plugins-dev-chat

[celest thorn]

You can use that?

It WASN'T MY FAULT

[upper vapor]

i made it so you can

[celest thorn]

THEY HANDED ME THIS and told me it was something else

then i got really drunk fast after just a drink

and that night was the worst one

i started laughing so hard for no reason

i totally don't remember anything other than 2 stuff

[upper vapor]

bruhhhhh

[celest thorn]

1. me changing to my pijama

2. me telling my friend im trans

lol

[upper vapor]

[celest thorn]

didn't end well

do i care what he said, nope

[upper vapor]

some people can't grasp the existence of bisexual people

so

don't be surprised if they don't get gender identity

[celest thorn]

nah he knows

trust me

[upper vapor]

i fought a freaking war in youtube comments

[upper vapor]

https://tenor.com/view/rock-agreed-rock-sus-agreed-gif-23818297

[celest thorn]

he knows because his best childhood friend is gay

[upper vapor]

[celest thorn]

crazy part he told me "yea don't do it, you know you would be the worst example"

do i care? no

[upper vapor]

don't do what

.?

[celest thorn]

transition

[upper vapor]

right

[celest thorn]

or starting hrt

even tho he knows what i went through

[random scaffold]

it exists

lol

[upper vapor]

good luck :3

[celest thorn]

HumanRole is base game stuff

[random scaffold]

then why need Exiled.CustomRoles

[celest thorn]

idk

[upper vapor]

i extend the base-game HumanRole type

[celest thorn]

i didn't know HumanRole was real and could be used only on server

[random scaffold]

and why i made #1371740029449146389

bruh

[upper vapor]

as long as what's written to the network is consistent with base-game, you're fine

[upper vapor]

why is the Interfaces namespace not in the API namespace

also naming it LabApi.CustomRoles is uhh

kinda makes it seem like it's official

[random scaffold]

used official framework

xd

[upper vapor]

but the plugin itself isn't

oh well

[random scaffold]

in anyway labapi dont will be add custom roles

so its normal

[random scaffold]

that easier access

[upper vapor]

you have one ICustomRole interface
and a CustomRole abstract class split up into 10+ parts

[random scaffold]

designed for possible expansion

[upper vapor]

putting every 3 members is

why

😭

if you already have those then why // ------------ PUBLIC INSTANCE METHODS

[random scaffold]

?

[upper vapor]

you could use #region

but

you already have a million parts

[random scaffold]

i dont love 3000 lines in one file

but instead of a comment, it would probably be better

[upper vapor]

that's fair, one shouldn't be 3K lines

but navigating this file tree also isn't very dev-friendly

it's your project though, just giving an opinion

maybe i should't be the one speaking
-# ternary operators are really cool :3

[random scaffold]

[grand flower]

Regions are the devil

If you need em you screwed up somewhere

[upper vapor]

why

collapsible sections in your IDE

[grand flower]

Just an organizational nightmare

[worthy rune]

@random scaffold https://github.com/northwood-studios/LabAPI/pull/139/files we might offer a type enum in the future thats not based of the name(for both door and room)

[grand flower]

IDE already allows you to collapse methods

Add a healthy amount of comments and it doesn't matter if you have a huge class

Assuming you can't split it up, I usually split up code into components to place on a gameobject if I can (and it makes sense)

For stuff like a base class, if it makes sense for it to be one class, doesn't matter if it's 500 or 5000 lines long as long as it's well documented

[worthy rune]

when ever i need to find something is a class i just use this, its all in alphabetical order so its way quicker than scrolling over the entier file

[grand flower]

yeah

[upper vapor]

i wanna collapse regions

[grand flower]

I pay for Rider, I'm using all of Rider

I mean hey if you wanna use regions feel free to obviously, but it's just that my mentor and a few other people at the time considered them clutter and I kinda see it

[hearty shard]

ive never once used this 😭

[terse bone]

idk, i use ctrl + ,

[upper vapor]

structure view can be goated

double shift

[upper vapor]

https://github.com/Axwabo/SCPSL-Helpers/blob/main/Axwabo.Helpers/PlayerInfo/PlayerInfoBase.cs
this one for example
i can hide the obtainer stuff in my IDE

or the properties

[grand flower]

I don't see the point of hiding stuff because I generally know what I'm looking for and the rest won't bother me

[upper vapor]

less distractions

[grand flower]

I also have a very specific order for class members

so I know exactly where I would've placed something

[upper vapor]

yeah code arrangement is nice

i don't use it in Rider cuz i change my mind sometimes lol

[terse bone]

i've got tortured with StyleCop couple years ago and from that time i always write clean code

[upper vapor]

define "clean code"

https://tenor.com/view/running-home-bugs-bunny-jump-hiding-gif-8509983

[true cedar]

i wish stylecop enforced more rules tbh

i have to do it in my ide

things like file scoped namespaces, ordering of methods

[upper vapor]

you can't make it handle file-scoped namespaces???????????

[true cedar]

well you can't make it enforce one way or another

anyways

if anyone cares i figured out how to bypass the hint size limit

[upper vapor]

let me guess
TextHintParameter

multiple

[true cedar]

string text = new('m', ushort.MaxValue - 10);
player.SendHint("{0}{1}", new Hints.HintParameter[] { new StringHintParameter(text), new StringHintParameter(text) }, duration: 10);

yea lol

[upper vapor]

that

[terse bone]

ig placement of all class members based on modifiers, naming conventions, hardly ever use var etc.

[true cedar]

idk if this is like new knowledge

prob not

[upper vapor]

um

so

var makes sense in contexts where the type is concrete

[true cedar]

var is great for linq and complicated type bullshit

[upper vapor]

clean code is opinionated

[upper vapor]

ye

[grand flower]

I always use var

[terse bone]

i only use it for complicated types

[true cedar]

it depends for me

[terse bone]

but that's just me, and i keep my code as simplest as possible

[upper vapor]

not writing a 100+ line method
minimizing indentation where possible

[true cedar]

i only rarely use var bc its kind of frowned upon in c# but in rust you use type inference as much as possible

[true cedar]

"minimizing indentation"?

