Just place a slab of any kind on the upper side of a block many times, and you'll sometime get one that converts to being on the bottom side of a block.
Description from MC-89850
When placing top- or bottom aligned blocks like slabs and stairs server and client disagree about the alignment of placed blocks
Doesn't happen in single player (or isn't visible), but on dedicated server or Realms.
Steps to reproduce:
Place a block with a well defined center (e.g. stone bricks) at ca. eye level
Take a stair or slab
Aim slightly above the center of the block
Place block
The results:
Client places the block at the correct top alignment
Server overwrites the block a split second later at incorrect bottom alignment
Screenshots show before and after placement (the in-between position is quite difficult to capture)
Code analysis by @unknown in this comment.
Linked issues
is duplicated by 12
relates to 1
Attachments
Comments 26
Yes, it has happened when I placed it aside another half-slab that was placed correctly.
It happened to myself and it keep happening to pretty much every youtuber I'm watching.
Confirmed in 13w03a on multiplayer. Aiming at the very bottom of an upper slab like in the screenshot places the upcoming slab as lower slab.
Edit: Works fine in singleplayer.
Agh, this is so annoying for the non-host when bridging with half-slabs in LAN play.
edit: Version 1.6.2 pre-release.
I came across this issue when looking into MC-4 and found that it basically has the same cause.
When you place a block the packet CPacketPlayerTryUseItemOnBlock
is sent to the server. It encodes the facing into bytes to transmit them which causes this inaccuracy.
The packet is just send when the player right clicks on a block holding an item, so it probably isn't an issue to just change the transmission of these values to floats.
It would probably be good to add this to the description 🙂
Wrongly marked. This is a duplicate of MC-8
@@unknown MC-8 can be a duplicate of this report because this one was created with more details, details that allowed the devs to actually fix the issue. Just because MC-8 was created first doesn't mean it's the best dupe-target for similar issues.
In some cases the content from the newer report will be moved to the older report if the fact that there is an older issue that is the same thing is found, but (especially if there are a lot of linked issues on the newer one and the older one is basically abandoned for 3 years) it won't usually be changed. It doesn't matter, anyways - the issue number is insignificant for most purposes. If having the smallest issue ID mattered, then one could just constantly replace old dupes with other unrelated issues - not necessary and generally a waste of time.
And you're absolutely sure the point of the block you were pointing at didn't change slightly while you were moving so you pointed at the lower half of the block?