[1.8.9] [SOLVED] PlayerTickEvent not applying damage when it should

[Daeruin]

I'm working on a little feature that is supposed to apply damage to the player when they are walking or sprinting across a specific kind of block on a semi-random basis. As a test, I have created a PlayerTickEvent:

 

 

@SubscribeEvent
public void onPlayertick(PlayerTickEvent event) {

	ticksCounted += 1;

	Entity entity = event.player;
	BlockPos posBelowPlayer = new BlockPos((int) Math.floor(entity.posX), (int) Math.floor(entity.posY) - 0.2D, (int) Math.floor(entity.posZ));
	Block blockBelow = entity.worldObj.getBlockState(posBelowPlayer).getBlock();

	if (ticksCounted % 100 == 0
			&& entity instanceof EntityPlayer
			&& blockBelow == Blocks.dirt
			&& entity.onGround
			&& !entity.isSneaking()
			&& entity.worldObj.getDifficulty() != EnumDifficulty.PEACEFUL) {
		System.out.println("ALL CHECKS PASSED");
		event.player.attackEntityFrom(DamageSource.generic, 1.0F);
	}

 

 

As you can see, I'm not actually checking if the player is moving yet. I'm using dirt for now (because it's plentiful). I haven't applied any kind of randomization. I'm just trying to get it to happen every 100 ticks as a test. But it isn't happening. I'm keeping my player just standing there on dirt blocks, and can see my debug statement "ALL CHECKS PASSED" occurring every 100 ticks like clockwork, but the damage is only being applied every 20th time or so. It seems random. Any idea what is preventing the damage from being applied?

 

As a secondary question, how would you recommend checking if the player is walking or sprinting over the block? I was thinking that as long as the player is on the ground (event.entity.onGround == true) and not sneaking (event.entity.isSneaking() == true) and their X movement isn't 0 (event.entity.motionX != 0), then that's enough to apply the damage. I've implemented the first two, but not the last. Does that sound like it would work?

 

Oh, and a third, very minor, question. This is how I'm registering this event:

 

        FMLCommonHandler.instance().bus().register(new PlayerTickEventHandler());

 

But Eclipse is telling me that "bus()" is deprecated. It seems to be working as is, but I'm curious what is the correct way to do this for 1.8.9. I did some searching and couldn't find a quick answer, so I thought I'd tack it onto this post.

 

Thanks in advance!

 

Edit: Added clarification to subject title

[Daeruin]

Duh. Just figured out the last one right after I posted. Still wondering about the first two questions, though.

[coolAlias]

Since 1.8.9, all of the event buses have been consolidated so there is only one bus, the MinecraftForge.EVENT_BUS.

 

As for the block damage, yeah, that can work, or you could check if motionX^2 + motionZ^2 is greater than a certain threshold and not worry about sneaking at all.

[Daeruin]

The motion part isn't the problem—the check for sneaking is working OK for now. The problem is that the damage isn't getting applied for some reason.

[Daeruin]

I didn't mean to sound dismissive. I did ask for advice on player movement, and I appreciate the tip. I'm just really frustrated that I can't get the damage to be applied consistently.

[Daeruin]

Could this problem have something to do with whether the event is firing on client vs. server side? Or maybe the tick phase? I'm at work so I can't check, but I'll try tonight. Any further insight on this would be appreciated.

[Daeruin]

Further testing shows the event is only happening on the client side (event.player.worldObj.isRemote is always TRUE, and the event doesn't fire at all if I register it only in the server proxy). Does that mean I need to set up a packet handler to make sure the damage gets communicated to the server? I haven't delved into packet handling yet.

[coolAlias]

PlayerTickEvent should occur on BOTH client and server - where are you registering your event handler from?

[Daeruin]

OK, I kind of figured out what's causing this weirdness. To answer your question, I am registering the event handler in my Common Proxy during post initialization. I tried registering it in Server Proxy and Client Proxy, in pre, init, and post, just to see what would happen. It's interesting that the event doesn't fire at all when registered in Server Proxy, but I think that's a red herring.