[upper vapor]

guard clauses

[true cedar]

oh yea

[terse bone]

ah, never nesting

[upper vapor]

instead of if in if in if in if in if in if

in for

in foreach

in try

in using

gosh, people should know about block-scoped usings

[true cedar]

i only use guard classes for simple shit like parameter validation

[upper vapor]

shh it's guard clause

[true cedar]

clauses

whatever!

[upper vapor]

basically

[true cedar]

like using x = y;?

or

[upper vapor]

i meant

using var stream = File.OpenRead

instad of using (var stream = File.OpenRead()) {}

[true cedar]

yea thats what im talking about

i only rarely see people using idisposable so

[grand flower]

try/catch in game dev ouch

[upper vapor]

the latter has some use cases
but most of the time you need the variable in the scope

[grand flower]

anytime I see one I throw it out

[upper vapor]

https://tenor.com/view/robertdowneyjr-wide-eye-tony-stark-iron-man-avengers-gif-7766118

[still thistle]

a

[true cedar]

dont look at how i implemented nuke radiation

[upper vapor]

you don't use try-catch anywhere?

[grand flower]

nope, if you're using it you can probably just fix your logic so it doesn't throw

[still thistle]

i have 51,2 hours on scpsl

[true cedar]

okay

[true cedar]

do u guys add a null check like everywhere

[upper vapor]

you still have to catch system exceptions though

[true cedar]

shoudl i do that

thats what the std lib does

[grand flower]

part of the try/catches I threw out

[true cedar]

yeah so

it would randomly throw errors

nres i should clarify

and break the entire coroutine

[grand flower]

doesn't anymore, but I did go through like, the entire codebase to add some null/destroyed object checks

[upper vapor]

unity lifetimes 💞

you can use the destroyCancellationToken now

[true cedar]

it was a coroutine that was running 24/7

which admittedly was not the best idea in the whole world but

[upper vapor]

in your own code, yeah

but there are cases where you need a try-catch

[grand flower]

What kind of cases

Anything Unity related you don't need to

[terse bone]

OS related stuff

[upper vapor]

yeah I/O

json parsing

yml

[true cedar]

it still pisses me off that c# doesn't have an elegant way to automatically throw nres if something is null

[upper vapor]

sad

could add it yourself

[true cedar]

well you have the argument null exception

like it exists but its

annoying to write the ifs

[upper vapor]

yeah

so you can make an extension method

kinda like j*va's Objects.requireNonNull

[grand flower]

I guess

[upper vapor]

i mean

[grand flower]

for one offs like that

[upper vapor]

if you dereference the null

NRE

[grand flower]

but if you have try/catch in gameplay code, you probably screwed up somewhere

[upper vapor]

true

[true cedar]

i like the rust approach where nres don't exist (for the most part)

[upper vapor]

TrySomething(..., out var result) everywhere

[grand flower]

pretty much

[upper vapor]

[grand flower]

null/validity checks

and you're good to go

make a Player.IsValid() extension method and it's all i need

[upper vapor]

IsValid as in..?

IsOnline?

there's an IsOffline prop but it uses the gameobject which throws...

[worthy rune]

isnt there a IsDestroyed prop

[grand flower]

not on players no

[grand flower]

hub is valid & gameobject is valid

wtf

well

[upper vapor]

yeah uh

that

[grand flower]

still would be needed since GameObject doesn't check the hub is valid before accessing

[true cedar]

you only have Option<T>s

[upper vapor]

lol

[grand flower]

I made a PR for that though

[upper vapor]

yup

saw

[grand flower]

yeah idk what kind of naming choices were made here but like

bruh

[upper vapor]

get merged pls

[true cedar]

no the fuck they didn't

did they seriously add the dumb offline api from nwapi

[upper vapor]

i don't think Player.cs was written by one person

so they didn't check

there are a lot of inconsistencies

[grand flower]

return player != null && player.ReferenceHub && player.ReferenceHub.gameObject;

IsValid

[upper vapor]

usages of null conditionals on unity objects

[grand flower]

yeah I made that mistake and was going insane as to why I was still getting NREs

fuck Unity

[true cedar]

god i hate the if (player.ReferenceHub) thing idk why

[grand flower]

WDYM NULL PROPAGATION DOESN'T WORK AAAAAAAAAAAAAAAAAA

[upper vapor]

:3333333

[grand flower]

idk you get used to it ig

I'm used to it because of Unreal since everything's a pointer anyway hah

if (IsValid(SomeObject)) == if (SomeObject) for unity's monobehaviours

[true cedar]

oh ik

[worthy rune]

is the gameObject check necessary? since if the component is valid the gameobject would be too

[upper vapor]

not sure why you'd check if the gameobject is null

riptide was faster than me

[grand flower]

Potentially not necessary but I might as well throw it in there to be certain

oh

right

I see now

yeah I can get rid of it heh

[worthy rune]

only brining it up because in a future PR i change all the IsDestroy method in LabAPI from Base == null || GameObject == null to just Base == null

[grand flower]

smart

[upper vapor]

🙏

[grand flower]

fixed

[worthy rune]

also about those weird properties in Player, ill look into removing/obsoleting them

[grand flower]

I'd probably just rename it to Player.IsDestroyed

to be in line with all of the other wrappers

[worthy rune]

i agree, thats should be coming soon

[grand flower]

o7

sweet

[true cedar]

are any of u good at design i cant figure out how i want to do this shit

[grand flower]

what kind of design

[upper vapor]

:plot:

wdym

[true cedar]

i have a class that is intended to be readonly (Element), and its supposed to let you have hintparameters
but you're not supposed to be able to modify or change the hint parameters once you add them (ill make a wrapper for each parameter to ensure that)
problem is how the fuck do i make that happen? i suppose i could just make an Element.AddParameter(Parameter) but that feels a bit awkward

actually the latter idea wouldn't be awful

[upper vapor]

so

the uh

[true cedar]

the uh what

[upper vapor]

you have an elements list

and you wanna "seal it"?

