archived-code-general

[tulip spruce]

and the settingsScript control the settings menu

which pops up when I press options

[rigid island]

okay but they do that already no? because its in DDOL scene as singleton

your error isnt related to settings though, its because you need a way to assign pauseMenu

[tulip spruce]

they do, if I in the Main menu first click on options so the settingsCanvas has been activated once

then everything works as I intended it to

[tulip spruce]

ok

[rigid island]

where does that happen exaclty? i dont see it in these scripts

[plucky inlet]

Sounds like you just drag dropped it to your script in inspector and the gameobject holding the UI is not a child of the scripts gameobject. In that case, switching scene will destroy the canvas and your reference in inspector will turn null. You can check that, if its exposed in your inspector of the script.

A solution would be, to keep your UI a static thing over all scenes by making it a child of your scripts gameobject.

[full canopy]

can someone explain to me why you aren't allowed to have a prefab reference itself correctly? the easiest workaround is to just have two of the exact same prefab but with the reference pointing to each other, but (1) that's stupid, and (2) I'm now spawning alternating objects using the reference and any changes I make to one prefab I have to remember to bring to the other.

its mega annoying and has been a pain to debug for forever all to find out its forcing the reference to be something i didn't put in it in the first place
it broke so many features that i was trying to work around

[rigid island]

what are you trying to do exactly ? prefabs are mainly an editor thing not ingame

[full canopy]

you don't use prefabs in runtime?

[rigid island]

prefabs are just "templates" to gameobject

[full canopy]

exactly

spawn in new objects from prefabs during runtime

what else would i do lol

[rigid island]

yeah the one spawned is a gameobject

so what do you mean prefab referencing itself

you want the spawned object ?

[full canopy]

oh i see the confusion

the spawned object needs a reference to its original prefab because the spawned object can spawn another copy of itself (its a lot more complicated then that but to shorten the whole process down thats sufficient lol)

[rigid island]

and object spawning itself sound like a recursion nightmare lol

[full canopy]

its not a recursion nightmare

[plucky inlet]

It sounds like you are trying to force your system into it instead of taking a step back and for example using a static manager that can be fetched from instead of let the object do runtime stuff on its own

[full canopy]

if you want more details-
its a magazine for a gun. its a vr game so there is an ammo pouch, the game checks what items you have picked up last, and associates a prefab from them. picking up a magazine or the gun it goes in associates that magazine prefab with the ammo pouch.

[knotty sun]

Ah, I see the problem, when you spawn from a prefab all of the references are updated from editor instances to runtime instances. So a prefab referencing itself the prefab reference would also be updated to the new gameobject instance

[plucky inlet]

you might wanna look into addressables and use references instead of prefabs

[full canopy]

yes, but i dont think this should be forced - the workaround is to just have an identical second prefab that alternates between itself and the other

[full canopy]

what do you mean use references instead of prefabs? it uses a reference to a prefab to spawn the new one

[knotty sun]

no, best way is to use dependency injection and inject the original prefab reference into the new game object

[plucky inlet]

When you reference a prefab, it will get loaded into memory. If you use addresables, you just have a "pointer" to it and can spawn it from any script as a raw object instead of a reference of a reference

[full canopy]

thx for the tips
still though, why is this a forced feature?

[plucky inlet]

You can still pass in the prefab you are instantiating to the new instantiated object by script with a simple Init() method

[knotty sun]

because the instantiation routine does not know the difference

[slate fiber]

Has anyone ever worked with easy save 3 and webgl? I cant seem to make Easy save work on webgl games

[knotty sun]

Contact Easy Save, we do not provide support for 3rd party assets

[slate fiber]

Alrite

[copper rose]

Hi there, i have some trouble trying to get working my game smoothly, its first time i do it, so im searching in documentation and some threads in unity forum and this is what i got, it works but in unity editor works smooth with moouse or gamepad, but not in real device with my gamepad , there is my input manager and my cam controller

[thin aurora]

Your !ide does not look properly configured. Is this correct?

[tawny elk]

💡 IDE Configuration

If your IDE is not autocompleting code
or underlining errors, please configure it.

