[Solved] What might cause tile entities shares same result?

[poopoodice]

As title, my custom furnace (with 3 inputs and 1 output) will sync all the custom furnace automatically. When one of the furnaces finish cooking, all the output slot of each custom furnace in the world will add by one. 

 

any reply will be appreciated

 

Github

https://github.com/poopoodice/Last-Day-on-Earth

Edited August 5, 2019 by poopoodice

[jaminv]

You're probably using a static variable somewhere. The TileEntities for each block should be all be different instances of your TileEntity class. I can't think of any reason why that might be so. But if one of the variables inside your TileEntity was declared as `static` then every TileEntity would share that same state. From the way you describe it, I'd guess that more than one of your variables is declared static.

[DavidM]

Post your code.

[poopoodice]

1 hour ago, jaminv said:

You're probably using a static variable somewhere. The TileEntities for each block should be all be different instances of your TileEntity class. I can't think of any reason why that might be so. But if one of the variables inside your TileEntity was declared as `static` then every TileEntity would share that same state. From the way you describe it, I'd guess that more than one of your variables is declared static.

Is it only apply on tileentity class or include block class as well?

[poopoodice]

github link at the above

[jaminv]

One thing I noticed is that you shouldn't be calling BlockWorkbench.setState() from your TileEntity. Instead, you should add a method called getActualState() from your Block. In that method you can call world.getTileEntity(pos), get data from the TE, and return it in an IBlockState. Here's how I've done it in the past.

 

	@Override
	public IBlockState getActualState(IBlockState state, IBlockAccess worldIn, BlockPos pos) {
        TileEntity tileentity = BlockHelper.getTileEntity(worldIn, pos);

        EnumFacing facing = EnumFacing.NORTH;
       	boolean active = false;

        if (tileentity instanceof TileEntityMachineBase) {
        	TileEntityMachineBase te = (TileEntityMachineBase)tileentity;
        	facing = te.getFacing();
        	active = te.isProcessing();
        }
        
        return state.withProperty(FACING, facing).withProperty(ACTIVE, active);
	}

 

I don't think this is causing your current problem, it's just the correct way to do things. I don't think you should really be calling World.setBlockState() the way that you are. It could cause you problems down the line.

 

Quote

Is it only apply on tileentity class or include block class as well?

The Block class uses a singleton pattern. Minecraft only creates one Block instance no matter how many blocks of that type there are in the world. So Block classes are basically already de facto static. You should never store anything in a Block class, but instead use block states and tile entities, which you are doing correctly. TileEntities are not singleton. Minecraft creates a separate TileEntity for every instance of a block in the [loaded] world. I don't see you doing anything wrong here. It's just important as a modder to understand what Minecraft is doing here.

 

private WorkstationEnergyStorage storage = new WorkstationEnergyStorage(6000);
...
private int cookTime, energy = storage.getEnergyStored();

 

You've got some weirdness with your energy handling. You probably shouldn't be storing the energy value in your TileEntity, but instead relying on WorkstationEnergyStorage storage to handle it for you. I'm not even sure how your machine actually works, because it doesn't look like `energy` should ever be more than 0. The above code is only called once, and it storage.getEnergyStored() will always be 0 when it's called. I don't see anything else syncing your `energy` value to the energy storage. Even if there is some code I'm not seeing somewhere, it's going to cause you problems down the line. There's no reason to store the energy value as an integer in your TileEntity. It's stored in your energy storage. Use that.

 

					ItemStack output = WorkbenchRecipes.getInstance().getWorkbenchResult(inputs[0], inputs[1]);
					
					if(!output.isEmpty())
					{
						output = WorkbenchRecipes.getInstance().getWorkbenchResult(inputs[0], inputs[1]);
						smelting = output;
						cookTime++;
						energy--;
						BlockWorkbench.setState(true, world, pos);
						if(cookTime == 200)
						{
							cookTime = 0;
							output = WorkbenchRecipes.getInstance().getWorkbenchResult(inputs[0], inputs[1]);
							smelting = output;
							inputs[0].shrink(1);
							inputs[1].shrink(1);
							handler.setStackInSlot(0, inputs[0]);
							handler.setStackInSlot(1, inputs[1]);
							if(handler.getStackInSlot(2).getCount() > 0)
							{
								handler.getStackInSlot(2).grow(1);
							}
							else
							{
								handler.insertItem(2, smelting, false);
							}
							smelting = null;
							output = null;
							cookTime = 0;
							System.out.println("finish");
							this.markDirty();
							return;
						}
					}

 

You've got some weird duplicate code in your update() method. You're initializing `output` 3 times to the exact same value. You're just wasting processor cycles. `smelting` is initialized twice but never used. `output` is set to null at the end of your code, but it's a local variable and this isn't necessary. Perhaps you were just trying different things out, but I would clean this up. It makes it easier to see what going on.

Although I've spotted a few potential issues that might cause you problems down the road, nothing really stands out as the actual problem. Are you sure you've pushed your most recent code to Github? With the energy issue, I'm not sure how you get to the point where this is actually outputing items. Perhaps I'm looking at an older version of the code?

[jaminv]

In WorkbenchRecipes:

	private final Table<ItemStack, ItemStack, ItemStack> workingList = HashBasedTable.<ItemStack, ItemStack, ItemStack>create();

 

ItemStack doesn't have a hashcode() method, so it's just using the default. There's no real way to guarantee that it will arrive at the same hash value for similar stacks, especially when you factor in item counts, NBT, etc. I'm not really sure how this is working for you at all (unless you're only putting 1 of each item in at a time), but it will almost certainly cause you problems down the line.