[true cedar]

nvm i explained it poorly

i think i know what i'm gonna do anyways

maybe

[true cedar]

basically:
every player has a Display, this display contains a list of Elements - each element provides a string to be combined into a final hint (so its like an individual hint)
but to prevent weird behavior elements are, for the most part, readonly - if you want to update the content of one you need to create a new element
the problem is thus how do i let you add parameters to an element if its supposed to be readonly?

so you can't retroactively add it

if i make it in take in a list, you could update the list

[upper vapor]

create an ElementList class
only expose a ReadOnlyList of the elements
and make the constructors create a new list

you can implement IReadOnlyList or just wrap it

or IReadOnlyCollecction

[true cedar]

yea thats what i was thinking but

not a big fan of creating a new list every time

if only we had rust ownership

[upper vapor]

well

you can add an IEnumerable overload

[true cedar]

wdym

[upper vapor]

for the constructor

[true cedar]

ig

but no

[upper vapor]

a constructor with 1 element param, 2 element params

[true cedar]

because you could then pass in a list and modify the list

which the ienumerable would reflect

[upper vapor]

you .ToList it

[true cedar]

OH you mean still make the list

[upper vapor]

yea

[true cedar]

also the elements store the parameters

[upper vapor]

make those immutable

maybe use records

[true cedar]

im using init only setters for the most part

[upper vapor]

that's good too

[grand flower]

next up on my optimization tests, gonna modify the server position distributor to do work round robin

odd players one frame, even players the next

[random scaffold]

why not better just rename

[grand flower]

PITA to have to update large codebases

[worthy rune]

because thats enum is for the name

it doesnt fit well for type, and alot of rooms by default dont have one assigned so imo its abit anoying

[upper vapor]

also maybe don't rename the whole type

like they did with RoleType...

[random scaffold]

just add .None

[worthy rune]

yeah renaming the enum type is a breaking change so kinda pointless to put a obselete on the prop

[worthy rune]

?

[random scaffold]

most of the humans dont need type for each room

only like hcz049, gr18, scp127 and other

[worthy rune]

did you read what i said earlier

[buoyant oyster]

is there an event that fires whenever a player damages another player? I can't find it

[worthy rune]

PlayerHurt event

[hearty shard]

Hurt

[buoyant oyster]

was looking under PlayerDamage

[hearty shard]

or Hurting if you want before damage happens

[buoyant oyster]

ty

[grand flower]

wonder how much of a performance boost this would provide

[unique crane]

Feel free to share any results

[grand flower]

will do

Taking inspiration from Fortnite, they do the work round robin for player states because of how expensive doing updates for 100 players is

[true cedar]

what do you mean "round robin"

[true cedar]

oh

ngl i dont know how much you'd gain from that

esp since patching a method makes it slower naturally

at least calling it

[grand flower]

can't be slower than like

the fpcserverpositiondistributor atm anyway

Already implemented caching for RelativePosition in FpcServerPositionDistributor.LateUpdate and it did give me some nice gains

[true cedar]

like if the player hasn't moved?

or

[hearty shard]

david...

[unique crane]

eve..

[restive turret]

you...

[grand flower]

#plugins-dev-chat message

[hearty shard]

can u make it so overwatch players dont count for players to start the round

[modern plover]

(Guys how the fuck do I find out why the player died)

[upper vapor]

"i just wanna say.. uhhhhhhh..
i.. I HATE you :3"
-# -eve, probably sometime

[restive turret]

over on my watch

[upper vapor]

PlayerDiedEvent or something

PlayerDeathEvent

[true cedar]

oh

what the fuck why does it do that

[upper vapor]

smartwood

[true cedar]

no wonder its tanking tps for higher player counts thats 2500 a frame for 50 players

[modern plover]

Doesn't have the cause anywhere

Or does it..?

[grand flower]

Got a nice boost, not enormous but enough to be visible on graphs

[upper vapor]

check the DamageHandler

[true cedar]

im impressed all my optimizations were micro optimizations

[upper vapor]

next up: LabEngine

[modern plover]

There isn't one for environmental damage

[upper vapor]

UniversalDamageHandler

[modern plover]

How should I have known

[upper vapor]

contains the translation id

[unique crane]

How about..gameplay wise?

[upper vapor]

um

yes

:3

[modern plover]

Yes :3

Okay lemme try that

[upper vapor]

also the handler should have a ServerLogsText if you want natural language output

[upper vapor]

the DeathTranslations class has the death translations

you canc check against those

[grand flower]

No change in visible behaviour if that's what your asking. It's not very noticeable to players but the servers raised their average TPS on higher player counts a bit

[unique crane]

Yes but how about when firing

Im pretty sure that might cause issues

[grand flower]

How so

[upper vapor]

uhhhhhh

backtracking might take care of that?

not sure to what extent though

[grand flower]

Are we talking about round robin or the relative position caching

[unique crane]

Probably..

[worthy rune]

the FPC tracer is seperate from the distributer, it creates a new relative position every 4 fixed updates

[grand flower]

tbh at a send rate of 60hz I don't think round robin would really break backtracking

I'll look into the FPC tracer and see if I need to patch it too

[worthy rune]

i would say its unlikely to break anything with out SL has backtracking set up

[upper vapor]

that means
you could make them use vector3 instead

[worthy rune]

backtracking is super lenient in SL\

[grand flower]

yeah it should be anyway

We'll find out once I'm done writing the transpiler for this

[unique crane]

true..

