itemSize fails in some cases

[Hakoyu]

TFC version: 3.2.20
KubeJS-TFC version: 1.3.4

I am playing the modpack ATMG².
Try using KJS-TFC and modify the size of all items so that they can be stacked to 64, but the itemSize method doesn't seem to work.

TFCEvents.data(event => {
  event.itemSize('minecraft:iron_ingot', 'very_small', 'very_light');         // fail
  event.itemSize('minecraft:copper_ingot', 'very_small', 'very_light');    // fail
  event.itemSize('minecraft:gold_ingot', 'very_small', 'very_light');       // success
})

If I modify all the items, none of the items are modified successfully

TFCEvents.data(event => {
  // all fail
  Item.getList().forEach(e => {
    // console.log(e.id);
    event.itemSize(e.id, 'very_small', 'very_light');
  })
})

So how can I correctly modify the itemSize of all items ?

[Notenoughmail]

It's likely there is another item size definition whose id has a higher 'priority' than the autogenerated id of the one made in scripts. Use the search command to see what the ids are and specify an id that is prioritized.

I'm not sure why the loop didn't work for any of the items, could you try using Ingredient.all for the first parameter instead of using a loop?

[Hakoyu]

I tried event.itemSize(Ingredient.all, 'very_small', 'very_light') but it didn’t work either.
I checked again the first few items that appeared in the loop and they were all modified successfully, but when I checked the first mineral ingot I found that it had not been modified.

[Hakoyu]

Hmm... maybe kubejs has some caching mechanism? Or a problem with sequential execution?
I will get different results every time I restart the game using the same piece of code. It is possible that all three modifications are successful, or some modifications are successful, or all modifications fail.

  event.itemSize('minecraft:iron_ingot', 'very_small', 'very_light');
  event.itemSize('minecraft:gold_ingot', 'very_small', 'very_light');
  event.itemSize('minecraft:copper_ingot', 'very_small', 'very_light');

[Notenoughmail]

There is no caching. The issue is almost certainly due to the id-based load order of the definitions, the randomization is likely due to the method for automatically generating ids of definitions being based on the hash code of the ingredient, which are not guaranteed to be stable across JVM runs

Can you try specifying the id of the size definitions so they are prioritized over whatever other definitions exist for those items (again, the search command can be helpful here)

Unfortunately, I'm not exactly sure of how the ids are sorted aside from 'order loaded from disk if all files were true datapack files' and potentially that the path of the ids are compared before their namespaces