mojira.dev
MCPE-244378

Hold-to-place bridging: client stops placing, server keeps placing where the crosshair points

Short description:

While jump-bridging with the Place button held, block placement sometimes enters a state where the client refuses every placement (the first-person hand shows no placing animation), but the held Place input keeps sending placements and the server places blocks wherever the crosshair points. The hold-to-place direction lock is lost: blocks appear in front of the player's face, 2-3 in quick succession, and the player gets pushed sideways. It lasts a few seconds and then fixes itself. It does not happen on 26.42 under the same conditions.

Disclosure: this was observed with a client-side mod loaded through LeviLauncher. The measurements below come from read-only hooks (return values were only logged, never changed). The mod does not modify inventory, GameMode or block placement. It has not been reproduced yet on an unmodified 26.63 install.

Steps to reproduce:

1. Join a featured server (The Hive, BedWars) on 26.63 (NetherNet connection, about 100 ms ping).

2. Hold wool, sprint and jump-bridge while holding the Place button.

3. Keep playing normally; after a while the issue starts and lasts a few seconds.

Expected result:

Same as 26.42: the bridge continues in the locked direction and the placing animation plays.

Actual result:

- No placing animation; the held block stays still in the hand.

- Holding Place puts blocks where the crosshair points, often in front of the player's face, instead of continuing the bridge.

- Switching hotbar slots and back does not fix it.

Technical findings (logged on 26.63/26.65):

- In the bugged state, GameMode::startBuildBlock and GameMode::buildBlock return false on every attempt while the player holds a placeable block (e.g. yellow_wool x16). GameMode::continueBuildBlock keeps being called on the crosshair target, about 15 times per tick.

- The server still places those blocks: an UpdateBlockPacket with the block arrives about 100-130 ms later at the targeted position, and the held stack drops from 16 to 11 with zero client-side placements.

- Outside that state, client placement succeeds and the server confirms each block 100-130 ms later.

- Hold-to-place otherwise behaves the same on 26.42 and 26.63 (blocks per click, repeat interval); the per-tick continueBuildBlock calls are about 10 on 26.42 vs about 15 on 26.63 (.65).

- Hold-to-place logic changed in 26.60 (snappable placement) and block selection changed in 26.63 , which is this code path.

A video of the issue is attached.2

Attachments

Comments 2

Thank you for helping us improve Minecraft! We saved your files:

[media]

Please edit your report to change the Affected Version to the version shown on the Minecraft title screen

活TH£PHĄND£R死

(Unassigned)

Unconfirmed

Windows

26.63 Hotfix

Retrieved