Well try it and see

[grand flower]

kinda wish this loop was a for loop so I didn't have to insert an index heh

[true cedar]

AllHubs is a hashmap

set*

[grand flower]

what

[upper vapor]

var index = 0;
foreach ()
{
var position = list[index++]
}

[upper vapor]

you just realized

[true cedar]

its kind of funny cause like

its Never used as a hashset and doesn't really have a reason to be one

but i guess whoever designed it wanted fast removals

[grand flower]

😭

NW please tell me you're gonna fix this

dunno if I can even transpile it to something else

probs not if there's no other list of hubs

[upper vapor]

well uh

create an int local
add 1 every iteration

if someone gets removed between calls
well crap

[true cedar]

u cant really

[grand flower]

forward iteration won't matter here

i meant replacing AllHubs but like, yeah nah

also for index I can probably just get away with player id

[true cedar]

u can make ur own List<ReferenceHub> ig

[upper vapor]

probably don't

[grand flower]

praying NW fixes this ouch

[upper vapor]

i doubt they would

[grand flower]

but yeah for index I'll just create an index local and increase it every iteration

[upper vapor]

not really an internal breaking change but for plugins using it, oof

[grand flower]

I mean

[upper vapor]

or a bool

[grand flower]

LabAPI plugins should use Player.List/ReadyList

[upper vapor]

benchmark

[grand flower]

shouldn't matter aren't C# bools 64 bit anyway

[upper vapor]

old code though

[upper vapor]

what

[grand flower]

might be confusing it with something else

but yeah a bool works too

[upper vapor]

it's 1 byte

[grand flower]

yeah I know

[upper vapor]

oh

arenn't

okay

i'm blind

wait no

[grand flower]

hahaha

[upper vapor]

nvm

[grand flower]

i meant more like

the operations needed to set a bool and increment an int

but yeah I misworded it

dw

[true cedar]

32

all types less than an int are widened

[grand flower]

right

checked and the IL for flipping a bool and the IL for incrementing an int are both 4 operations

but by using a bool I don't have to modulo it

so better

[upper vapor]

or you can bitmask the int with 0b1

more operations though

[grand flower]

need to figure out how to add a condition to an if statement

[upper vapor]

labels

[true cedar]

what do you mean

[upper vapor]

there's no if in IL

[grand flower]

I know

:p

[upper vapor]

oh you mean which jump

[grand flower]

need to add the odd/even check there

[upper vapor]

instruction?

[true cedar]

add another brfalse

or brtrue

[grand flower]

aye

[true cedar]

wheres this code

[icy knoll]

me when id just do the if call inside of my method

[grand flower]

FpcServerPositionDistributor.LateUpdate

[upper vapor]

not performant enough

[grand flower]

yeah I'm transpiling this because it's already a hot path and I don't wanna make it any hotter

[upper vapor]

brfalse to the "skip" label

but wait

[upper vapor]

ye

find that instruction

and then instruction.labels[0]

brfalse to that

[true cedar]

or use a codematcher (the best way of doing transpilers0

[grand flower]

yeah im using a codematcher atm

[true cedar]

yippee

another person converted

[grand flower]

I mean it's the most update-compatible way

if I start using offsets, NW updates that method and I get headaches

[true cedar]

yea

plus just nicer to look at it

nobody knows what 18 instructions ahead is

[grand flower]

so how do I get a label from a codematcher

[icy knoll]

what is code matcher?

i just do what exiled does for their transpilers

[grand flower]

guessing I MatchEndForward these 3 instructions and then...?

[true cedar]

iirc its like

[grand flower]

"Find those 4 instructions, then insert my instructions afterwards"

[true cedar]

codematcher.instruction.labels

or wtv

[upper vapor]

why nop

[icy knoll]

ohhh that makes so much sense and actually so useful

[grand flower]

idk not a wizard that was generated in the IL for a function call

not needed?

[upper vapor]

nop means "no operation"

literally does nothing besides wasting time

[grand flower]

👌

[icy knoll]

how do we return this back tho?

[grand flower]

return matcher.Instructions();

[true cedar]

.InstructionEnumeration()

[upper vapor]

https://harmony.pardeike.net/articles/patching-transpiler-matcher.html

[grand flower]

or that ig

[true cedar]

plus

its super easy to make extensions for a codematcher

[icy knoll]

ah

[true cedar]

i think they're objectively better than normal transpilers

[upper vapor]

"i think they're objectively better"

they're probably objectively better

don't think :3

[true cedar]

no

my opinion is always right

tehy are objectively better

[upper vapor]

but if it's objective then it's not really an opinion, more of a fact

[true cedar]

yeah

its a fact

they're better

[hearty shard]

pai is objective pai

[icy knoll]

me when

[upper vapor]

specify the created type

[random scaffold]

why someone using transpiler

if exists labapi

[upper vapor]

because you can't do everything with labapi..?

[icy knoll]

im not used to transpiler, im not sure what the type should be

[random scaffold]

example?

[upper vapor]

says it there

CodeMatch

[icy knoll]

ohh

[upper vapor]

#plugins-dev-chat message

[grand flower]

if you have a single op to match it fails yeah

[icy knoll]

instead of this

new (OpCodes.Ldc_I4_1)

new Code.Ldc_I4_1_()

?

[upper vapor]

wha

no

[icy knoll]

ok

[upper vapor]

new CodeMatch()

also

Rider should let you do that

[icy knoll]

ohhh

right

ok

[upper vapor]

[icy knoll]

me when it's not

[upper vapor]

use the IDE, people

[true cedar]

1 sec

[upper vapor]

alt+enter
specify created type

does it not show that suggestion

[icy knoll]

so ur saying that this

does the same as this?

[icy knoll]

should change to MatchEndBackward or smth for the -1

[true cedar]

MatchEndBackward puts the index at the last instruction

that matches that

don't use backwards

[icy knoll]

ok

[true cedar]

backwards looks

backwrads

[icy knoll]

so what do you recommend for index - 1 then?

[true cedar]

.Advance(-1)

or just match two instructions

or wtv

[icy knoll]

peak

this is so much better

why does exiled do it in the other way

[true cedar]

tradition

[icy knoll]

this may make it so i actually do transpilers more

wtf

[true cedar]

i wrote all my transpilers for brights with them

u can do some really complex shit

[icy knoll]

i usually only do prefix or postfix

[hearty shard]

Im taking you

[icy knoll]

that's it

[hearty shard]

Ur my dev now thanks

[icy knoll]

LET HER FOCUS ON RUEI V3

[true cedar]

them but

[hearty shard]

negative

[icy knoll]

OR ILL TAKE AWAY UR SERVER

:3

[hearty shard]

focus on evehints

[hearty shard]

U cant

[icy knoll]

me when cables :3

[hearty shard]

sending hr report

[grand flower]

sanity check, if I want to pass in a local to a function call, I load the local and then call the function right

[hearty shard]

jk too lazy to do that again

[icy knoll]

hr cant do anything about me eating ur cable!! teehee

[hearty shard]

What

[true cedar]

yw

yea

[grand flower]

someclass.function(myLocalBool)

[hearty shard]

Im too braindead to understand what ur trying to say

[grand flower]

aight awesome

writing a transpiler

[true cedar]

is the method static

[hearty shard]

Oh

[true cedar]

if not u need to load the class to call it on first

[grand flower]

static

also how do I tell Harmony I want to load a local declared earlier

[true cedar]

huh

[grand flower]

oh

LocalIndex

nvm

[upper vapor]

Ldloc(index)

[grand flower]

yeah

[upper vapor]

ye

[hearty shard]

Transpilers are interesting

I gotta write one...

