A fix for heavy_core failing to get it's texture

[SpaceshipCaptain]

Before:

After:

Currently the heavy_core model is falling back to the texture at 0, 0 on the texture atlas because it's trying to use "all" as a texture and that doesn't exist. The heavy core's block model has it's face's textures defined like "texture": "all" but they should be defined like "texture": "#all". Without a '#' getTexture will skip the texture lookup for heavy_core.

I am not sure this is the best approach to solve this so let me know if you have a better idea. Added an extra check to make sure the textureRef string doesn't exist in this.textures before returning it. I put a cap of 7 on the loop out of an abundance of caution but it rarely reaches 3. The other option I thought of would just be an explicit if === 'all' check to only catch this one case.

If Mojang fixes how they define the heavy_core model this would be unnecessary but as far as I can tell this is how they have defined it since they added the block.

[jacobsjo]

For reference: This is reported in MC-270059.

[SpaceshipCaptain]

For reference: This is reported in MC-270059.

Ah, so it has been known about for a while. I think it's reasonable to not want to add a work around for a Mojang bug so no hard feelings if that determination is made. But it's also been a problem for over year and there is no way to know when they will get around to fixing it.

[jacobsjo]

I've looked at the code. The # seems to only be optional in the texture field of a face. That's because you can't put a texture id there directly. In the textures object, a # is required to mark it as a variable, while its interpreted as a texture id otherwise.

I think we should handle it the same way. It should simplify this logic a bit.

[SpaceshipCaptain]

I think this is what you meant but I am not 100% sure so let me know if you had something else in mind. I changed it to always assume it's getting handed a key and then only go into the while loop if gets returned another variable("#") from that initial lookup. This does catch the wrong heavy_core model and it also gets rid of the extra lookups that my first version was doing.

Sorry for the git nonsense I am a bumbling fool.