>>>

close

Stops accepting work, waits for submitted jobs, then closes workers idempotently.

public fun close(): Future<Result<Unit, TaskPoolError>>