[1.12] Modular Recipes

[Draco18s]

I have a set of recipes, crafting 9 tinydust into 1 largedust. There's about...16-18 of these for different ores (iron, gold, flour, etc). Not all of the ores exist in the game (e.g. copper) so I would prefer not registering those recipes (and their associated advancement).

 

In 1.10 and 1.11 I had this:

    private void addExtraOre(String oreName, EnumOreType oreType, Block oreBlock, Block dummyOre, int sluiceWeight, boolean canFindDefault) {

        if(!oreType.set) {
            if(config.get("SLUICE", "canFind"+oreName, canFindDefault).getBoolean()) {
                for(int i = sluiceWeight; i > 0; i--) {
                    HardLibAPI.oreMachines.addSluiceRecipe(oreBlock);
                }
            }
            String oreIn = "dustTiny"+oreName;
            ItemStack rawOreIn = new ItemStack(rawOre, 1, oreType.meta);
            List<ItemStack> dustStack = OreDictionary.getOres("dust"+oreName);
            List<ItemStack> tinyDusktStack = OreDictionary.getOres(oreIn);
            List<ItemStack> nuggetStack = OreDictionary.getOres("nugget"+oreName);
            if(dustStack.size() > 0 && tinyDusktStack.size() > 0) {
                GameRegistry.addRecipe(new ShapelessOreRecipe(dustStack.get(0),
                        oreIn, oreIn, oreIn,
                        oreIn, oreIn, oreIn,
                        oreIn, oreIn, oreIn));
                ItemStack instk = tinyDusktStack.get(0).copy();
                if(nuggetStack.size() > 0) {
                    GameRegistry.addSmelting(instk, nuggetStack.get(0), 0.1f);
                    GameRegistry.addSmelting(rawOreIn, nuggetStack.get(0), 0.1f);
                }
                instk.setCount(2);
                HardLibAPI.oreMachines.addMillRecipe(rawOreIn, instk);
                HardLibAPI.oreMachines.addSiftRecipe(oreIn, 8, dustStack.get(0));
            }
            List<ItemStack> ingotStack = OreDictionary.getOres("ingot"+oreName);
            if(ingotStack.size() > 0) {
                GameRegistry.addSmelting(dummyOre, ingotStack.get(0), 0.7f);
            }
            HardLibAPI.oreMachines.addPressurePackRecipe(rawOreIn, new ItemStack(dummyOre));
            OreDictionary.registerOre("ore"+oreName, dummyOre);
            OreDictionary.registerOre("ore"+oreName+"Hard", oreBlock);
            oreType.set = true;
        }
    }

 

Which set up an arbitrary ore for all the various things that could be done with it in my mod: smelting, grinding, sifting, packing, crafting...

Now with 1.12 and the json recipe system, how would I convert this? Is there an easier way than 16+ recipe json files?

[Choonster]

You will unfortunately need to create a recipe file for each recipe.

 

You can use the minecraft:item_exists condition to enable a recipe only when a specific Item has been registered.

Edited August 29, 2017 by Choonster

[Draco18s]

4 hours ago, Choonster said:

You can use the minecraft:item_exists condition to enable a recipe only when a specific Item has been registered.

Oh, that's good. Will that condition accept an oredict reference the way other Item specification does?

(This was the bigger deal, I can suffer 16 json files, but disabling them was more important).

[Draco18s]

26 minutes ago, Draco18s said:

Will that condition accept an oredict reference the way other Item specification does?

Nope! It does not!

[Choonster]

On 30/08/2017 at 2:15 AM, Draco18s said:

Oh, that's good. Will that condition accept an oredict reference the way other Item specification does?

(This was the bigger deal, I can suffer 16 json files, but disabling them was more important).

 

No, it only accepts an Item registry name.

 

You could create your own condition for that, but ore dictionary registration is usually done in init; which is after recipes have been loaded and conditions tested.

