Support padding positions in fitBounds

[tristen]

• ref fitBounds padding options #995

cc @jfirebaugh

[jfirebaugh]

For the tests, try to find some bounds and padding values that make an appreciable difference in the result, and then assert equality using fixedLngLat (see other tests in the file for example). You could use fixedLngLat with the current values, and your tests would probably pass, but they wouldn't really be testing the different between not passing bounds, passing a single number, and passing an array, because the resulting difference is so small (< 0.0000000000001 degree of latitude / longitude).

[lucaswoj]

@tristen Do we still want to add this functionality? Is there anything I can do to help you 🚢 this?

[tristen]

@lucaswoj this functionality would be amazing and I'd love help with it by basically making this commit legit - my change above isn't robust enough to flexibly adjust padding on all four sides.

[lucaswoj]

Ok. In an effort to clean out stale PRs, lets close this for now, re-open #1339, and resume work when this feature becomes top-of-mind.