For smth..

[grand flower]

don't write transpilers unless absolutely necessary

[upper vapor]

maybe tomorrow

[grand flower]

I find them cool but also don't like them in terms of maintenance cost

[hearty shard]

It makes me learn them

[grand flower]

if NW updates something a transpiler modifies it could just break

[hearty shard]

However

Why would nw update code for round start

[grand flower]

what are you trying to do

[upper vapor]

me when local indexes change
and the numbers in compiler-generated classes' names change

[hearty shard]

Learn transpilers :3

[grand flower]

no i mean in round start

[hearty shard]

Also modify ScpSpawner and also

Modify the round start player count

Make ow not count

:)

Its just to a reason to learn transpilers

[grand flower]

it already shouldn't

[hearty shard]

However

False

It does

[grand flower]

interesting

[hearty shard]

Has been doing that since ages

[true cedar]

example

[hearty shard]

It was never changed to not count OW afaik

[grand flower]

then yeah have fun modifiying RoundSummary._ProcessServerSideCode

[hearty shard]

Uh actually

[true cedar]

i tried do that

[hearty shard]

Is that it?

[grand flower]

yeah it is

[true cedar]

it didnt work well

[hearty shard]

I dont think thats it

[hearty shard]

No

Thats the after round was started

[grand flower]

that's the coroutine that checks if a round should end

[hearty shard]

Im referring to before

[grand flower]

OH

you mean starting

[hearty shard]

Thats not it

yes

[grand flower]

okay yeah I see

[grand flower]

yeah it's hell

but it works now

[hearty shard]

When round is starting if 2 overwatch players are only players on round will start and then end immediately

Hence why i want to change that

[grand flower]

sec

[true cedar]

my patch on it?

[grand flower]

might be? what was the patch again

[true cedar]

mayhem victory thing

[grand flower]

New LabAPI event does it iirc

you can change the winning team

[true cedar]

oh

yea i love codematchers idk

i wish people used them more

newer harmony versions make them even better

[grand flower]

@hearty shard CharacterClassManager.Init

[hearty shard]

I know

Thats not the issue

The issue is

Im in bed

And ive never written a transpiler before

So this will be fun

[grand flower]

good luck

@true cedar how do I get the return value of a method call in IL

[true cedar]

what do u mean

its automatically loaded onto the stack

if thats what ur asking

[grand flower]

yep, awesome

[true cedar]

ldarg.0
ldarg.1
Z call Method(X, Y) 
// z is now on stack

[hearty shard]

pai so smart

[true cedar]

so true

the #1 iler

[grand flower]

ah wait

got the wrong label

for skipping the current iteration

label i'm currently matching is ending the loop

[hearty shard]

#1 nerd

[true cedar]

see and with a normal transpiler

[hearty shard]

Pai

[true cedar]

that would've taken 3 years to figure out

[hearty shard]

Can you rewrite harmonyyyy :3

[true cedar]

i WISH

[grand flower]

why rewrite perfection

[hearty shard]

Never speak again

[grand flower]

the thing even spits out IL with [HarmonyDebug]

i love it

[hearty shard]

Harmony is hell on earth

[hearty shard]

It does???

[grand flower]

yea

[hearty shard]

Damn

[grand flower]

on your desktop

[true cedar]

oh i didnt realize that LOL

[true cedar]

i always made a LogTranspiler method

[grand flower]

