This constructor is used to copy an old context to the new one after a db failure occurred.

 ServiceWorkerContextCore(
      ServiceWorkerContextCore* old_context,
      ServiceWorkerContextWrapper* wrapper);

It seems really complex to understand the state of the world after the storage was cleared and live objects copied from the old context.

I'm thinking we can just stop supporting this, and start with a clean context after the storage is wiped. Ideally we just terminate or stop listening to all running service workers and make new ones as needed.

Some backgroundn at  issue 371675  and https://docs.google.com/document/d/1eXdgnAOZC4dDDybmRpXT0t0lOYGhFf7xajOSanEb91Y/edit#