Does replace/reloadAll() affect uncontrolled pages?

[jakearchibald]

• Page registers a SW for the first time (so .active is null)
• That SW installs and calls .replace()

Should .active in the page now be the new ServiceWorker?

The algorithms say yes, but I've seen some comments in other threads suggest otherwise.

/cc @kinu @michael-nordman @alecf

[jakearchibald]

@slightlyoff what's your understanding here?

[michael-nordman]

jake, my comments didn't mean to suggest otherwise, see #198 for details.

[alecf]

+1 for .active changing - especially since you'll be able to observe the sw itself changing via the onactivate event. I think this is a really nice feature for developers, that once a user "installs" / keeps the service worker, then all open windows can be "upgraded"

[jakearchibald]

@michael-nordman ah, sorry! (although it wasn't just your comments)

Settled at .active changing

[slightlyoff]

Sorry for the delayed response. What @alecf suggests is correct. Pages which match the extent of scope have their "implicit controller" replaced in these cases and wind up being controlled (and therefore, .active changes).