Select one:

• Visual Studio (Installed via Unity Hub)
• Visual Studio (Installed manually)
• VS Code
• JetBrains Rider
• :question: Other/None

[thin aurora]

Namely the fact methods like LateUpdate are unknown, and in general syntax highlighting is very off

Please configure it first, then we can help you

[copper rose]

is visualcode, ok let me see what i have wrong

[vestal arch]

you mean visual studio code?

[copper rose]

i got installed visual code with the common but still looking like this

[copper rose]

im using macos yes

[thin aurora]

If you use Visual Studio Code, I suggest you use Visual Studio or Jetbrains Rider instead. I believe support has generally improved for Unity, but the other two editors have very good support

[vestal arch]

visual studio for mac is dead, visual studio code is the recommended replacement (idk about rider)

[thin aurora]

In that case I suggest you check out Jetbrains Rider. It recently got a license that allows it to be free

[copper rose]

that is

[thin aurora]

That it, depending on a few prerequisites obviously

[vestal arch]

but also visual studio code does work fine in my experience

[copper rose]

for me works fine, then only i have found is lateupdate and other methods from unity are grey

[vestal arch]

if you're familiar with it and can set it up properly then you should just use the one that's more comfortable 🤷

[copper rose]

im comfortable with that

[thin aurora]

Yes, so I think you're still missing some configuration steps. I suggest you follow each step once more to verify

[copper rose]

this is not legible?

[thin aurora]

These methods should be known Unity methods, but they are not

Right now they are listed as being unknown methods which is not correct

If you lack the intellisense here, you likely have to manually write each method, do you?

[copper rose]

just disable-enable the extension

[thin aurora]

That looks even worse 😄

Now ALL the intellisense and highlighting is gone

If you want/can change then you can also try Jetbrains Rider. It has a single plugin you need to install I believe

[copper rose]

ok uninstalled extension and installed again

[thin aurora]

Weird. You have auto completion but it doesn't know LateUpdate is a reserved method name

[copper rose]

i will check later jetbrains, but i want to try to improve this code 🙂

[thin aurora]

Generally I would expect it to work properly but maybe this is just VSC being weird as always

[thin aurora]

I would definitely consider it, yeah. An actual IDE for C# is always better than a text editor with plugins 😛

Regardless, I suppose this is how it works then

[vestal arch]

LateUpdate and other messages are recognized as proper messages in my ide

[copper rose]

ok im just installed for now jetrider

im earing about that long time ago so maybe is a good moment for testing

[thin aurora]

Jetbrains is skyrocketing when it comes to supporting different c# frameworks

Which is both surprising and generally weird considering you'd expect applications like VSC and VS to get the same treatment since they are not from a 3rd party company

[copper rose]

ok, about the code to manage my gamepad, any suggestion? i have to deal also with the bad support on macos with gamepad, my ps4 gamepad works, but some times when it comes to standby i have to re-plug again

[plucky inlet]

are you using the old input or the new input system?

[copper rose]

new input system

[copper rose]

this is how i manage the input from gamepad

[plucky inlet]

And your issue is your gamepad movement is not smooth in runtime but in editor?

[copper rose]

yes

[plucky inlet]

Did you ever test another gamepad?

[copper rose]

i only have this one

do you think the code is fine right?

[plucky inlet]

I mean, it works in editor, so either macos is trying to do something in "game mode", maybe disable that, or your gamepad is trying to do something ingame mode or is broken entirely, if its having connection issues. macos on latest updates?

[copper rose]

yes, but i use unity editor in macos but real device is android, with an steelseries stratus

[plucky inlet]

well... that is some minor information you should put in your introduction of the question 😄

How is it connected, BT, wired?

[copper rose]

bluetooth

[grand aurora]

Hi, gyus. I am trying to bake a light. After baking has been completed, the light on the wall seem to be correct, but after importing a lightmap, everything becomes too bright. Just moved to unity 6 and have never had this problem before. Also, it does not seem that baked light maps are so bright..

The problem was that, for some reason, unity marks all lightmaps as default in the scene folder. I changed it to lightmap, dir lightmap and shadowmask accordingly and it fixed the problem. But looks like unity set it as default every time..

