Force move to pick center even if max bounds does not fit

[jetpeter]

If bounds do not fit first attempt to zoom up to the max zoom.
If center still cannot work then just set center with max zoom
and live with the grey borders.

[ibrierley]

This feels like it "jumps" too much, rather than sliding into place like it does currently (though this would fix it on an init better). Quick video https://drive.google.com/file/d/1m5oDdjuuvtE10PIKd3EjBRWp-jzWe3Hv/view to try and explain it a bit!

[jetpeter]

Yes I agree, It works good with a really small maxBounds but breaks down with larger areas. Im working on making an adjustment.

[ibrierley]

Does the original work for you apart from the init phase ? Just wondering if that could be an option, rather than changing behaviour for people that exists already.

[ibrierley]

(hmm I guess you never get past that to know!)

[jetpeter]

It does work actually. The problem is init() calls move or setBounds which also eventually calls move. I was avoiding adding a new boolean param to the move method. Im experimenting now with only zooming if _lastCenter is null.

[jetpeter]

Has anyone else ran into an issue where you cannot pan the map at some locations when zoomed out fully and and rotated. I assumed it was a bug I created, but its showing up in the main branch as well. If you zoom and and then zoom out again you can pan but if you just rotate from all the way zoomed out it does not work. I'll try to address that too.

[jetpeter]

I think the map being frozen is an expected side effect of being zoomed all the way out with maxBounds fixing it would require adjusting zoom when panning which would be unpleasant.

[JaffaKetchup]

Sorry @jetpeter, bit out of the loop, is there anything you need me to do?

[jetpeter]

@JaffaKetchup This just needs re-review. @ibrierley made a good point about the map jumping when you hit bound which was caused by the re-center zoom adjustment being called when it didn't need to be. I have since changed the code to only adjust zoom for max bounds if the late non null variables that are set in the move method are null. This means that the code should only be run during the first call to move which is from the init method. The change now should have not negative impact on existing users. The only change users will see is the map is zoomed more on devices where the view port size is smaller than the maxBounds, or in the worst case they will see their max bounds in the center of the map the most zoomed in it can be. There should never be an exception when using maxBounds when using this change.

[JaffaKetchup]

@jetpeter I'm giving priority to #1333 for now, as it's a breaking change. Then I'll review and merge also as part of v3.0.0. Thanks for your contributions so far, and sorry for the delay :)

[ibrierley]

I think I'm ok with this. It seems to test ok in the example, which was its original intent. I'm not necessarily a fan of changing a zoom that isn't solicited (wondering if there's some odd edge cases then), but neither am I fan of getting exceptions :D.

[JaffaKetchup]

Hi @jetpeter, I'm struggling to resolve the conflicts after #1333. If possible, could you resolve the conflicts please? I've had a go myself, but the _lastCenter variable seems to have vanished into thin air.

[jetpeter]

Yea, I'll get it done today

[JaffaKetchup]

Closing as discussed in Discord server.