For non-synthetic NavigateEvents, the spec explicitly checks if there is an any related to any info/state and grabs it out of the incoming NavigationOptions

Let info be options["info"], if it exists; otherwise, undefined.

• https://html.spec.whatwg.org/multipage/nav-history-apis.html#performing-a-navigation-api-traversal
• https://html.spec.whatwg.org/multipage/nav-history-apis.html#dom-navigation-navigate
• https://html.spec.whatwg.org/multipage/nav-history-apis.html#dom-navigation-reload

However, for synthetic events, there's no actual initialization defined?

The NavigateEventInitData dictionary leaves the initial value not set:

dictionary NavigateEventInit : EventInit {
  NavigationType navigationType = "push";
  required NavigationDestination destination;
  boolean canIntercept = false;
  boolean userInitiated = false;
  boolean hashChange = false;
  required AbortSignal signal;
  FormData? formData = null;
  DOMString? downloadRequest = null;
  any info;
  boolean hasUAVisualTransition = false;
};

The spec text for NavigateEvent just says that it has "the value it was initialized to"

https://html.spec.whatwg.org/multipage/nav-history-apis.html#dom-navigateevent-info

The navigationType, destination, canIntercept, userInitiated, hashChange, signal, formData, downloadRequest, info, and hasUAVisualTransition attributes must return the values they are initialized to.

If it's supposed to be implied that any info; creates an ECMAScript undefined value, then why do all of the places that look at NavigationOptions check for the existence of info in the dictionary first?

After chatting with @Lubrsi I think in order for the WPT test to be correct: https://github.com/web-platform-tests/wpt/blob/master/navigation-api/navigate-event/event-constructor.html#L95 there needs to be some spec language to initialize the info property to undefined if it's not provided in the NavigateEventInitData struct.

cc @domenic