Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Add networking documentation #42

Merged
merged 2 commits into from
Dec 31, 2023
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
122 changes: 122 additions & 0 deletions docs/networking/configuration-tasks.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,122 @@
# Using Configuration Tasks

The networking protocol for the client and server has a specific phase where the server can configure the client before the player actually joins the game.
This phase is called the configuration phase, and is for example used by the vanilla server to send the resource pack information to the client.

This phase can also be used by mods to configure the client before the player joins the game.

## Registering a configuration task
The first step to using the configuration phase is to register a configuration task.
This can be done by registering a new configuration task in the `OnGameConfigurationEvent` event.
```java
@SubscribeEvent
public static void register(final OnGameConfigurationEvent event) {
event.register(new MyConfigurationTask());
}
```
The `OnGameConfigurationEvent` event is fired on the mod bus, and exposes the current listener used by the server to configure the relevant client.
A modder can use the exposed listener to figure out if the client is running the mod, and if so, register a configuration task.

## Implementing a configuration task
A configuration task is a simple interface: `ICustomConfigurationTask`.
This interface has two methods: `void run(Consumer<CustomPacketPayload> sender);`, and `ConfigurationTask.Type type();` which returns the type of the configuration task.
The type is used to identify the configuration task.
An example of a configuration task is shown below:
```java
public record MyConfigurationTask implements ICustomConfigurationTask {
public static final ConfigurationTask.Type TYPE = new ConfigurationTask.Type(new ResourceLocation("mymod:my_task"));

@Override
public void run(final Consumer<CustomPacketPayload> sender) {
final MyData payload = new MyData();
sender.accept(payload);
}

@Override
public ConfigurationTask.Type type() {
return TYPE;
}
}
```

## Acknowledging a configuration task
Your configuration is executed on the server, and the server needs to know when the next configuration task can be executed.
This is done by acknowledging the execution of said configuration task.

There are two primary ways of achieving this:

### Capturing the listener
When the client does not need to acknowledge the configuration task, then the listener can be captured, and the configuration task can be acknowledged directly on the server side.
```java
public record MyConfigurationTask(ServerConfigurationListener listener) implements ICustomConfigurationTask {
public static final ConfigurationTask.Type TYPE = new ConfigurationTask.Type(new ResourceLocation("mymod:my_task"));

@Override
public void run(final Consumer<CustomPacketPayload> sender) {
final MyData payload = new MyData();
sender.accept(payload);
listener.finishCurrentTask(type());
}

@Override
public ConfigurationTask.Type type() {
return TYPE;
}
}
```
To use such a configuration task, the listener needs to be captured in the `OnGameConfigurationEvent` event.
```java
@SubscribeEvent
public static void register(final OnGameConfigurationEvent event) {
event.register(new MyConfigurationTask(event.listener()));
}
```
Then the next configuration task will be executed immediately after the current configuration task has completed, and the client does not need to acknowledge the configuration task.
Additionally, the server will not wait for the client to properly process the send payloads.

### Acknowledging the configuration task
When the client needs to acknowledge the configuration task, then you will need to send your own payload to the client:
```java
public record AckPayload() implements CustomPacketPayload {
public static final ResourceLocation ID = new ResourceLocation("mymod:ack");

@Override
public void write(final FriendlyByteBuf buffer) {
// No data to write
}

@Override
public ResourceLocation id() {
return ID;
}
}
```
When a payload from a server side configuration task is properly processed you can send this payload to the server to acknowledge the configuration task.
```java
public void onMyData(MyData data, ConfigurationPayloadContext context) {
context.submitAsync(() -> {
blah(data.name());
})
.exceptionally(e -> {
// Handle exception
context.packetHandler().disconnect(Component.translatable("my_mod.configuration.failed", e.getMessage()));
return null;
})
.thenAccept(v -> {
context.replyHandler().send(new AckPayload());
});
}
```
Where `onMyData` is the handler for the payload that was sent by the server side configuration task.

When the server receives this payload it will acknowledge the configuration task, and the next configuration task will be executed:
```java
public void onAck(AckPayload payload, ConfigurationPayloadContext context) {
context.taskCompletedHandler().onTaskCompleted(MyConfigurationTask.TYPE);
}
```
Where `onAck` is the handler for the payload that was sent by the client.

## Stalling the login process
When the configuration is not acknowledged, then the server will wait forever, and the client will never join the game.
So it is important to always acknowledge the configuration task, unless the configuration task failed, then you can disconnect the client.
22 changes: 7 additions & 15 deletions docs/networking/entities.md
Original file line number Diff line number Diff line change
@@ -1,26 +1,18 @@
Entities
========
# Entities

In addition to regular network messages, there are various other systems provided to handle synchronizing entity data.

Spawn Data
----------

In general, the spawning of modded entities is handled separately, by Forge.

:::note
This means that simply extending a vanilla entity class may not inherit all its behavior. You may need to implement certain vanilla behaviors yourself.
:::
## Spawn Data
Since 1.20.2 Mojang introduced the concept of Bundle packets, which are used to send entity spawn packets together.
This allows for more data to be sent with the spawn packet, and for that data to be sent more efficiently.

You can add extra data to the spawn packet Forge sends by implementing the following interface.

### IEntityAdditionalSpawnData

### IEntityWithComplexSpawn
If your entity has data that is needed on the client, but does not change over time, then it can be added to the entity spawn packet using this interface. `#writeSpawnData` and `#readSpawnData` control how the data should be encoded to/decoded from the network buffer.
Alternatively you can override the method `sendPairingData(...)` which is called when the entity is paired with a client. This method is called on the server, and can be used to send additional payloads to the client within the same bundle as the spawn packet.

