DynamicLayer - wrong bounds with scrollFactor

[Zax37]

Test steps:

1. Open and run this example:
   http://labs.phaser.io/edit.html?src=src/game%20objects/tilemap/tests/visual%20culling%20check.js

2. Look at the small red, pink and yellow tiles. Move camera right. They disappear.

Any ideas why the bounds are not computed correctly? As I test that on different Phaser versions, it seems to be fine in 3.6, but broken in 3.7.1 and above.

[photonstorm]

The issue is the tile culling (rather than the scrollFactor) because it was assumed that the LayerData baseTileWidth and baseTileHeight properties were authoritative, but it appears we need to use those from the Tilemap itself, not any of the layers, to get a true bounds. I've updated the Cull function to use this and the test works fine now.

Just pinging @mikewesthad - can you see anything wrong with the current approach in https://github.com/photonstorm/phaser/blob/master/src/tilemaps/components/CullTiles.js ? It passes whatever I throw at it, but there's so many weird edge-cases in tilemaps who knows.

[Zax37]

This did not fully fix the issue. Open same example, add this line:

layer3.setScrollFactor(0.5);

Same thing happens.

[photonstorm]

This is taking up far too much time, so, for now, I'm disabling culling entirely on a layer if you have a scroll factor <> 1. Renders fine and will do for now.

[mikewesthad]

I can take a look at cull tiles tomorrow and see if anything jumps out.

Edit: Yeah, what you did with disabling culling entirely for now seems reasonable.

[Zax37]

Seems to work fine in v3.12.0-beta2 :)