Chrome has a bug where a service worker can postMessage to itself to stay alive indefinitely: https://bugs.chromium.org/p/chromium/issues/detail?id=647943

You can also register at two scopes and have those workers ping-pong messages.

I'm wondering if this should be prevented at the spec level, for example:

• postMessage to a service worker that has no in-scope (or same-origin) client won't wake it up, or
• postMessage from a service worker doesn't wake up the destination service worker

If this is an implementation detail, would something heuristic like a SW can be kept alive by at most 5 messages events be reasonable?