You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The original implementation is inherently flawed because padding and fill is applied after only checking if there are tiles in all surrounding positions, which completely disregards custom definitions or "not this tile" masks.
Tileset masks are also incorrectly sorted in priority, which does not match the vanilla implementation.
Expanding the scan size for padding and fill levels to match a custom scan width/height should also be properly trialed to determine if it actually has a notable impact on performance.
At a minimum, a vanilla tileset parsed using the custom parser and handled by the custom autotiler should behave correctly, which it currently does not.
The original implementation is inherently flawed because padding and fill is applied after only checking if there are tiles in all surrounding positions, which completely disregards custom definitions or "not this tile" masks.
Tileset masks are also incorrectly sorted in priority, which does not match the vanilla implementation.
Expanding the scan size for padding and fill levels to match a custom scan width/height should also be properly trialed to determine if it actually has a notable impact on performance.
At a minimum, a vanilla tileset parsed using the custom parser and handled by the custom autotiler should behave correctly, which it currently does not.
A very unfinished attempt at fixing some of these issues: coloursofnoise@f9e5582
The text was updated successfully, but these errors were encountered: