HashMapping issue has me stumped

[Draco18s 2417]

So I'm having an issue with a wrapper class I wrote, so I can map a block+metadata to another piece of information data.  It works find in Eclipse, no problems, none.

 

Yet I've got reports from players (of a specific modpack, none the less, I haven't had reports from anyone else) of some console spam that only shows up if there is no match between the block and the info in the hashmap.  The console spam is just debug logging info trying to resolve the underlying problem.

 

Here's a section of that logging info, note the 4th listed item:

 

[15:43:27][com.draco18s.ores.block.ores.BlockHardOreBase:func_149726_b:201]: Hard Ore block, tile.ore_diamond has no flower data.  This is a bug!
[com.draco18s.ores.block.ores.BlockHardOreBase:func_149726_b:203]: Listing all mapped blocks:
[com.draco18s.ores.block.ores.BlockHardOreBase:func_149726_b:206]:     tile.ore_iron:-1 false
[com.draco18s.ores.block.ores.BlockHardOreBase:func_149726_b:206]:     tile.ore_gold:-1 false
[com.draco18s.ores.block.ores.BlockHardOreBase:func_149726_b:206]:     tile.beyondrealitycore:oreCopper:0 false
[com.draco18s.ores.block.ores.BlockHardOreBase:func_149726_b:206]:     tile.ore_diamond:-1 true
[com.draco18s.ores.block.ores.BlockHardOreBase:func_149726_b:206]:     tile.beyondrealitycore:oreTin:0 false
[com.draco18s.ores.block.ores.BlockHardOreBase:func_149726_b:206]:     tile.ore_redstone:0 false
[com.draco18s.ores.block.ores.BlockHardOreBase:func_149726_b:206]:     tile.ore_redstone:1 false

 

Here's the code that generates the logging:

		BlockWrapper bw = new BlockWrapper(this,-1);
	OreData dat = HardLibAPI.recipeManager.getOreList().get(bw);
	if(dat != null) {
		//do stuff
	}
	else {
		System.out.println("Hard Ore block, " + this.getUnlocalizedName() + " has no flower data.  This is a bug!");
		System.out.println("Listing all mapped blocks:");
		Map<BlockWrapper,OreFlowerData> map = HardLibAPI.recipeManager.getOreList();
		for(BlockWrapper b : map.keySet()) {
			System.out.println("    " + b.block.getUnlocalizedName() + ":" + b.meta + " " + bw.equals(b));
		}
	}

 

Here's the wrapper class:

public class BlockWrapper {
public Block block;
public int meta;
private int fHashCode;

/**
 * the block and metadata to match, -1 matches all
 **/
public BlockWrapper(Block block, int meta) {
	this.block = block;
	this.meta = meta;
}

@Override
public int hashCode() {
	if (fHashCode == 0) {
	      int result = HashUtils.SEED;
	      result = HashUtils.hash(result, Block.getIdFromBlock(block));
	      fHashCode = result;
	}
	return fHashCode;
    }

@Override
public boolean equals(Object aThat) {
	if(aThat instanceof BlockWrapper) {
		BlockWrapper other = (BlockWrapper)aThat;
		return other.block == this.block && (other.meta == -1 || this.meta == -1 || other.meta == this.meta);
	}
	return false;
}
}

 

I don't get what could possibly be going wrong to cause map.get(key) to be returning null here.

[diesieben07 7706]

Block IDs are not consistent, hence your hash code will not be consistent. That is a BAD idea.

[Draco18s 2417]

Block IDs are not consistent, hence your hash code will not be consistent. That is a BAD idea.

 

Even when the mods don't change? I thought it was at least stable under those conditions. And the hash is built during startup, IDs shouldn't change while the program is running and nothing about this data structure persists between runs (or even be relevant), it shouldn't even matter that the IDs changed due to a new mod install.

[Ernio 600]

IDs are not generated on startup, but per-world.

 

Whenever new world is generated - (or there are missing entries) dictionary is being created in world's data.

Dictionary binds names with newly generated IDs.

 

In world1 245 can be Copper_Ore

In world2 245 can be Tungsten_Rod

[Draco18s 2417]

Ahh, that's what I didn't know.