Why unity marks lightmaps as default? it takes some time to correct it...

[copper rose]

you add the scene as additive?

[late lion]

Could you explain what you mean by importing after baking? Are you talking about just a few seconds after it has finished baking it automatically starts reimporting the textures?

[copper rose]

im testing the game pad in stumble guys (i dont know another game) and my gamepad works smoothly, so the problem should be my code

[plucky inlet]

And its connected to BT to your editor as well?

[grand aurora]

yes

[plucky inlet]

Ah, good to know. But there is a lot going on from editor to android. So not sure, its the code itself, as it works in editor.

[copper rose]

yes

[wheat spruce]

private void Update()
{
  _speed += (_moveDirection != Vector2.zero ? 1 : -1) * _speedRate * Time.deltaTime;
  _speed = Mathf.Clamp(_speed, 0f, _maxSpeed);
}

private void FixedUpdate()
{
  Vector2 move = _moveDirection * _speed * Time.deltaTime;
  _rb.AddForce(move, ForceMode2D.Force);
}```Does this seem like a decent way to implement a simple vehicle controller? It's a little stiff which is mostly because the code hasnt been fine tuned yet

[leaden ice]

deltaTime shouldn't be factored into the force

(the one in FixedUpdate)

Also all of it should really go in FixedUpdate

[cyan ivy]

Hey :)
When you're working in the FixedUpdate block, you want to use Time.fixedDeltaTime as opposed to Time.deltaTime. That probably helps smooth out some of the jitter you're seeing.

[heady iris]

Time.deltaTime is set to Time.fixedDeltaTime within FixedUpdate

It would still be incorrect to factor in the length of the frame or the length of the physics update when using AddForce, either way

[wheat spruce]

private void FixedUpdate()
{
  _speed += (_moveDirection != Vector2.zero ? 1 : -1) * _speedRate * Time.fixedDeltaTime;
  _speed = Mathf.Clamp(_speed, 0f, _maxSpeed);

  Vector2 move = _moveDirection * _speed;
  _rb.AddForce(move, ForceMode2D.Force);
}```This definitely feels much smoother

[heady iris]

That looks correct.

[wheat spruce]

plus move doesnt end up being really small

[heady iris]

You accelerate towards a target velocity, moving by Time.fixedDeltaTime * _speedRate each physics update

Well, it's not really a speed, but close enough

you accelerate towards a target force value

you'll wind up getting a sort of "double-smoothing" effect from this -- you accelerate towards a force value, and that force value causes you to accelerate!

You might compare this to just doing _rb.velocity = move; and seeing how that feels

[wheat spruce]

Yeah, the _speed property is mostly so I can control the initial acceleration. But it does seem counter-intuitive considering the rigidbody does the same thing

[leaden ice]

make sure interpolation is enabled on your Rigidbody too

[wintry crescent]

Is there a many-to-many dictionary?

because I can only have 1 instance of a key

actually no, I'll just use a list as the value

[vestal arch]

dict to set/list of value

[magic harness]

Hello guys. Im having a bit of a gigantic trouble with loading some data from Json.

My current project has a single entry point for Data, which is the Profile class. We load a Json file containing an object of type Profile at startup and then save it on application quit.

Profile has a list of characters, characters have an inventory, inventory has inventory entries and inventory entries have and item.

Item is polymorphic so it can be an Equipment.

Equipments have two lists, one for PrimaryAttributeModifiers and a Second one for SecondaryAttributeModifiers.

An Equipment is created whenever the player collect loot from the ground(example) and it calls the Equipment constructor which accepts an EquipmentSO.

This far, everything works great. It serializes the data correctly and by debugging the load call on the Json, also reads everything correctly. All the lists work in exception of the attribute modifiers.

I've tried a plethora of solutions but it seems to not be able to deserialize the attribute modifier list by any means at all.