### Patch: static PlayerRoles.FirstPersonControl.NetworkMessages.FpcSyncData PlayerRoles.FirstPersonControl.NetworkMessages.FpcServerPositionDistributor::GetNewSyncData(ReferenceHub receiver, ReferenceHub target, PlayerRoles.FirstPersonControl.FirstPersonMovementModule fpmm, System.Boolean isInvisible)
### Replacement: static PlayerRoles.FirstPersonControl.NetworkMessages.FpcSyncData PlayerRoles.FirstPersonControl.NetworkMessages.FpcServerPositionDistributor::PlayerRoles.FirstPersonControl.NetworkMessages.FpcServerPositionDistributor.GetNewSyncData_Patch0(ReferenceHub receiver, ReferenceHub target, PlayerRoles.FirstPersonControl.FirstPersonMovementModule fpmm, System.Boolean isInvisible)
IL_0000: Local var 0: PlayerRoles.FirstPersonControl.NetworkMessages.FpcSyncData
IL_0000: Local var 1: PlayerRoles.FirstPersonControl.NetworkMessages.FpcSyncData
IL_0000: Local var 2: PlayerRoles.FirstPersonControl.NetworkMessages.FpcSyncData
IL_0000: // start original
IL_0000: ldarg.0
IL_0001: ldarg.1
IL_0002: call       static PlayerRoles.FirstPersonControl.NetworkMessages.FpcSyncData PlayerRoles.FirstPersonControl.NetworkMessages.FpcServerPositionDistributor::GetPrevSyncData(ReferenceHub receiver, ReferenceHub target)
IL_0007: stloc.0
IL_0008: ldarg.3
IL_0009: brtrue =>  Label0
IL_000E: ldloc.0
IL_000F: ldarg.2
IL_0010: callvirt   PlayerRoles.FirstPersonControl.PlayerMovementState PlayerRoles.FirstPersonControl.FirstPersonMovementModule::get_SyncMovementState()
IL_0015: ldarg.2
IL_0016: callvirt   System.Boolean PlayerRoles.FirstPersonControl.FirstPersonMovementModule::get_IsGrounded()
IL_001B: ldarg.1
IL_001C: callvirt   UnityEngine.Transform UnityEngine.Component::get_transform()
IL_0021: callvirt   UnityEngine.Vector3 UnityEngine.Transform::get_position()
IL_0026: call       static RelativePositioning.RelativePosition GameOptimizer.Patches.RelativePositioningCache::GetOrCreate(UnityEngine.Vector3 position)
IL_002B: ldarg.2
IL_002C: callvirt   PlayerRoles.FirstPersonControl.FpcMouseLook PlayerRoles.FirstPersonControl.FirstPersonMovementModule::get_MouseLook()
IL_0031: newobj     System.Void PlayerRoles.FirstPersonControl.NetworkMessages.FpcSyncData::.ctor(PlayerRoles.FirstPersonControl.NetworkMessages.FpcSyncData prev, PlayerRoles.FirstPersonControl.PlayerMovementState state, System.Boolean bit, RelativePositioning.RelativePosition pos, PlayerRoles.FirstPersonControl.FpcMouseLook mLook)
IL_0036: br =>      Label1
IL_003B: Label0
IL_003B: ldloca.s   2 (PlayerRoles.FirstPersonControl.NetworkMessages.FpcSyncData)
IL_003D: initobj    PlayerRoles.FirstPersonControl.NetworkMessages.FpcSyncData
IL_0043: ldloc.2
IL_0044: Label1
IL_0044: stloc.1
IL_0045: ldsfld     System.Collections.Generic.Dictionary`2<System.UInt32, System.Collections.Generic.Dictionary`2<System.UInt32, PlayerRoles.FirstPersonControl.NetworkMessages.FpcSyncData>> PlayerRoles.FirstPersonControl.NetworkMessages.FpcServerPositionDistributor::PreviouslySent
IL_004A: ldarg.0
IL_004B: callvirt   System.UInt32 Mirror.NetworkBehaviour::get_netId()
IL_0050: callvirt   virtual System.Collections.Generic.Dictionary`2<System.UInt32, PlayerRoles.FirstPersonControl.NetworkMessages.FpcSyncData> System.Collections.Generic.Dictionary`2<System.UInt32, System.Collections.Generic.Dictionary`2<System.UInt32, PlayerRoles.FirstPersonControl.NetworkMessages.FpcSyncData>>::get_Item(System.UInt32 key)
IL_0055: ldarg.1
IL_0056: callvirt   System.UInt32 Mirror.NetworkBehaviour::get_netId()
IL_005B: ldloc.1
IL_005C: callvirt   virtual System.Void System.Collections.Generic.Dictionary`2<System.UInt32, PlayerRoles.FirstPersonControl.NetworkMessages.FpcSyncData>::set_Item(System.UInt32 key, PlayerRoles.FirstPersonControl.NetworkMessages.FpcSyncData value)
IL_0061: ldloc.1
IL_0062: // end original
IL_0062: ret
DONE

[true cedar]

u can prob do that without codematchers

[hearty shard]

Oh....

My desktop ....

Is hidden....

[grand flower]

F

[hearty shard]

it looks so much nicer...

[grand flower]

wait nvm the label is correct

goes to the end of the loop but jumps back to the start if we've still got elements

im stoopid

[true cedar]

also have u ever seen the internals of harmony

[grand flower]

no and i won't ever look at them

[true cedar]

yea

[grand flower]

it's beautiful on the outside and it's all that matters

[true cedar]

even then

NOTHING IS DOCUMENTED!!

and the things that are have the most useless documentation

[grand flower]

also can I tell the codematcher to set its position to a specific label

[true cedar]

um

one sec

probably not

i mean u can

[grand flower]

eh it's fine I'll just inject the instructions when I find the label

[true cedar]

instructions.FindIndex(x => x.Labels.Contains(label))

[grand flower]

guess that'd work yeah

aight time to test it

matcher.InsertAndAdvance(
    new(OpCodes.Ldloc, isOdd.LocalIndex),
    new(OpCodes.Ldc_I4_0),
    new(OpCodes.Ceq),
    new(OpCodes.Stloc, isOdd.LocalIndex)
);

[true cedar]

why are u

huh

oh thats flipping i

[grand flower]

isOdd = !isOdd

wtf

no harmony patching exception

[true cedar]

https://learn.microsoft.com/en-us/dotnet/api/system.reflection.emit.opcodes.not?view=net-9.0#system-reflection-emit-opcodes-not

[grand flower]

oooh ty

ah fuck

the IL looks fucky

[true cedar]

can i see ur codematcher

[grand flower]

dm'd

guess i need to move to the next instruction

matcher.Advance(1).Insert...

huh

[restive turret]

Cyn and it's transpilers

[grand flower]

im stupid it couldn't match anymore because I added my instructions

[true cedar]

lol

[grand flower]

IL looks perfect now

i'd test it out but dummies won't cut it since they don't walk much bleh

uh oh

[restive turret]

Haha

254 player?

[grand flower]

100

how am I making it worse with this

nvm

it is better

without

okay it's gotta be some other code

it's only because they were captains

[hearty shard]

What did bro do

[grand flower]

probs another plugin

without, lemme try with

[restive turret]

Try with 254 player

[grand flower]

with

seems to be a bug

one of the QA testers doesn't see other players move

but another sees us fine

i do see some wearablesync exceptions

probs the dummies

[2025-05-25 14:44:19.063 -07:00] [STDOUT] InvalidOperationException: Unable to create WearableSyncMessage with payload of 183.
[2025-05-25 14:44:19.063 -07:00] [STDOUT]   at PlayerRoles.FirstPersonControl.Thirdperson.Subcontrollers.Wearables.WearableSyncMessage..ctor (ReferenceHub hub, PlayerRoles.FirstPersonControl.Thirdperson.Subcontrollers.Wearables.WearableElements flags, Mirror.NetworkWriter payloadWriter) [0x00046] in <2343be033e9f4e37923f780ece756d8e>:0
[2025-05-25 14:44:19.063 -07:00] [STDOUT]   at PlayerRoles.FirstPersonControl.Thirdperson.Subcontrollers.Wearables.WearableSync.OnHubAdded (ReferenceHub hub) [0x00066] in <2343be033e9f4e37923f780ece756d8e>:0
[2025-05-25 14:44:19.063 -07:00] [STDOUT]   at (wrapper delegate-invoke) System.Action`1[ReferenceHub].invoke_void_T(ReferenceHub)
[2025-05-25 14:44:19.063 -07:00] [STDOUT]   at ReferenceHub.Start () [0x00009] in <2343be033e9f4e37923f780ece756d8e>:0

no exceptions but still not moving hmmm

ah odd players can't see anyone move lol

probs fucked up my logic

[hearty shard]