 

I created a new event without all the extra stuff. All it does is increment the tick value and print to the console the tick value, phase, and whether the world is remote. I found that I get ticks from client and server in roughly equal numbers, but in irregular patterns—sometimes it's 2 ticks in a row from the server, then 2 from client, for 20 or 30 ticks; then I'll get 10 ticks in a row from the server and 10 from the client for a little while. So it's a toss of the dice whether any given tick is coming from the client or the server. Either way, the phase alternates every other tick regularly.

 

So next I set it up to check if the world is remote every X number of ticks. For example:

 

public class TickTest {

int tick;

@SubscribeEvent
public void onTick(PlayerTickEvent event) {
	tick += 1;
	if (tick % 50 == 0) {
		System.out.println("Tick: " + tick + " - Phase: " + (event.phase) + " - World remote? " + event.player.worldObj.isRemote);
		event.player.attackEntityFrom(DamageSource.generic, 1.0F);
	}
}
}

 

The code above has a tick firing from the server about 5 out of 15 times. I guess that's not too weird, given the uneven patterns I was seeing when checking every tick. I'm guessing it would have evened out to 50% of the time if I waited long enough. I also checked the tick phase, and it was always END, no matter what. Sample from log output:

 

 

[22:47:54] [Client thread/INFO] [sTDOUT]: [com.daeruin.primalcraft.events.TickTest:onTick:17]: Tick: 50 - Phase: END - World remote? true
[22:47:55] [Client thread/INFO] [sTDOUT]: [com.daeruin.primalcraft.events.TickTest:onTick:17]: Tick: 100 - Phase: END - World remote? true
[22:47:56] [server thread/INFO] [sTDOUT]: [com.daeruin.primalcraft.events.TickTest:onTick:17]: Tick: 150 - Phase: END - World remote? false
[22:47:56] [Client thread/INFO] [sTDOUT]: [com.daeruin.primalcraft.events.TickTest:onTick:17]: Tick: 200 - Phase: END - World remote? true
[22:47:57] [server thread/INFO] [sTDOUT]: [com.daeruin.primalcraft.events.TickTest:onTick:17]: Tick: 250 - Phase: END - World remote? false
[22:47:58] [Client thread/INFO] [sTDOUT]: [com.daeruin.primalcraft.events.TickTest:onTick:17]: Tick: 300 - Phase: END - World remote? true
[22:47:58] [server thread/INFO] [sTDOUT]: [com.daeruin.primalcraft.events.TickTest:onTick:17]: Tick: 350 - Phase: END - World remote? false
[22:47:59] [Client thread/INFO] [sTDOUT]: [com.daeruin.primalcraft.events.TickTest:onTick:17]: Tick: 400 - Phase: END - World remote? true
[22:48:00] [server thread/INFO] [sTDOUT]: [com.daeruin.primalcraft.events.TickTest:onTick:17]: Tick: 450 - Phase: END - World remote? false
[22:48:00] [Client thread/INFO] [sTDOUT]: [com.daeruin.primalcraft.events.TickTest:onTick:17]: Tick: 500 - Phase: END - World remote? true
[22:48:01] [Client thread/INFO] [sTDOUT]: [com.daeruin.primalcraft.events.TickTest:onTick:17]: Tick: 550 - Phase: END - World remote? true
[22:48:02] [Client thread/INFO] [sTDOUT]: [com.daeruin.primalcraft.events.TickTest:onTick:17]: Tick: 600 - Phase: END - World remote? true

 

 

Every time the tick comes from the server, the code to damage the player works (well, it takes a couple hundred ticks upon starting up the world before the damage starts registering). When the tick comes from the client, no damage is applied.

 

If I change the above code to check every 100 ticks, I get almost every single tick occurring on the client (in one test I got a single tick from the server after waiting 4000 ticks, in another test it took 2000 ticks). The reason my previous tests seemed to be occurring only on the client is because I was only checking every 100 ticks. If I change it to every 101 ticks, I get roughly every other one from the server.

 

So it appears if I check every X ticks, I just need to randomly pick the right value for X to get the code to execute regularly. Unless someone else can see a pattern that I can't.

[coolAlias]

The player doesn't take any damage on the client because it's not supposed to - changes to data should only ever occur on the server, which then, if the data is important for the client to know about (e.g. for rendering GUIs), notifies the client of the current values.