Things i've tried:
-Simpliying attribute modifer inheritance to not include vector 2(newtonsoft doesnt like the normalize properties)
-Changing serialization and deserialization settings
-Removing the Unity's SerializeField properties and just using JsonProperty instead.
-Adding parameterless constructors with the JsonConstructor attribute.
-Saving and loading the items while equipped (changing the list type)

[heady iris]

for the first point, you can add converters that correctly handle things like Vector2

you'll need to show us how an attribute modifier is defined and how you're storing them

[magic harness]

Ok.

namespace Attributes
{

    [Serializable]
    public abstract class AttributeModifier
    {
        [JsonIgnore] public Action OnValueChanged;


        [field: SerializeField, JsonProperty] public bool isPercentage { get; set; }
        [field: SerializeField, JsonProperty] public float timeActive { get; private set; }
        [field: SerializeField, JsonProperty] public int minRoll { get; private set; }
        [field: SerializeField, JsonProperty] public int maxRoll { get; private set; }
        [field: SerializeField, JsonProperty] public bool isRolled { get; private set; }
        [field: SerializeField, JsonProperty] public float value { get; private set; }
        [JsonIgnore]
        public float Value
        {
            get
            {
                if (!isRolled) return RollModifier();
                return value;
            }
        }


        [JsonConstructor]
        public AttributeModifier() { }

        public AttributeModifier(int minRoll, int maxRoll, bool isPercentage = false, float timeActive = 0)
        {
            this.minRoll = minRoll;
            this.maxRoll = maxRoll;
            this.isRolled = false;
            this.timeActive = timeActive;
            this.isPercentage = isPercentage;
        }


namespace Attributes
{
    [Serializable]
    public class PrimaryAttributeModifier : AttributeModifier
    {
        [field: SerializeField, JsonProperty] public PrimaryAttributes stat { get; set; }


        [JsonConstructor]
        public PrimaryAttributeModifier() : base() { }

        public PrimaryAttributeModifier(PrimaryAttributes stat, int minRoll, int maxRoll, bool isPercentage) : base(minRoll, maxRoll, isPercentage)
        {
            this.stat = stat;
        }
    }
}

CLASS EQUIPMENT CONSTRUCTOR

public Equipment(EquipmentSO preset) : base(preset)
    {
        Debug.Log(this);
        enchantItems = new();
        upgradeLevel = 0;
        primaryModifiers = new();
        secondaryModifiers = new();
        this.slot = preset.slot;
        this.enchantSlots = preset.enchantSlots;
        foreach (PrimaryAttributeModifier modifier in preset.primaryModifiers)
        {
            PrimaryAttributeModifier newMod = new PrimaryAttributeModifier(modifier.stat, modifier.minRoll, modifier.maxRoll, modifier.isPercentage);
            newMod.RollModifier();
            primaryModifiers.Add(newMod);

        }
        foreach (SecondaryAttributeModifier modifier in preset.secondaryModifiers)
        {
            SecondaryAttributeModifier newMod = new SecondaryAttributeModifier(modifier.stat, modifier.minRoll, modifier.maxRoll, modifier.isPercentage);
            newMod.RollModifier();
            secondaryModifiers.Add(newMod);

        }
    }

and its stored here

also just tried sving the equipment while it is equipped, so the list is List<Equipment> instead of List<item> and the same result, so. yeah

[wintry crescent]

wait, so the issue is that this class doesn't get serialized/deserialized properly, yes?

[magic harness]

The equipment class doesnt get deserialized properly. The Serialized Json data is totally fine.

The only things missing are the modifiers

doesnt matter what happens, the list storing the modifiers gets set to empty

[wintry crescent]

but where exactly is the issue - those foreach loops in Equipment constructor, do they run properly?

[magic harness]

they do. If i kill a mob, loot the item everything is fine

After i close the game and check the serialized data

everything is fine

when i open the game and debug.log the data that i will deserialize, everything is fine

when rebuilding the Equipment, lists are empty

[wintry crescent]

So wait, let me clarify - deserialization happens here, right? in this constructor you sent here?

EquipmentSO is your serialized data and you make your equipment from it

so those foreach loops, in that constructor, do they run?

[magic harness]

no. i have a default parameterless constructor in the same class.

