pub async fn offload<F, T>(f: F) -> TExpand description
Run f off the orchestrator thread and await its result.
Wrap CPU-bound or blocking work in offload(...) and .await it: the
orchestrator stays free to keep polling — including the system lane — while
the work runs elsewhere, so interrupt barge-in stays responsive.
§Native
Runs f on a fresh std::thread and returns its result over a
futures::channel::oneshot. Runtime-agnostic and tokio-free; a runtime
adapter can later back this with a pooled spawn_blocking.
Dropping the returned future before it resolves detaches the worker thread —
it still runs to completion, but its result is discarded. If f panics, the
worker unwinds and drops the sender; awaiting the returned future then panics
(a fresh panic noting the worker produced no result — the original payload is
not propagated).