 

What I've had to do is create a wrapper class that implements hashcode() in a consistent manner:

	@Override
	public int hashCode() {
		// Although it might cause hash collision, we don't hash meta or NBT data. Hash collision should be fairly minimal.
		// We don't hash meta because wildcard meta data wouldn't work.
		// We don't hash NBT because then items with simple NBT data (like renaming in an anvil) would be rejected by the recipe.
		return item.getRegistryName().hashCode();
	}

 

[poopoodice]

i know all the codes are messy because Ive been working on these for quite a long time, how I wrote these code is basically read other mod's source code, but different mod had different setups, which makes me really confused. In the update() function, I tried multiple times to get it to work, so it's really messy but works. For me, the primary target is to make everything to work and then the efficient of the code runs. Thank you for you patients and all replies. I will record a video and what the problems looks like.

 

And no, the code on the github is the latest

 

Edit:

Here is the video

 

As you can see, 2 aluminium bars can produce a aluminium plate. To test the bug, I put 4 aluminium bars into the left workbench, which produces 2 plates, and I put 2 bars into the right one, which should produce only 1 plate, but in the vd, they share the results.

 

Edited August 4, 2019 by poopoodice

[jaminv]

That is strange. It's even more strange that it the count is 1 at first and then changes it to 2. Since the update() routine is never performed on the client, it's likely not a client/server desync at the tile entity. It seems like it's happening somewhere in the container, but I don't see it.

Unfortunately, I don't have time to debug through all of your code and I can't really run it yourself. You should trace through it and watch it happening step by step. I suggest subclassing ItemStackHandler and override the onContentsChanged. You can even narrow it down by creating an if (slot == 2) {} statement with a dummy operation in it, and add a breakpoint there. When you hit the breakpoint, you can look at the stack trace and see everything that it causing an update to the slot.

 

	@Override
	protected void onContentsChanged(int slot) {
		if (slot == 2) {
			int a = 0; // Breakpoint here
		}
	}

 

[Animefan8888]

10 hours ago, jaminv said:

output = WorkbenchRecipes.getInstance().getWorkbenchResult(inputs[0], inputs[1]);

Whenever you are interacting with ItemStacks in recipes you need to do ItemStack#copy to create a new instance otherwise you are using the same one for all of them.

[poopoodice]

1 hour ago, jaminv said:

That is strange. It's even more strange that it the count is 1 at first and then changes it to 2. Since the update() routine is never performed on the client, it's likely not a client/server desync at the tile entity. It seems like it's happening somewhere in the container, but I don't see it.

Unfortunately, I don't have time to debug through all of your code and I can't really run it yourself. You should trace through it and watch it happening step by step. I suggest subclassing ItemStackHandler and override the onContentsChanged. You can even narrow it down by creating an if (slot == 2) {} statement with a dummy operation in it, and add a breakpoint there. When you hit the breakpoint, you can look at the stack trace and see everything that it causing an update to the slot.

 

	@Override
	protected void onContentsChanged(int slot) {
		if (slot == 2) {
			int a = 0; // Breakpoint here
		}
	}

 

Thank you, I will try it out

[poopoodice]

1 hour ago, Animefan8888 said:

Whenever you are interacting with ItemStacks in recipes you need to do ItemStack#copy to create a new instance otherwise you are using the same one for all of them.

Is copyItemStack(ItemStack stack)  What Im looking for?

[Animefan8888]

3 minutes ago, poopoodice said:

Is copyItemStack(ItemStack stack)  What Im looking for? 

Looks like that method is from 1.9.4 so no. Use ItemStack#copy

[poopoodice]

4 minutes ago, Animefan8888 said:

Looks like that method is from 1.9.4 so no. Use ItemStack#copy

How do I use it? Since I can't find something looks similiar

[Animefan8888]

Just now, poopoodice said:

How do I use it? Since I can't find something looks similiar

ItemStack stack = ...;
newStack = stack.copy();

[jaminv]

AnimeFan8888 is right about returning ItemStacks. That would definitely cause an issue like this. You should have getWorkbenchResult() return like this:

return ((ItemStack)ent.getValue()).copy();

There's a pretty good chance that will solve your problem. It certainly can't hurt.

 

---

 

I don't know if it helps because it might be total information overload, but the recipe management system for my mod is here: https://github.com/jaminvanderberg/AdvancedMachines/tree/master/src/main/java/jaminv/advancedmachines/util/recipe. It includes a three-ingredient recipe manager that handles ingredients in any order, ore dictionary, etc. There's a lot there because recipe management can get quite complicated. But if it helps, great.

[poopoodice]

9 minutes ago, Animefan8888 said:

ItemStack stack = ...;
newStack = stack.copy();

 

9 minutes ago, jaminv said:

AnimeFan8888 is right about returning ItemStacks. That would definitely cause an issue like this. You should have getWorkbenchResult() return like this:

return ((ItemStack)ent.getValue()).copy();

There's a pretty good chance that will solve your problem. It certainly can't hurt.

 

---

 

I don't know if it helps because it might be total information overload, but the recipe management system for my mod is here: https://github.com/jaminvanderberg/AdvancedMachines/tree/master/src/main/java/jaminv/advancedmachines/util/recipe. It includes a three-ingredient recipe manager that handles ingredients in any order, ore dictionary, etc. There's a lot there because recipe management can get quite complicated. But if it helps, great.

Wow, they works! Thank you guys so much, Animefan8888 Ive subscribe to ur channel hope you can upload some nice tutorials, and jaminv thanks for all ur suggestions!

[Animefan8888]

20 minutes ago, poopoodice said:

Ive subscribe to ur channel hope you can upload some nice tutorials

I'll try, but it will likely start with the basics ?