Handle current_path patch via LiveViewWatcher

[ftes]

If the LiveView path is patched by an async event (e.g. pubsub) it seems that %Live{current_path: _} is not updated.
So even if one manually awaits the pubsub event that triggers the patch, a subsequent assert_path fails.

Should assert_path be wrapped with the same timeout mechanism used for assert_has and refute_has?

[idabmat]

To help illustrate this issue I've added #212 with currently failing tests.