Interesting

[grand flower]

Gonna fix my logic but @unique crane since you wanted results of round robin ^

those results do include relative position caching which gives like 4-7 ticks by itself

but the rest is just round robin

[unique crane]

I would put it to normal round test

with real players

[grand flower]

yeah we will once I get that odd players logic fixed

[unique crane]

If its gonna work

I might ask beryl and hubert what they think about it

[hearty shard]

Hubert...

David when fix

[restive turret]

fix deez

[hearty shard]

BE QUIET

[grand flower]

now to figure out why the fuck my even/odd logic is screwy

    IL_007E: ldloc      2
    IL_0082: not
    IL_0083: stloc      2

kinda wonder if it's not

[unique crane]

@grand flower If you want, next up you can look at animations (IK)

Maybe you can optimize that

(thats like top 1 performance heavy currently)

[hearty shard]

Oh

Can you show everything :3

[unique crane]

Uhhh tomorrow

Ive got final matura exam tomorrow

and I gotta eep well

[hearty shard]

Im gonna do horrible things just u wait

[unique crane]

):

[grand flower]

serverside? D:

https://cdn.discordapp.com/attachments/1348879302036557885/1376321532036644934/SCP-_Secret_Laboratory_-_2025-05-25_15-10-10.mp4?ex=6834e6d0&is=68339550&hm=b589f292cfad02156f5ac9aed38b9dc03f3023cfe87d1606b83eb1bf9593ed69& at high speeds round robin does provide some jitter but like, 255% speed. 100% movement boost has no issues/

https://cdn.discordapp.com/attachments/1348879302036557885/1376321826715861012/SCP-_Secret_Laboratory_-_2025-05-25_15-11-21.mp4?ex=6834e716&is=68339596&hm=d86ee53abed1a495992492dc9eba9049a7525b95c604b724579f95b58fde42df& also has a lot of jitter for 173 - but like, it's 173, it doesn't happen when you're looking at it.

for the performance boost it's worth it IMO

Will test with actual players soon™️

[hearty shard]

interesting

[grand flower]

also backtracking was still working, we'll have to see with actual players (as always) but it seems like there's very little cons to this

[hearty shard]

turn on cheats and test

[grand flower]

inb4 fix cheating problem with round robin /s

hah jk

[hearty shard]

Whats back tracker used for again

Desync / weird movement iirc

[grand flower]

rewinds people for hitreg

iirc

[hearty shard]

Or that.

[grand flower]

i'll have to check how expensive it is and if i need to patch it

[hearty shard]

I forgot about shooting

[grand flower]

but i doubt it

hell with round robin I don't think I'm gonna be doing much more optimization after this

[plain gazelle]

Using a raycast to look at a Pickup on the ground, how would go about to get the Pickup as a reference?
GetComponent<Pickup> on the collider?

@hearty shard Pickup.Get only accepts a ushort or an ItemPickupBase
-# not sure what an ItemPickupBase is

[hearty shard]

ItemPickupBase is SLs Pickup

Pickup is the labapi wrapper of it

You might be able to Pickup.Get(GetComponent<ItemPickupBase>)

[plain gazelle]

Will try that, thank you+

[icy knoll]

if that doesn’t work, usually what works for me on colliders is collider.transform.root.GetComponent<ItemPickupBase>()

every time I try to do GetComponent on the collider i get null so, yeah

[buoyant oyster]

ev.IsAllowed = true in Hurting event is not working for two members of the same team, I'm def missing something but idk what it is

[teal junco]

Is this with FF on?

[buoyant oyster]

FF has to be on?

man

[teal junco]

uh yeah

the game wont even attempt hurting between two players if FF isnt on

[hearty shard]

^^

Hurting is when damage is attempted

IsAllowed is always true in that case i believe

You'll need a work around

[grand flower]

patch in your own event

[grand flower]

@unique crane Tested with a few players, no discernible gameplay changes other than the jitter in the videos

The performance boost is also tied to the resources the machine has

If the server is given a single core, the performance boost (while there) isn't as good

But otherwise still a nice thing for servers that need a little boost

[grand flower]

Especially high player counts, I think you'd be able to play at a stable 30 tps w/ 100 players.

The issue would likely be client performance at that point

This is still useful for 48/64 player servers as long as you don't starve it of resources

[upper vapor]

Try setting ForceFullFriendlyFire (FFFF) in the damage handler to true

[heady turret]

What did you do?

[upper vapor]

Wizard level trickery

[grand flower]

Implemented round robin to the FpcServerPositionDistributor, helps a lot with server performance on higher player counts

[grand flower]

^

[upper vapor]

How did you handle even/odd frames jf you got around to it?
I have an idea but I wanna know if you have a solution

[grand flower]

Added a bool to the method and just flipped it before it looped

And within the loop it keeps track if the current hub is odd or even with another bool

[upper vapor]

How do you track it between calls tho

[grand flower]

Wdym between calls

The bool that tracks whether the current frame should treat odd/even is a static if that's what you mean

[upper vapor]

A field?

[grand flower]

I added a condition to the loop that checks if the update should skip the player

Yeah

Sec

[upper vapor]

Ye

So like
ldfld State
ldc.i4.0
ceq
dup
stfld State
stloc localState

Not sure if the bitwise not instruction is faster than ceq, have you benchmarked it?

