Uli Bubenheimer
Uli Bubenheimer
@Del-S What I see in the case of the workflow from the video is the GoogleMap View lifecycle ending up in CREATED state after the user clicks the map navigation...
I've looked at it once more and I see that there is more complexity involved. Will look into it some more.
What I see now (in the case from the video) is that the LifecycleOwner changes. The View does see the ON_DESTROY lifecycle event from the old lifecycle owner, and around...
@Del-S for the time being you could try applying the `key(lifecycleOwner)` workaround yourself, wrap it around the GoogleMap(). Thinking about it, you will lose everything you remembered in the `GoogleMap...
@Del-S the other workaround that you are suggesting may be feasible. I did not mention it, because the exact desired implementation and behavior at the time of View seeing ON_DESTROY...
@kikoso are you open to reviewing this PR?
@wangela can this PR receive a review, please?
@kikoso The types in `model` are not generally immutable, which commonly would make them non-stable for Compose purposes. They would be covered by the upcoming strong skipping changes for handling...
@kikoso Actually, `maps-compose` uses some mutable types from `model` as well (e.g. Marker), but they need to be handled via strong skipping changes, not stability configuration file, for the the...
To be clear: replacing `MarkerState.dragState` (START/DRAG/END) with `MarkerState.isDragging` does not lose information, because the information was not valid in the first place, due to the State vs. Event mismatch in...