mojira.dev
MCPE-21038

Spreading speed of water is not always faster than that of lava

Spreading speed of water should be guaranteed to be faster than that of lava in the Overworld. However in the current implementation, water is only statistically faster and has a small chance of getting overtaken by lava.

This is usually not a problem and could even be WAI but for smooth stone generators the difference of speed is crucial.

Steps to reproduce:

  1. Open the attached test world.

  2. Push the "reset" button to initialize the state of the contraption.

  3. Grab some coffee while looking at the machine absent-mindedly. It may take several tens of minutes to pose the problem.

What I expect to happen:
The machine keeps generating stone blocks indefinitely.

What actually happens:
Flowing lava occasionally lands on the air block in front of the piston, before the water stream pours into the said space. When this happens water ends up flowing into the lava stream horizontally, which does not generate any blocks as expected, and finally the lava stream tries to flow horizontally into water, generating a cobblestone block blocking the water stream and the machine is dead.

Linked issues

Attachments

Comments 9

[Mojang] Mega_Spud (Jay)

Cleaning up old tickets: This ticket has not been updated recently (1 year+) so is being closed as Cannot Reproduce. If you feel this is still a valid issue then please comment, or create a new ticket following the Issue Guidelines.

Quick Links:
📓 Issue Guidelines – 💬 Community Support – 📧 Feedback – 📖 Game Wiki

[Mojang] Mega_Spud (Jay)

Reopening as this does still seem to be an issue some of the time.

Affects Minecraft Bedrock 1.12.1

I've been able to reliable reproduce this in 1.14.30, but only with lava dropping in to water.

Just break the stone repeatedly and in just a few tries the lava will update before the water and run down without converting neither to stone nor cobblestone

[media]

I have reproduced this in the test world in 1.16.20. It took over half an hour for the machine to break, and it may have only broken while I was switching windows to put Minecraft in the background so I could update other tickets.

[Mojang] Mega_Spud (Jay)

Thank you for your report!
However, this issue has been temporarily closed as Awaiting Response

Is this still an issue in the latest version? If so, please make sure the ticket description contains the following information:

Steps to Reproduce:
1.
2.
3.

Observed Results:
(Briefly describe what happens)

Expected Results:
(Briefly describe what should happen)

If your ticket does not look like the example given here, then it's likely to be closed as incomplete.

This ticket will automatically reopen when you reply.

Quick Links:
📓 Issue Guidelines – 💬 Mojang Support – 📧 Suggestions – 📖 Minecraft Wiki

just tried to make a stone generator in Window bedrock edition and I'm having this exact same problem, lava randomly flows faster than water when you break multiple blocks in rapid succession...

 

This report is currently missing crucial information. Please take a look at the other comments to find out what we are looking for.
If you added the required information and a moderator sees your comment, they will reopen and update the report. However, if you think your update to this report has been overlooked or you want to make sure that this report is reopened, you can contact the Mojira staff on Discord or Reddit.
-- I am a bot. This action was performed automatically! If you think it was incorrect, please notify us on Discord or Reddit

I have been running the machine in the test world for a half hour or more in 1.20.10.21 Preview and have not reproduced the issue with the preset clock circuit timings. However, the issue does sometimes reproduce if you adjust the repeater delays in the clock circuit while the machine is running. Therefore, I am not sure if the originally reported "statistical" random variance in spreading speed was fixed at some point, or, alternatively, if the description overlooked user input. It is possible that the original random variance was due to random ticks of lava being treated the same way as the block updates described below, and at some point the lava random tick logic was refactored but that refactoring was not applied to block update lava ticks.

In the current release (1.19.83) and Preview (1.20.10.21) there is no randomness in lava and water spread speed: lava always attempts to spread 30 game ticks after receiving a block update from an adjacent block, and water attempts spread 5 game ticks after receiving a block update from an adjacent block.

What causes unexpected results for players is that the delay does not get reset by subsequent block updates. This means that lava can appear to spread faster than water if you time block updates carefully. Thus, the unexpected behavior described in the comment by Twilac can be reproduced by breaking the stone 26-29 ticks after it forms (or just under 1.5 seconds):

[media]



It is even possible to create the appearance of instant spread with precise delay:

[media]

This behavior is part of the underlying cause of MCPE-48128 / MCPE-73342. The issues described in those reports also rely on an additional factor: in some cases it seems that the delayed block updates stack. By "stack" I mean that the liquid "remembers" each update it receives and makes an attempt to spread at the expected delay from each block update. As a result, the liquid can appear to spread faster if it has been repeatedly updated for a period time before it is actually able to spread. This "warm up" can be done by a fast clock circuit updating blocks as described in MCPE-73342, or just by mining and generating blocks as in shown in the video attached to MCPE-48128.

When considering whether or how to patch this issue, please keep in mind

  • The ability to stack liquid spread updates is used to increase rates in stone, cobblestone, and moss farms. A basic stone generator is 2.5x faster using stacked updates.

  • The ability to stack fire spread updates is used to run portal ticking gold farms. Ticking a portal with stacked updates is about 2x faster than using lava spam and random ticks.

  • Full Java parity. Stacking liquid updates/pending ticks does not seem to occur in Java Edition, but it would be bad for the community to nerf the above-mentioned farms in Bedrock without considering whether Bedrock players will be able to switch to Java designs or have comparable alternatives after the patch.

PHO

(Unassigned)

236936

Confirmed

Multiple

lava, water

1.20.10.21 Preview, 1.0.5.54, 1.2.1.1, 1.11.4, 1.13.0.2 Beta, ..., 1.16.20, 1.19.83 Hotfix, 1.20.0, 1.20.62 Hotfix, 1.21.51 Hotfix

Retrieved