Client-side NPE in ClientboundSableUDPActivationPacket when UDP channel is null (tunnel/proxy environments) #1080

Describe the bug

When playing behind an NAT traversal tunnel (e.g., EasyTier-based Terracotta/陶瓦联机), the client crashes with a NullPointerException in ClientboundSableUDPActivationPacket.handle() because the UDP channel is null. This causes the client to be kicked with "Invalid player data" (无效的玩家数据).

The issue is client-side, distinct from #175 (server-side NPE in SableUDPServer.getServer()).

Steps to reproduce

1. Use any tunneling/proxy tool that creates a virtual network interface (EasyTier, frp, ngrok, etc.)
2. Join a server with Sable installed
3. The server sends sable:udp_activation → client's UDP channel is null → NPE

Crash log

[Render thread/INFO] [dev.ryanhcode.sable.Sable/]: Client UDP channel inactive
[Render thread/INFO] [dev.ryanhcode.sable.Sable/]: Closed UDP channel!
[Render thread/INFO] [dev.ryanhcode.sable.Sable/]: Received authentication request, sending response over UDP to localhost/127.0.0.1:25565
[Render thread/ERROR] [net.minecraft.util.thread.BlockableEventLoop/FATAL]: Error executing task on Client
java.util.concurrent.CompletionException: java.lang.NullPointerException: Cannot invoke "io.netty.channel.Channel.eventLoop()" because "channel" is null
    at TRANSFORMER/sable@1.2.2/dev.ryanhcode.sable.network.packets.tcp.ClientboundSableUDPActivationPacket.handle(ClientboundSableUDPActivationPacket.java:51)
    ...
[Render thread/WARN] [net.minecraft.client.multiplayer.ClientCommonPacketListenerImpl/]: Client disconnected with reason: 无效的玩家数据

Then the player reconnects and it crashes again identically.

Environment

• Mod version: sable-neoforge-1.21.1-1.2.2.jar
• Minecraft: 1.21.1 NeoForge
• Launcher: HMCL with Terracotta (EasyTier-based tunneling)
• Relevant config: sable-common.toml

Root cause

ClientboundSableUDPActivationPacket.handle() (line 51) calls channel.eventLoop() without checking if channel is null. When the UDP channel fails to initialize (common in tunneled/VPN environments where the underlying Netty Bootstrap cannot bind properly), the field remains null and the handler crashes instead of falling back gracefully.

Workaround

Setting disable_udp_pipeline = true in sable-common.toml prevents the crash, but this disables UDP entirely for the affected client.

Suggested fix

Add a null guard at the top of ClientboundSableUDPActivationPacket.handle():

if (channel == null) {
    LOGGER.warn("UDP channel is not available, skipping UDP activation");
    return;
}

This allows the sub-level data to fall back to TCP without crashing the client.

Related

• I can't log in to the server; I'm getting an "invalid player data" error #175 — similar NPE but server-side (SableUDPServer.getServer() returning null)