The following exception prevents tracing the worker threads:
Error while calling actor 'tracer's method 'startTracing'" "Services.scriptSecurityManager is undefined"

This is especially easy to reproduce when enabling tracing from the browser toolbox.

This requires to manually toggle "dom.worker.console.dispatch_events_to_main_thread" to false
in order to spawn targets for workers.

I had to disable throttling for workers because of missing setTimeout/clearTimeout methods
in worker modules. Hopefully we could get access to these methods when migrating to ESM?

(In reply to Alexandre Poirot [:ochameau] from comment #1)

I had to disable throttling for workers because of missing setTimeout/clearTimeout methods
in worker modules. Hopefully we could get access to these methods when migrating to ESM?

WorkerDebuggerGlobalScope doesn't implement setTimeout, just setImmediate[1]. We can add one, but it would be good to understand the use case, especially if it's for something like batching where we might be able to add a custom API that could be potentially more power-saving aware if early and/or late transmission are both fine. By default we'd probably also cancel the timeouts when canceling the worker with no recourse.

1: This is implemented as a runnable and so will happen strictly after the microtask queue is processed but before any content runnables run. (We always process debugger runnables in preference over content runnables.)