Service Worker audit needs to test controlling scope

[brendankenny]

Our current audit only checks for an activated SW on the test page's origin. We need to look specifically at the SW's scope to make sure it can actually control the test page (e.g. by default a service worker at https://example.com/dir/sw.js can only control pages under ./dir/ and can't be the SW for https://example.com/).

Looking at the protocol response, I'm not sure if we'll be able to test the scope directly and so we may need to resort to testing that the target page's TargetID is in controlledClients.

[mutdmour]

Was looking into working on this one. But I have yet to find a service worker in the wild that has any TargetIds in controlledClients. Thoughts? Also, should it be it's own separate audit for meta to return an accurate message of the issue? Or should I try to add some property that controls what the failureDescription is?

[brendankenny]

fixed in #6609 and #6678