 

Since you are testing on single player, the client and server are not really separated - the integrated server runs in the same program instance as the client - so your event handler only has one instance shared between the two. When you increment the tick counter on either side, the increased value is also seen on the other side, so both sides are incrementing the same counter. You would not notice this behavior when running on a dedicated server.

 

Anyway, to fix your issue, nest all of your event code in an 'if (!event.player.worldObj.isRemote)' statement - that will ensure your code only runs on the logical server side, which is what you want.

[Daeruin]

Yes, I tried that after I posted as it seemed the next logical step. The behavior I'm seeing is that I can't predict whether any given tick is occurring on the server or the client. So if I check for whether the world is remote, THEN check whether a specific number of ticks have occurred (e.g., tickCount % 50), chances are good that specific tick was on client side (for example, see ticks 50 and 100 from the sample output in my last post), and nothing happens. In fact, when I'm checking every 100 ticks, the vast majority of those ticks are occurring on client side--embedding that kind of check inside a check for whether the world is remote will practically guarantee that the event code never occurs, especially after adding the other conditions I want to add (walking or sprinting on a specific kind of block).

 

Maybe instead of checking every 100 ticks, I need to just generate a random number that should let the code execute 1% of the time. Same net result, hopefully.

[Choonster]

If the tick is happening on the server side and the

Phase

is correct (it doesn't really matter whether it runs in

START

or

END

as long as you pick one), increment your tick counter and check if the appropriate amount of ticks have passed. If you don't check the side and

Phase

, the handler will be called twice per tick per side for every player.

 

You can't store the tick counter as a field of your event handler, since the same handler will be called for every player on the server. The more players there are, the more often one of them (there's no guarantee as to which one) will take damage. Event handlers are essentially singletons like

Block

s or

Item

s, so you can't store data for a specific object in their fields.

 

The

Entity#ticksExisted

field stores the number of ticks the entity has been alive for. Use this instead of maintaining your own tick counter.

[Daeruin]

Good point. I really haven't been thinking about server use at all, but if the event gets called every time ANY player ticks . . . yeah. Not quite what I was imagining.

 

So I check if the tick is happening on server side, in the correct phase, and whether Entity#ticksExisted % 100 is zero, then apply the damage. That sounds like it would be more reliable. I'll try it when I get home tonight.

 

I just realized that I probably don't need to check if the entity is a player, because I'm using PlayerTickEvent. In my earlier code I created an Entity variable named "entity" and assigned it the value of event.player, then later on I checked to see if "entity" is an instance of EntityPlayer. Seems a little redundant, no?

[Choonster]

So I check if the tick is happening on server side, in the correct phase, and whether Entity#ticksExisted % 100 is zero, then apply the damage. That sounds like it would be more reliable. I'll try it when I get home tonight.

That should work, yes.

 

I just realized that I probably don't need to check if the entity is a player, because I'm using PlayerTickEvent. In my earlier code I created an Entity variable named "entity" and assigned it the value of event.player, then later on I checked to see if "entity" is an instance of EntityPlayer. Seems a little redundant, no?

That is indeed redundant. As the name suggests,

PlayerTickEvent

is only fired for players.

[Daeruin]

Awesome, that worked perfectly!

 

However, now my check for movement is failing. It appears that whenever the world is not remote, event.player.motionX is always zero. Any other ideas how I can accomplish this? I need the event to trigger when the player is walking or sprinting across a certain type of block, and I need it to apply damage.

[coolAlias]

Player motion is unfortunately handled almost entirely on the client. What you can do is listen on the client for the player movement, then send a packet to the server to perform the damage.

[Draco18s]

Or you can compare the player's lastPos values with its pos values and see if they have changed.

[Daeruin]

Draco, thanks for the suggestion. I tried it out. Unfortunately, after making sure that the world is not remote and the phase is START (or END, doesn't make a difference), (event.player.posX == event.player.lastTickPosX) only seems to be false about one tick out of twenty, even when I'm moving constantly.

 

I'm trying to learn about packets so I can try coolAlias's suggestion. It's taking a while, but I'll try to report on that. In the meantime, any other suggestions are welcome.