[grand flower]

bitwise not didn't work

for some reason

I could probably not make it a function call but like

negligible

[celest thorn]

what is even the round robin?

never heard of it

[grand flower]

multiple ways of doing it but one is you update odd players one frame, and then even the next

so you cut the amount of updates per tick in half

and then inside the loop @upper vapor

_isEven is the frame state, isEven is whether or not the current hub is even

then I just check if we should skip the hub by adding a condition

which is just this atm

although in the future I will be skipping dummies/the host player probably and other stuff
also got some plans for adaptive update rates based on the current amount of players alive

[celest thorn]

but what this is like?

[grand flower]

what do you mean

[celest thorn]

what is this?

im confused

[upper vapor]

not surprised

[grand flower]

A way to optimize servers with high player counts

yeah I did like

Ldloc, not, Stloc, just didn't do anything, even players would never get updates and saw everybody as standing still

[celest thorn]

oh but for what aspect?

[grand flower]

define aspect

It optimizes the update rate of player positions

[celest thorn]

Oh

[grand flower]

by only sending updates to half of the players every tick

meaning your 100 player server would act as a 50 player server in that regard

[celest thorn]

interesting

[grand flower]

if you can keep a high tickrate you get minimal side effects from doing that

[celest thorn]

where did you even patch for this?

[grand flower]

FpcServerPositionDistributor.LateUpdate

[celest thorn]

interesting

[grand flower]

I've got some other patches in that function that give it some boosts

[celest thorn]

never thought of that

[teal junco]

fire

[grand flower]

You can thank Fortnite for it

[celest thorn]

wait fr?

[grand flower]

It's one of the many tricks they do to run 100 player servers smoothly enough

[teal junco]

does it only cause issues for high speeds?

[grand flower]

Yeah Epic Games are pretty damn good at that stuff

[celest thorn]

Cool asf

[grand flower]

yes and somewhat minor, i've got videos

[upper vapor]

What if you dup before stloc and brfalse after stloc

[grand flower]

^

[upper vapor]

In the loop

[grand flower]

wouldn't that just be more instructions though

compared to just the current four

[upper vapor]

Wouldn't you need to load the local again?

[teal junco]

that doesnt even look terrible

[grand flower]

that's 255% movement speed, so like, hardly ever happens for most servers

and 173's jitter isn't even that bad

[grand flower]

ah i see

[celest thorn]

wouldn't this cause issues with like hitreg

[grand flower]

maybe if I get rid of the function call

[celest thorn]

etc...

[grand flower]

nope, was brought up by davidsebesta but it works perfectly fine

backtracking is lenient in SL and still holds up

[celest thorn]

but im interested to know what else you did

[upper vapor]

Yeah that would eliminate some nanoseconds

[celest thorn]

to optimize it

[grand flower]

tbf the hot path in that method is the looping, calculating relative positions and sending the data

and I wanna keep the function call & pass in the hub actually

so I can do more checks if I want to be aggressive and do stuff like skipping dummies, the host player, and other custom checks

[grand flower]

relative position caching

[celest thorn]

what you mean?

like you do half and half

for it

thats it?

[grand flower]

#plugins-dev-chat message

[upper vapor]

OptiFine for SL

Plugin

[grand flower]

clientside defo needs some optimization at higher counts

[teal junco]

you mean it steals code?

[grand flower]

but we can't do it so w/e

[celest thorn]

so each player is going to create a new relative position?

[teal junco]

yes because I will make optifine for SL

[celest thorn]

and you just strip them of it?

[grand flower]

I cache relative positions in that method so it doesn't compute them more than once per player

so at 48 players, it's 48 computed relative positions, not 48^2

[celest thorn]

im really into this like how do you manage backtrack and how it works in general

because desync if done improprely would be crazy

[grand flower]

backtrack isn't affected by either of those fixes

[celest thorn]

so if im getting right

waypoints get duplicated?

for each player?

[grand flower]

those calculations do yeah

even though they won't change in that loop

for each player combo

so playercount^2

48 players means 2304 relative position calculations

give or take

[celest thorn]

So how do you manage them?

[grand flower]

with caching it's just playercount, so 48

[celest thorn]

im so confused

on how you remove them

without causing issues

[grand flower]

I just have a cache that I clear at the start of LateUpdate

[celest thorn]

on the FPC distrubution?

[grand flower]

and I patched the code to call a function that gets the cached relative position for the given Vector3 if any, or creates and caches it

yeah

The cache is allocated once, I think I set it to like, 128 elements, so in theory it never gets re-allocated

[celest thorn]

im confused

so you send each data

i just don't see issue with this code

im being serious

[grand flower]

check the FpcPositionMessage constructor

[celest thorn]

OH

i see now

[grand flower]

yeah

bad

[celest thorn]

so you just cache them

if they are missing

and re use them?

am i right?

[grand flower]

yep

[celest thorn]

        private static void LateUpdate()
        {
            if (!NetworkServer.active || !StaticUnityMethods.IsPlaying)
            {
                return;
            }
            FpcServerPositionDistributor._sendCooldown += Time.deltaTime;
            if (FpcServerPositionDistributor._sendCooldown < FpcServerPositionDistributor.SendRate)
            {
                return;
            }
            FpcServerPositionDistributor._sendCooldown -= FpcServerPositionDistributor.SendRate;
            foreach (ReferenceHub referenceHub in ReferenceHub.AllHubs)
            {
                if (referenceHub.Mode != ClientInstanceMode.Unverified && !referenceHub.isLocalPlayer)
                {
                if(!Dictionary.TryGetValue(player, out FpcPositionMessage message)
   Dictionary.Add(referencehub, new FpcPositionMessage(referenceHub))
referenceHub.connectionToClient.Send<FpcPositionMessage>(message);
                }
            }
        }

so something like this?

[grand flower]

nope

the patch for caching the relative position isn't done in that method

the only thing done in that method is clearing the cache at the very start

[celest thorn]

so you do it on the ctor?

[grand flower]

in GetNewSyncData

[celest thorn]

48^2 because each player target each others

[grand flower]

yep

[celest thorn]

so playerA will send a packet

and PlayerB does the same

2^2 for example

[grand flower]

You send every player the position of every player

[celest thorn]

So you just create a dictionary in there add them

[grand flower]

You could if you wanted to, I just made it easier with a method

[celest thorn]

I just dont get how you save the data