[TIMOB-8746] Android: Geolocation e.coords.longitude is returning undefined.

[vishalduggal]

Associated ticket in JIRA

Verified that neither IOS nor MobileWeb have co-ordinates in the locationEvent when success is false. Updated docs to reflect this.
Updated error event to match the definition in the docs.

[rusticphilosopher]

This should be kept in place. the error this represents stems from a failure when interacting with the native LocationManager which is wrapped by TiLocation. Since the error is not generated by interaction with LocationProviderProxy, moving the error constant to LocationProviderProxy creates a disconnect.

[rusticphilosopher]

This is a good change but it was a mistake when we originally put this method inside TiConvert since this is used only by Geolocation. If possible, I think this should be moved to GeolocationModule.java along with the existing buildLocationEvent() method and events should be built in a more generic way.

For example:
-success property should exist in all events for Geolocation events and ideally be set as part of a "buildBaseEvent" or "buildGeolocationEvent" perhaps that is then invoked by the "buildLocationEvent" and "buildErrorEvent" calls
-likewise, provider should always be present

I know this is arguably scope creep, but since the fix involves changing how the event is built we can (and should IMO) take this opportunity to update the mechanism to something that will cause less problems in the future by maintaining better encapsulation, code re-use and event consistency.

[vishalduggal]

Have moved the method to the geolocation module but am uncomfortable about the scope creep. I'll file a separate ticket as an API enhancement which we will pick up in a later sprint

[vishalduggal]

JIRA isue filed TIMOB-8882

[rusticphilosopher]

FR in progress

[rusticphilosopher]

Code reviewed and FR passed. docgen and validate also passed. Accepted