Closed jrandolf closed 5 months ago
I don't think I agree with the arguments here in general; we have other cases where information is dynamic and we don't have corresponding lifecycle events, and I think there are clear use cases for questions like "why is this service worker active", which knowing its owner set helps answer.
But I don't object to the idea of removing this for now, if no one is actively planning to implement it. I would at least like us to make it easy to reland this feature once we do have interest.
I don't think I agree with the arguments here in general; we have other cases where information is dynamic and we don't have corresponding lifecycle events, and I think there are clear use cases for questions like "why is this service worker active", which knowing its owner set helps answer.
But I don't object to the idea of removing this for now, if no one is actively planning to implement it. I would at least like us to make it easy to reland this feature once we do have interest.
I should rephrase; there are clear use-cases, but these use-cases are low priority and have hardly ever been requested, so with our current roadmap, I suggest we remove this feature so we can focus implementing the core features and reland if there is interest as you've stated.
This PR removes the
owners
property for shared and service workers.For shared and service workers, the
owners
property makes little sense since the set of owners is dynamic. In particular, to properly implement theowners
property, we would need another lifecycle method such asscript.RealmUpdated
which contains updates to the list of owners. As of now, we've never come across the need inspecting ownership of service and shared worker owners, so removing this and adding it when the use-case arises may be more beneficial for implementors.Preview | Diff