Remove dynamic padding way name adjustment for MapWayname

[danesfeder]

Closes #1465

As reported by @hurrba, we are dynamically adjusting the map top padding based on the visibility of the way name pill. If a developer want's to adjust the top padding themselves, or just wants to center the icon, we should make this easier and not override with unexpected behavior.

This PR also adds a more straightforward API to center the location icon, which also let's us account for centering when rotating the device:

// To center the icon
int[] customPadding = {0, 0, 0, 0};
navigationView.retrieveNavigationMapboxMap().updateLocationIconWith(customPadding);

[hurrba]

looks fine, I've added a single comment in the code.

Also, while this solves the issue with centering the puck (which is what we need), I can't help but think that it could(should?) be a bit more generalized. if you want the puck to be in the top of the screen (e.g. in order to get a better overview when travelling south-bound with NavigationCamera.NAVIGATION_TRACKING_MODE_NORTH), the puck would still jump between top of screen and either center of screen (if updateLocationIconPosition(true)) or bottom of screen.

[hurrba]

should probably use mapPaddingAdjustor.isCentered() instead of false? Otherwise, clicking the recenterButton will reset the padding.

[danesfeder]

@hurrba yeah great catch, thanks!

[danesfeder]

I can't help but think that it could(should?) be a bit more generalized.

@hurrba you're saying the API should be more specific rather than just a toggle?

[hurrba]

you're saying the API should be more specific rather than just a toggle?

my point was that instead of keeping track of isCentered, we could have an array, say customMapPadding, that keeps the custom padding set by the developer (if any).

the logic would be similar, but instead of checking for isCentered here, we check if customMapPadding is set and use that value if it is.

this would account for people setting the padding to anything other than center or default (bottom of screen). again, the current solution solves our problem, but others might have other needs.

edit: this is only an issue when rotating the screen.

[danesfeder]

@hurrba yeah I see what you mean - let me work a bit more and see if I run into any snags implementing. If I do, we can ticket as a future improvement. I don't want this PR to sit / continue to block you all for too long.

[danesfeder]

@hurrba I update the APIs here more aligned with what we discussed. A custom int[] can now be provided and we will consider this padding for the life of the view thereafter. With this current setup, to return to the default, you'd need to cache [the default] on your side and set again.

This PR also considers the custom padding during rotation / when clicking the re-center button. We will always zero out the padding when clicking the overview button.

[hurrba]

@danesfeder I've tested the latest changes and it seems to be working as it should! :)
not sure if I can/should approve, but you have my thumbs up 👍

[Guardiola31337]

I've left a couple of minor comments.

Also, I'm wondering how's this SEMVER. I think we're not breaking any APIs here 🤔 Could you clarify? If it's because of NavigationMapboxMapInstanceState#isCameraTracking, NavigationMapboxMapInstanceState is already package private so without an instance you cannot access to it so I believe that wouldn't break SemVer. Actually, the "public" methods included in NavigationMapboxMapInstanceState should be package private - see comment below re Access can be package-private.

[Guardiola31337]

Minor comment - What about "exposing" (package private) updatePaddingWith instead and removing this method?

[Guardiola31337]

Access can be package-private

The same thing happens with isWaynameVisible, retrieveWayname and retrieveMapPadding methods in NavigationMapboxMapInstanceState.

[danesfeder]

Glad to hear it @hurrba, thanks for the great feedback.

@Guardiola31337 this is ready for another round 👀

[Guardiola31337]

Thanks for addressing the feedback @danesfeder 🚢

[Guardiola31337]

Hey @hurrba 👋 friendly reminder that OP landed and as you mentioned fixed your issue. You can retest with 0.23.0-SNAPSHOT to re-reconfirm 😂 This will be in the next release. Thanks!