No EquipmentSO is a scriptable object that i use to create normal instances of equipments(since i want to have randomized stats per drop)

    [JsonConstructor]
    public Equipment() : base()
    {
    }

this is what gets called for deserializing

My Json settings are as follows:
Serializing
TypeNameHandling = TypeNameHandling.All,
ObjectCreationHandling = ObjectCreationHandling.Replace,
DefaultValueHandling = DefaultValueHandling.Ignore,
Formatting = Formatting.Indented,
PreserveReferencesHandling = PreserveReferencesHandling.Objects,

Deserializing:

TypeNameHandling = TypeNameHandling.Auto,
ObjectCreationHandling = ObjectCreationHandling.Replace,
DefaultValueHandling = DefaultValueHandling.Ignore,
PreserveReferencesHandling = PreserveReferencesHandling.Objects

[eager tundra]

don't you want to use the other constructor with the actual parameters? this way you're telling the deserializer to simply create a new instance with default values

[magic harness]

but i am using ObjectCreatingHandling.Replace and PreserReferencesHandling.Objects.

It works fine with all the other data that i'm saving and loading

not sure how i can pass a scriptable object parameter to the actual deserializing construtor tbh

[wintry crescent]

why TypeNameHandling.Auto and not TypeNameHandling.All? not saying that's the issue, just asking what's the reason

[leaden ice]

Auto is basically an optimization where the type name isn't included when it's not ambiguous.

[wintry crescent]

I'd double-check with .All anyway, had stranger bugs than that in the past

[magic harness]

ive almost done all possible combinations of serializer settings in there

Yeah i've just removed all parameterless constructors from equipment, item and attribute modifiers

[eager tundra]

where exactly is deserialization happening?

[magic harness]

  public Profile LoadProfile(string name)
        {
            string fileLocation = GetPathToFile(name);
            var settings = new JsonSerializerSettings
            {
                TypeNameHandling = TypeNameHandling.Auto,
                ObjectCreationHandling = ObjectCreationHandling.Replace,
                DefaultValueHandling = DefaultValueHandling.Ignore,
                PreserveReferencesHandling = PreserveReferencesHandling.Objects
            };

            if (!File.Exists(fileLocation))
            {
                throw new ArgumentException($"No persisted GameData with name '{name}'");
            }
            Debug.Log(JsonConvert.DeserializeObject<Profile>(File.ReadAllText(fileLocation), settings));
            Debug.Log(File.ReadAllText(fileLocation));
            var profile = JsonConvert.DeserializeObject<Profile>(File.ReadAllText(fileLocation), settings);
        
            return JsonConvert.DeserializeObject<Profile>(File.ReadAllText(fileLocation), settings);
        }

[magic harness]

it just can't create items anymore because it doesnt know the parameter, of course

[tiny delta]

Howdy folks, quick question. Is it possible to change a setting or something to make it so FixedUpdate waits for Start before running? Maybe I am confused or something but it seems like it oftentimes a frame of FixedUpdate will occur before Start.

[wintry crescent]

use Awake instead of start, that's guaranteed to always run first, although ngl that's a new one for me

[tiny delta]

I did consider that, but it depends on a reference to another script which gets some stuff during Awake, so it would mess everything up

Currently I just have a variable called "started" which is set to true once Start runs, but it feels like bad practice to do

[heady iris]

Start runs before the first Update call

There aren't any other guarantees

[obsidian mulch]

It should run awake on all objects before any fixedupdate calls afaik

[heady iris]

Awake will run as soon as the object comes into existence, yeah

[wintry crescent]

sanity check... try serializing/deserializing an array instead of a list

[heady iris]

(as long as the game object is activated)

[obsidian mulch]

There’s also lateupdate, but not sure if there’s latefixedupdate?

[wintry crescent]

if you REALLY cannot figure out how to put your stuff inside Awake, then I guess that's the best you can do...

[wintry crescent]

not that I'm aware of

[obsidian mulch]

But you could probably also run if (whatever != null) in an update to make sure

[wintry crescent]

pretty sure I had similar issues with lists a looong time ago, so maybe that'll be the solution... although can't quite remember

[wintry crescent]

also a good option I guess

very weird that fixedupdate can run before start, that's funny

can it run before onenable? that could be a good idea as well

[heady iris]

You know, I was going to say LateFixedUpdate wouldn't make any sense, but it would be kind of handy to have code run after the physics update

[heady iris]

No, because OnEnable runs at the same time as Awake (immediately)

[obsidian mulch]

How can it run at the same time?

