Fix retain cycle causing crash in RouteOverlayController on NavigationMapView reuse

[irtemed88]

Description

This PR fixes a crash due to a retain cycle caused as part of NavigationMapView reuse across multiple instances of NavigationViewController which was made available with the injection work in #3186.

• Updated Changelog

Implementation

RouteOverlayController.navigationViewDidLoad defines two closures for handling map events. Each closure captures a strong reference to self which will result in a retain cycle if the referenced NavigationMapView object is used across multiple instances of NavigationViewController. The crash occurs on presentation of the subsequent NavigationViewController

Steps to reproduce this bug are:

1. Create an instance of NavigationMapView
2. Create and present an instance of NavigationViewController and inject the map view created in step 1.
3. Dismiss the instance of NavigationViewController ensuring the map view is still retained.
4. Create and present a second instance of NavigationViewController and inject the map view created in step 1.
5. Observe crash in RouteOverlayController.showRouteIfNeeded when RouteOverlayController.navigationViewData (which is implicitly unwrapped) is accessed. The crashing controller instance will be the one created by NavigationViewController in step 2.

Using weak self references breaks the retain cycle allowing RouteOverlayController to deinit when the owning NavigationViewController is deinit.

It's unclear to me at this point if the handlers added to navigationMapView.mapView.mapboxMap also need to be cleaned, but in the interim I believe this fix should address the more immediate crash.

[irtemed88]

Looks like CI failed for some reason? Is this something I need to address before merging?

cc @MaximAlien

[azarovalex]

In theory self can be deallocated between these lines, so as a rule of thumb I always add guard let self = self else { return } if self is used more than once.
Not sure if it's critical in this case.

[MaximAlien]

Looks like CI failed for some reason? Is this something I need to address before merging?

cc @MaximAlien

Looks like it's temporary glitch on CircleCI. I've restarted failed jobs. Besides this everything looks good.