server crash #1458

Context

Client crashes with a NullPointerException every single time when connecting to our dedicated NeoForge server, but the same client mod list works fine in singleplayer. The crash is 100% reproducible on multiplayer connect and has persisted through extensive troubleshooting (details below).

Crash Report

java.lang.NullPointerException: Cannot invoke "dev.ryanhcode.sable.mixinterface.entity.entity_sublevel_collision.EntityMovementExtension.sable$getTrackingSubLevel()" because "entity" is null
	at TRANSFORMER/sable@2.0.3/dev.ryanhcode.sable.ActiveSableCompanion.getTrackingSubLevel(ActiveSableCompanion.java:435) ~[sable-neoforge-1.21.1-2.0.3.jar]
	at TRANSFORMER/minecraft@1.21.1/net.minecraft.client.renderer.GameRenderer.handler$fpo000$sable$renderLevel(GameRenderer.java:13078)
	at TRANSFORMER/minecraft@1.21.1/net.minecraft.client.renderer.GameRenderer.renderLevel(GameRenderer.java)
	at TRANSFORMER/minecraft@1.21.1/net.minecraft.client.renderer.GameRenderer.mixinextras$bridge$renderLevel$321(GameRenderer.java)
	...

Full crash reports available on request (multiple identical crashes captured across different sessions, same stack trace and line numbers each time).

Environment

• Minecraft: 1.21.1
• NeoForge: 21.1.248
• Sable: 2.0.3 (neoforge)
• Sable Companion: 1.6.0
• Create: 6.0.10
• Create Aeronautics (bundled): 1.3.0 (includes Create Offroad 1.3.0, Create Simulated 1.3.0)
• CC: Sable: 1.3.4
• Java: 21.0.9 (Azul)
• OS: Windows 11
• Launcher: Modrinth App

Steps to Reproduce

1. Join the dedicated multiplayer server with the modpack installed.
2. Client crashes immediately during the first renderLevel call after connecting.
3. Crash is 100% reproducible — every connection attempt fails the same way.

What we've already ruled out

We've done extensive isolation testing before filing this report:

Additional context that may be relevant

Earlier in our troubleshooting, before this crash appeared, we saw repeated console log spam on the server:

[Netty Epoll Server IO #0/ERROR] [de.ry.sa.Sable/]: SableUDPPacket SNAPSHOT (ClientboundSableSnapshotDualPacket) was larger than expected, found 27 bytes extra

This suggested a UDP snapshot packet decode issue. We hypothesized this might be related to the current NPE crash, which is why we tried disabling the UDP pipeline entirely — but that had no effect, so the two issues may be unrelated (or the corruption happens somewhere UDP-disable doesn't reach).

We also separately hit an unrelated crash caused by Create: Vibrant Vaults (ItemVaultItem.place NPE on placing a specific block variant) which we've since worked around by removing that mod — this is likely a separate bug in Vibrant Vaults itself, not Sable, but mentioning for completeness since it was in the same troubleshooting session.

Working hypothesis

Since the crash is independent of the connecting player's own position/data, we suspect the null entity may belong to a different entity being rendered near spawn/player location — possibly an active Create Aeronautics contraption (airship, vehicle, etc.) with a corrupted or incomplete Sable sub-level tracking reference. We have not yet been able to test removing/landing all active contraptions to confirm this, but wanted to file this report with the full troubleshooting history in case it's useful regardless.

Expected behavior

Sable's renderLevel mixin should null-check entity before calling sable$getTrackingSubLevel() on it, similar to the pattern seen in other reported NPEs in Sable's mixins (e.g. #675, #1080).

crash-2026-08-09_16.47.49-server.txt

Crash Report

https://mclo.gs/yX7hXM4