Choosing between manual setTimeout and grunt.util.spawn.timeout for Grunt task cancellation

Goal: allow a long‑running Grunt watch task to stop promptly when an external cancellation signal is received, without leaving child processes alive.

Constraint: a manual timeout implemented with setTimeout can invoke the async callback and mark the task as failed, but it does not terminate any processes that the task may have spawned, potentially leaving orphaned workers. Conversely, grunt.util.spawn provides an options.timeout that automatically kills the spawned child process, yet this mechanism only affects work delegated to spawned children; pure‑JavaScript loops or synchronous code inside the task continue to run and cannot be stopped by the timeout.

Uncertainty: Grunt does not expose a generic abort signal that propagates to all asynchronous operations inside a task, so developers must decide which timeout strategy to adopt based on the nature of the work performed.

Which approach guarantees cleanup of both spawned children and inline JavaScript work?

How can a task that mixes spawned processes and pure‑JavaScript logic be cancelled reliably?

Is there a way to combine the two mechanisms or extend Grunt to provide a unified cancellation API?