isActive should not be on history

[taion]

It'd be a cleaner API - lets us keep the API to history stateless, and "active" state does belong naturally to the location object.

[taion]

If we have a concept of a "routing location" that supports isActive, we can also put routes (and params, but keep the current injection) on that location object.

[mjackson]

I'd like to keep location objects as simple as possible (i.e. not add a bunch of methods that make them impossible to serialize). Also, the current design of location objects is analogous to window.location which doesn't have an isActive method.

[taion]

Fair enough there - the issue is more that it's on history, which is otherwise stateless and doesn't provide any way to read information about the current location other than by listening.

[mjackson]

I agree it's a bit messy. isActive should basically be just a utility method like isActive(route, currentState). Then it has everything it needs in order to say whether or not a route is active.

[taion]

I'm re-opening this, then.

[taion]

Route matching is a property of the history object, though - so perhaps this should be history.isActive(locationSpec, currentLocation).

[taion]

Fold this into #2614 as well.