Server character PivotTo() inconsistencies

[ZNoob_x (Lucille)]

As the title suggests, there are some rare/occasional inconsistencies using the character PivotTo() on the server. Simply put, sometimes the character just does not pivot or pivots to the wrong location, which is a huge issue for my game. I was wondering if there is a direct fix to this. I have NOT been able to replicate any of the bugs that involve this error ,and it seems to be completely random.

[Downrest (Down)]

Is there any way that the pivot of your characters are incorrectly changed as you run :PivotTo()? Or maybe physics is somehow incorrectly interacting with your :PivotTo()?

[ZNoob_x (Lucille)]

No, pivots are correctly scheduled appropriately. There are also checks to ensure the pivots succeed but rare cases the bugs somehow bypass these checks.

[benjacara2]

provide proof please

or even better, some recorded context

edit:
i cant rlly tell you much without context okay… I MAY help with context ofc

[Ilucere]

pivotTo is not reliable and I recommend against it

Custom Shift Lock Causing Player to Slide

I’m trying to recreate shift lock, and it seems like pivoting the pivoting the character directly causes it to slide on the ground. So far all the information I’ve found seems to be about custom rigs, I’m using the base r6 rig. This is the code I’m using. if not shiftLockEnabled then human.AutoRotate = true return end human.AutoRotate = false local hPos = hrp.Position local x,y,z = camera.CFrame:ToOrientation() local shiftCF = CFrame.new(hPos) * CFrame.Angles(0,y,0) * CFrame.new(0,1.5,0)…

1. use setprimarypartcframe, which is ‘deprecated’ (bs)
2. use cframe with parts welded/connected to humanoidrootpart (this is best for your situation imo)

edit: oh shoot

Simply put, sometimes the character just does not pivot or pivots to the wrong location, which is a huge issue for my game. I was wondering if there is a direct fix to this. I have NOT been able to replicate any of the bugs that involve this error ,and it seems to be completely random.

probably a stretch, but sounds like you’re setting it at a constant rate? in a runservice loop? and you said there’s random spikes of the issue? then maybe you aren’t utilizing deltatime properly

https://www.construct.net/en/blogs/ashleys-blog-2/using-lerp-delta-time-924
a = lerp(a, b, dt) suffices (no need for a = lerp(a, b, 1 - f ^ dt))

eedit: start = lerp(start, goal, speed * 60 frames_per_second * dt seconds_per_frame

[EzraPalo (ezra)]

When i want to move the player i just move its humanoid root part and it works

[Downrest (Down)]

Curious, why are you against PivotTo()?

[Fietowski (Kira)]

Only reason I could think of would be that the RootPart is basically the center of the character, PivotTo() will use the Character’s PrimaryPart, which I believe is the head

Edit: I just rechecked and the PrimaryPart is the RootPart in R15 (Head in R6)

[Ilucere]

i linked an example
sure it works most the time
but i find setprimarypartcframe or .cframe to work perfectly

[Ilucere]

no thats not the reason you can easily just change the primarypart

[ZNoob_x (Lucille)]

Context is I’m using PivotTo() on a player character model >> RARE cases where players reported not teleporting properly.

[ZNoob_x (Lucille)]

@Ilucere

ate? in a runservice loop? and you said there’s random spikes of the issue? then maybe you aren’t utilizing deltatime properly

The pivot is run once, then there’s a check for correct pivot with a tolerance of 5 studs, it continues trying to pivot in 0.2s intervals + frame delay up to 15 times (~3 seconds).

[ZNoob_x (Lucille)]

After diving deeper into the problem and receiving some more information from players, it seems as if the bug is being caused by a huge lag spike right at the moment of teleports, causing the client to override the player’s position on reconnect (my theory). Pivot problems are completely RANDOM with no obvious pattern to them, however from 2/3 reports, this is all I can go off from.

[Ilucere]

oh man set the position of the player on server and client

-- server
playerInfo.currentValidPosition = newPosition
---- this playerInfo.currentValidPosition is changed only if the server is manually teleporting 
---- the player or if the change in the clients position between two 'ticks' (maybe 0.2s) is valid
---- via raycasting and distance checking (on server)

---- you can also uncomment the line below if it doesnt end up being choppy.
---- shouldn't though, because server position is replicated to the client almost the exact time they recieve the clientSetPosition remote
-- playerCharacter.HumanoidRootPart.Position = pos

clientSetPosition:FireClient(player, newPosition)

-- client
clientSetPosition.OnClientEvent:Connect(function(pos)
    localCharacter.HumanoidRootPart.Position = pos
end)

theres a few things in this code

• there is server authority
• i do not use pivotTo, i absolutely do not trust it
• you can set the position twice (server + client) for guarantee and to eliminate the bug. but rationally you would only do server

typically you would’nt even need the remote- setting position on server usually suffices. in your case it apparently doesn’t.

@ZNoob_x

Pivot problems are completely RANDOM with no obvious pattern to them, however from 2/3 reports, this is all I can go off from.

this is not normal.

memory leak? heavy computation?
go through your game and wrap every computationally-heavy function (loops, and wtver else) with debug.profilebegin(title) and debug.profileend()
it doesn’t seem like a memory leak- memory leaks are usually gradual decreases in frames. patterns sounds like its an issue with one of your games functions. or maybe map idk maybe theres a bunch of parts in 1 spot.

if you send me a copy of your games client-side i can go through it and find the bug for you.

@ZNoob_x

The pivot is run once, then there’s a check for correct pivot with a tolerance of 5 studs, it continues trying to pivot in 0.2s intervals + frame delay up to 15 times (~3 seconds).

also this is not a good band-aid at all

---

in sum, don’t use pivotTo, and also try setting position on client w/ server authority

[ZNoob_x (Lucille)]

@Ilucere

memory leak? heavy computation?
go through your game and wrap every computationally-heavy function (loops, and wtver else) with debug.profilebegin(title) and debug.profileend()
it doesn’t seem like a memory leak- memory leaks are usually gradual decreases in frames. patterns sounds like its an issue with one of your games functions. or maybe map idk maybe theres a bunch of parts in 1 spot.

There’s no memory leak or heavy computation in the game. I’ve put in a lot of effort to optimize the game as low-level as possible (btw 2/3 meant 2 or 3, not 2-thirds of all reports). I am going through with client verification, too. For any teleports, I will just be sending a client teleport request through my remote manager. Thanks!

[Ilucere]

hows it going? is the issue resolved? do u have a video of the issue

[ZNoob_x (Lucille)]

I have the functionality set up but I haven’t rolled it out to live servers yet. There are a couple of other bugs that I’ve delegated to testing. No additional video of the issue or reports. I have other work/projects, so I will update the servers on the next rolling update patch!