Uploaded an example world. It uses a flushing mob farm that's partially outside simulation distance instead, but it just shows how easy it is for farms to accidentally create a build-up of pending ticks.
You may need to afk for a bit.
Turns out minecraft didn't export the world properly.
(note: different types of components are powered in alternating gameticks, it may be useful to know to understand why it happens)
This seems hard to fix without major changes to redstone.
The issue isn't with the repeater. The issue is that when a world change happens, the components activated in that tick (often called consumers) aren't instantly powered/depowered.
You can either:
remove the delay and make chains of pistons instantly power/depower each other
add extra delay to repeaters from world changes (and other components known as producers)
fix it only for player interactions
The first two would break most redstone builds. The third wouldn't, but doesn't really fix the bug fully.
May I suggest that this is marked as "Won't fix"?
You did it so fast that the gravel couldn't fall.
Having a bit of trouble uploading the world. I'll try again later. Hopefully the issue will go away.
ok
I've added a test world demonstrating these issues.
Not properly fixed.
Note: this doesn't happen when using something like buttons, since they break using pending ticks and you don't get the chained instant updates, meaning that this is definitely not related to lag.
Also, last time I checked, blocks were updated in the following order from instant updates:
x- x+ z+ z- y+ y-
This might be useful while testing stuff.
Something similar applies to honey blocks. Their top surface is slightly higher up so that mobs can pathfind on them.
This is partly fixed in 1.16.20. It's still a little buggy though.
Probably because of the new despawning. There are way more mobs than usual because of it.
I'm 100% sure that there were changes. I slowed down the game using a lag machine and in 1.16.0 it launched the player when the slime block finished moving (but not when it started moving), plus the moving slime block wasn't bouncy and you could take fall damage. In 1.16.20.50 the opposite is true. This means players are now bounced 100% of the time (upwards only though).
Slime blocks should definitely still bounce you when they finish moving.
The desync thing does take into account the hitbox of pressure plates. The player slides on top of them. Client-side you move only 1 block, so something is definitely wrong. And pressure plates activating when you're far from them is definitely not normal.
Also, moving blocks will be way better once the bug is fixed client-side. Here are some reasons why they will be better once it's fixed client-side:
they take into account the hitbox of the block being moved
you no longer take fall damage from landing on moving slime blocks
you can stand on flying machines without falling through
you will be able to stand on slabs (and any moving block really) and still be moved along without falling through
The piston cuts off the redstone signal before the game can display it.
still affects 1.16.20.50
Also, it appears moving slime blocks are bouncy now, which is great.
Mojang, if you are planning on removing this, don't. Instead, make the tridents take damage from being used in trident killers. That way people have to make a drowned farm or kill drowneds if they want to use a trident killer, which is way more interesting.
Ok it seems like slime blocks no longer launch players in certain circumstances as seen in this video <removed due to auto play>
It appears that it won't launch them when they're more than 0.5 blocks above it when it starts pushing. I assume that slime blocks no longer launch players once they're finished moving (probably as a direct result of the fix in 1.16.20.50).
Moving blocks haven't been fixed client-side in the latest beta (1.16.20.50). Slime blocks still launch players to the side if they're standing near the edge (however this is less noticable). They still have a collision box of a full block.
Something did change however. Moving blocks now launch the player as soon as they start moving, instead of as soon as they finish moving. This is nice, but moving blocks are still kinda buggy client-side, which is arguably the most noticable.
Examples of problems that still exist:
Client-server desyncs, for both players and items. Items desync permanently, whereas players desync temporarily and then "correct" the server and end up on the wrong spot
Buggy movement when players are standing near the back of a block being pulled. The player will fall, which they wouldn't do server-side.
Players can get pushed to the side when pushed up by slime blocks (as mentioned before) or regular blocks
Players are not always bounced to the side properly (they are bounced server-side though, but the client corrects the server soon enough). You can test this by bouncing players to the side with slime blocks across ice with pressure plates on top.
Really looking forward to this being properly fixed once and for all.
This is because of random piston update order. The piston trying to pull the slime back is blocked from doing so because the other piston is still extended (50% of the time). Remove the lower piston and it should work. Random piston update order is WAI, so I suspect this may be WAI too.