 

If you use a forge:ore_dict ingredient with a non-existent ore name, the recipe will show in the recipe book and in JEI but won't be craftable. Attempting to fill the recipe using the recipe book will crash the game with an ArithmeticException: / by zero (I plan to report this to Forge). Attempting to fill the recipe using JEI will simply do nothing.

 

Edit: Reported the crash here.

Edited November 16, 2017 by Choonster

[Draco18s]

4 minutes ago, Choonster said:

If you use a forge:ore_dict ingredient with a non-existent ore name, the recipe will show in the recipe book and in JEI but won't be craftable. Attempting to fill the recipe using the recipe book will crash the game with an ArithmeticException: / by zero (I plan to report this to Forge). Attempting to fill the recipe using JEI will simply do nothing.

That's workable for now.

[jabelar]

11 hours ago, Choonster said:

but ore dictionary registration is usually done in init; which is after recipes have been loaded and conditions tested.

 

I just asked about this and for 1.12.x, with the new registry events, the recommendation (as I suspected and confirmed by LexManos and discussion with Draco) is that you should now do ore dictionary registration in your item registry event handler after your item registration, as that fires before the recipe registry event.

[Choonster]

3 hours ago, jabelar said:

I just asked about this and for 1.12.x, with the new registry events, the recommendation (as I suspected and confirmed by LexManos and discussion with Draco) is that you should now do ore dictionary registration in your item registry event handler after your item registration, as that fires before the recipe registry event.

 

Indeed, I saw that some time after posting in this thread. That was news to me, since I'd previously thought that init was the de-facto standard phase for Ore Dictionary registration.

 

Even with this recommendation, I suspect a lot of mods are still doing things the old way (i.e. registering Ore Dictionary entries in init).

[Alpvax]

Fortunately (ish) those mods who register OreDict entries in init will find that their items cannot be used in other mod's recipes, so will be encouraged to move the registration.

[Choonster]

4 minutes ago, Alpvax said:

Fortunately (ish) those mods who register OreDict entries in init will find that their items cannot be used in other mod's recipes, so will be encouraged to move the registration.

 

OreIngredients will match items registered for the ore name at any time, they don't take a snapshot of the Ore Dictionary when they're created.

[Alpvax]

19 hours ago, Choonster said:

 

If you use a forge:ore_dict ingredient with a non-existent ore name, the recipe will show in the recipe book and in JEI but won't be craftable. Attempting to fill the recipe using the recipe book will crash the game with an ArithmeticException: / by zero (I plan to report this to Forge). Attempting to fill the recipe using JEI will simply do nothing.

 

3 minutes ago, Choonster said:

 

OreIngredients will match items registered for the ore name at any time, they don't take a snapshot of the Ore Dictionary when they're created.

 

So this is actually just a  case of the reported issue. Other than that it all works correctly?

These two statements seem contradictory, I'm not sure what I'm missing here.

[Choonster]

21 minutes ago, Alpvax said:

So this is actually just a  case of the reported issue. Other than that it all works correctly?

These two statements seem contradictory, I'm not sure what I'm missing here.

 

In my first post, I was talking about an ore name that never had any items registered for it. This will always return a zero-length array from Ingredient#getMatchingStacks.

 

If a mod were to register an item for that ore name at any point (i.e. even in init, postInit or when there's a world loaded), the OreIngredient would now match that item and include it in the array returned from Ingredient#getMatchingStacks. This is because OreIngredient stores a read-only view of the Ore Dictionary's item list for the specified ore name rather than creating a snapshot of the list at construction time.

[Alpvax]

Oh I see now. The issue is that Draco's recipes wouldn't be enabled if other mods registered OreDict entries too late.

Again, if that issue is discovered by users and reported correctly, the offending mods would be inclined to update.

Is it worth suggesting an oreDict version of the item_exists condition for inclusion in forge? (Although it wouldn't solve the problem, it would prevent the need for multiple modders to create their own.)

[Draco18s]

3 hours ago, Alpvax said:

Is it worth suggesting an oreDict version of the item_exists condition for inclusion in forge? (Although it wouldn't solve the problem, it would prevent the need for multiple modders to create their own.)

I would love to have it.