Dynamic Data
------------

## Dynamic Data
### Data Parameters

This is the main vanilla system for synchronizing entity data from the server to the client. As such, a number of vanilla examples are available to refer to.
Expand Down
10 changes: 4 additions & 6 deletions docs/networking/index.md
Original file line number Diff line number Diff line change
@@ -1,5 +1,4 @@
Networking
==========
# Networking

Communication between servers and clients is the backbone of a successful mod implementation.

Expand All @@ -12,9 +11,8 @@ There are two primary goals in network communication:

The most common way to accomplish these goals is to pass messages between the client and the server. These messages will usually be structured, containing data in a particular arrangement, for easy sending and receiving.

There are a variety of techniques provided by Forge to facilitate communication mostly built on top of [netty][].

The simplest, for a new mod, would be [SimpleImpl][channel], where most of the complexity of the netty system is abstracted away. It uses a message and handler style system.
There is a technique provided by NeoForge to facilitate communication mostly built on top of [netty][].
This technique can be used by listening for the `RegisterPayloadHandlerEvent` event, and then registering a specific type of [payloads][], its reader, and its handler function to the registrar.

[netty]: https://netty.io "Netty Website"
[channel]: ./simpleimpl.md "SimpleImpl in Detail"
[payloads]: ./payload.md "Registering custom Payloads"
98 changes: 98 additions & 0 deletions docs/networking/payload.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,98 @@
# Registering Payloads

Payloads are a way to send arbitrary data between the client and the server. They are registered using the `IPayloadRegistrar` that can be retrieved for a given namespace from the `RegisterPayloadHandlerEvent` event.
```java
@SubscribeEvent
public static void register(final RegisterPacketHandlerEvent event) {
final IPayloadRegistrar registrar = event.registrar("mymod");
}
```

Assuming we want to send the following data:
```java
public record MyData(String name, int age) {}
```

Then we can implement the `CustomPacketPayload` interface to create a payload that can be used to send and receive this data.
```java
public record MyData(String name, int age) implements CustomPacketPayload {

public static final ResourceLocation ID = new ResourceLocation("mymod", "my_data");

public MyData(final FriendlyByteBuf buffer) {
this(buffer.readUtf(), buffer.readInt());
}

@Override
public void write(final FriendlyByteBuf buffer) {
buffer.writeUtf(name());
buffer.writeInt(age());
}

@Override
public ResourceLocation id() {
return ID;
}
}
```
As you can see from the example above the `CustomPacketPayload` interface requires us to implement the `write` and `id` methods. The `write` method is responsible for writing the data to the buffer, and the `id` method is responsible for returning a unique identifier for this payload.
We then also need a reader to register this later on, here we can use a custom constructor to read the data from the buffer.

Finally, we can register this payload with the registrar:
```java
@SubscribeEvent
public static void register(final RegisterPacketHandlerEvent event) {
final IPayloadRegistrar registrar = event.registrar("mymod");
registar.play(MyData.ID, MyData::new, handler -> handler
.client(ClientPayloadHandler.getInstance()::handleData)
.server(ServerPayloadHandler.getInstance()::handleData));
}
```
Dissecting the code above we can notice a couple of things:
- The registrar has a `play` method, that can be used for registering payloads which are send during the play phase of the game.
- Not visible in this code are the methods `configuration` and `common`, however they can also be used to register payloads for the configuration phase. The `common` method can be used to register payloads for both the configuration and play phase simultaneously.
- The constructor of `MyData` is used as a method reference to create a reader for the payload.
- The third argument for the registration method is a callback that can be used to register the handlers for when the payload arrives at either the client or server side.
- The `client` method is used to register a handler for when the payload arrives at the client side.
- The `server` method is used to register a handler for when the payload arrives at the server side.
- There is additionally a secondary registration method `play` on the registrar itself that accepts a handler for both the client and server side, this can be used to register a handler for both sides at once.

Now that we have registered the payload we need to implement a handler.
For this example we will specifically take a look at the client side handler, however the server side handler is very similar.
```java
public class ClientPayloadHandler {

private static final ClientPayloadHandler INSTANCE = new ClientPayloadHandler();

public static ClientPayloadHandler getInstance() {
return INSTANCE;
}

public void handleData(final MyData data, final PlayPayloadContext context) {
// Do something with the data, on the network thread
blah(data.name());

// Do something with the data, on the main thread
context.workHandler().submitAsync(() -> {
blah(data.age());
})
.exceptionally(e -> {
// Handle exception
context.packetHandler().disconnect(Component.translatable("my_mod.networking.failed", e.getMessage()));
return null;
});
}
}
```
Here a couple of things are of note:
- The handling method here gets the payload, and a contextual object. The contextual object is different for the play and configuration phase, and if you register a common packet, then it will need to accept the super type of both contexts.
- The handler of the payload method is invoked on the networking thread, so it is important to do all the heavy work here, instead of blocking the main game thread.
- If you want to run code on the main game thread you can use the `workHandler` of the context to submit a task to the main thread.
- The `workHandler` will return a `CompletableFuture` that will be completed on the main thread, and can be used to submit tasks to the main thread.
- Notice: A `CompletableFuture` is returned, this means that you can chain multiple tasks together, and handle exceptions in a single place.
- If you do not handle the exception in the `CompletableFuture` then it will be swallowed, **and you will not be notified of it**.

Now that you know how you can facilitate the communication between the client and the server for your mod, you can start implementing your own payloads.
With your own payloads you can then use those to configure the client and server using [Configuration Tasks][]

[Configuration Tasks]: ./configuration-tasks.md
Loading