Function spawn_remote 

pub fn spawn_remote(
    cx: &Cx,
    node: NodeId,
    computation: ComputationName,
    input: RemoteInput,
) -> Result<RemoteHandle, RemoteError>

Spawns a named computation on a remote node.

This is the primary entry point for distributed structured concurrency. The caller specifies:

• A target NodeId identifying where to run the computation
• A ComputationName identifying what to run (no closure shipping)
• A RemoteInput containing serialized arguments

The function requires a RemoteCap from the Cx, ensuring that remote operations are impossible without explicit capability.

Region Ownership

The returned RemoteHandle is conceptually owned by the region of the calling task. When the region closes, it waits for all remote handles to resolve (or escalates per policy).

Current Contract

Attached runtimes can use deterministic harnesses such as DistributedHarness or later real transports that implement the protocol defined in this module. When no runtime is attached, spawn_remote() resolves the handle to the configured explicit fallback error immediately. This keeps behavior deterministic and avoids detached wall-clock work outside structured concurrency.

Errors

Returns RemoteError::NoCapability if the context does not have a RemoteCap.

Example

use asupersync::remote::{spawn_remote, NodeId, ComputationName, RemoteInput};

let mut handle = spawn_remote(
    &cx,
    NodeId::new("worker-1"),
    ComputationName::new("encode_block"),
    RemoteInput::new(serialized_data),
)?;

let result = handle.join(&cx).await?;
if let RemoteOutcome::Success(data) = result {
    // process data
}