[Daeruin]

So I spent a while learning about packets and set up a very basic thirst system while following some tutorials by coolAlias, diesieben, and jabelar. Cool stuff.

 

After doing all that, I changed my PlayerTickEvent to listen on the client side and created a packet (called PlayerToeStub, haha) to perform the damage on the server side. Unfortunately, it doesn't seem to work. I registered the packet as a server message and set up the handler to run on the correct thread (I think) but a check for world.isRemote inside the runnable shows that it still seems to be running on the client, and the damage doesn't get applied. I must be doing something wrong, but I don't know what it is. Any ideas or advice?

 

PlayerTickEvent (listens on client for the right conditions, then sends the packet)

 

public class PrimalPlayerTickEvent {

@SubscribeEvent
public void onTick(PlayerTickEvent event) {

	Entity entity = event.player;
	BlockPos posBelowPlayer = new BlockPos((int) Math.floor(entity.posX), (int) Math.floor(entity.posY) - 0.2D, (int) Math.floor(entity.posZ));
	Block blockBelow = entity.worldObj.getBlockState(posBelowPlayer).getBlock();

	if (entity.worldObj.isRemote) {
		if (event.phase == Phase.END) {
			if (entity.ticksExisted % 100 == 0) {
				System.out.println("Block below: " + blockBelow);
				if (blockBelow == Blocks.dirt || blockBelow == Blocks.grass || blockBelow == Blocks.stone || blockBelow == Blocks.gravel) {
					System.out.println("On ground: " + entity.onGround);
					if (entity.onGround) {
						System.out.println("Sneaking: " + entity.isSneaking());
						if (!entity.isSneaking()) {
							System.out.println("Moving: " + entity.motionX);
							if (entity.motionX != 0) {
								if (entity.worldObj.getDifficulty() != EnumDifficulty.PEACEFUL) {
									System.out.println("ALL CHECKS PASSED");
									PrimalPacketHandler.INSTANCE.sendToServer(new PrimalToeStubPacket(event.player));
	}}}}}}}}

}
}

 

 

Packet (sends damage to be applied to player on server)

 

public class PrimalToeStubPacket implements IMessage {

private int playerId;

public PrimalToeStubPacket() {}

public PrimalToeStubPacket(EntityPlayer player) {
	this.playerId = player.getEntityId();
}

@Override
public void fromBytes(ByteBuf buffer) {
	this.playerId = buffer.readInt();
}

@Override
public void toBytes(ByteBuf buffer) {
	buffer.writeInt(playerId);
}

    public static class PrimalToeStubHandler implements IMessageHandler<PrimalToeStubPacket, IMessage> {

	@Override
    public IMessage onMessage(final PrimalToeStubPacket message, final MessageContext ctx) {
        IThreadListener mainThread = (WorldServer) ctx.getServerHandler().playerEntity.worldObj;
        mainThread.addScheduledTask(new Runnable() {
            @Override
            public void run() {
                Entity entity = ctx.getServerHandler().playerEntity;
    			if (entity instanceof EntityPlayer)
    			{
    				entity.attackEntityFrom(DamageSource.generic, 1.0F);
    			}
            }
        });
        return null;
    }
    
    }

}

 

 

(Side note: I just realized that I'm creating a playerId variable but not using it anywhere—I think that's a holdover from the thirst system that I started to implement based on something I read. But I'm not using the playerId in my thirst packet, either, and everything seems to be working fine in single player mode. Maybe it will become problematic with multiple players?)

 

Packet registration in CommonProxy:

 

public class CommonProxy {

public void preInit(FMLPreInitializationEvent e) {

	PrimalPacketHandler.INSTANCE.registerMessage(PrimalThirstHandler.class, PrimalThirstPacket.class, 0, Side.CLIENT);
	PrimalPacketHandler.INSTANCE.registerMessage(PrimalToeStubHandler.class, PrimalToeStubPacket.class, 1, Side.SERVER);

}

 

[Daeruin]

IT WORKS!

 

Also, I need to quit being a doofus and remember not to test this kind of stuff while in Creative Mode.