[wintry crescent]

I am pretty damn sure that from some testing I did, all Awakes run everywhere first, then OnEnable and Start run in said order per object

[heady iris]

as in, together

[obsidian mulch]

I get that it will run both on the same frame, but I’m pretty sure it will create and call references in Awake before it makes it to OnEnable

[heady iris]

not literally simultaneously

Awake and OnEnable execute together as a pair if the behaviour is enabled upon creation

[wintry crescent]

so
obj1.Awake
obj2.Awake
obj3.Awake
obj1.OnEnable
obj1.Start
obj2.OnEnable
obj2.Start
obj3.OnEnable
obj3.Start

that's how I remember it

[heady iris]

This is not the behavior I have observed.

[wintry crescent]

I might have it mixed up I admit

it was a while since I last checked it

[magic harness]

    public SecondaryAttributeModifier[] secondaryModifiers { get; private set } = new SecondaryAttributeModifier[10];

Json serializes correctly

but same thing, it deserializes the array with 10 null values

[wintry crescent]

but something similar to this was happening xd

[wintry crescent]

Oh that's significantly different!!! than an empty array!!!

[tiny delta]

Just checked, the reason awake has been causing problems is because there's info given to that other script when it's spawned in, and so running Awake actually runs before that information is given sometimes

[magic harness]

think i might have lost you here actualy

[leaden ice]

More like this:

obj1.Awake
obj1.OnEnable
obj2.Awake
obj2.OnEnable
obj3.Awake
obj3.OnEnable
obj1.Start
obj2.Start
obj3.Start 

[wintry crescent]

an array with 10 null values is a different problem than a null array

[heady iris]

Yeah. You often have problems like that

Awake is too fast. Start is too slow.

[wintry crescent]

it was a long time ago that I checked that xd ¯_(ツ)_/¯

[heady iris]

I'll often create an "Init" method that I manually call after doing some setup

[magic harness]

yeah, sure. But the underlining reason why it has 10 null values seems to be the same reason why the list wouldnt serialize

[tiny delta]

Yup, I've improved it a lot now by getting rid of like 90% of the stuff in my dumb CharacterInfo scriptable object

[wintry crescent]

wait, explain - was the list null and the array has null values?

[tiny delta]

It just has the stuff for input actions and starting positions

[wintry crescent]

can you show an example of an object that gets deserialized properly?

[tiny delta]

So maybe what I'll look at doing is moving the Start code to Awake, and then setting the starting position of that object in the other script

And I'm getting closer and closer to being able to get rid of the scriptableObject entirely

[magic harness]

the list was null because i wans't starting it on the property declaration, only in the constructor that actually takes parameters. and the array is empty and not null because i'm starting it on the class declaration

[magic harness]

sure. https://pastebin.com/51jpnpdK

thats my entire json file, as i said, im using a single entry point

inside it there are many objects being serialized correctly

including lists and polymorphic types

[wintry crescent]

uhhh am I seeing this right? all the primary/secondary attributes lists are empty or null in there...

[magic harness]

because thats one that has been loaded already and then saved again

let me send you another one then

https://pastebin.com/wDz8KaxW

check id 125 at line 660

[placid summit]

Hi I did not know this but the Mesh API with RecalculateNormals() has UV seam issues: Imported Meshes sometimes don't share all vertices. For example, a vertex at a UV seam is split into two vertices, so the RecalculateNormals function creates normals that are NOT SMOOTH at the UV seam!

I can understand from a speed perspective but they ought to offer a slower version that fixes this

[heady iris]

I know that using the "Mikktspace" option for calculating tangent vectors can cause meshes to be split along UV seams

(since it uses the UV map to figure out the tangent)

However, the resulting vertices should have identical normal vectors

[magic harness]

any idea?

[placid summit]

so this is posted in the forums, Unity will not blend across a UV seam as it is like a hard edge internally when 2 vertices share same pos/normal (in order to have different UVs) - oh well

known normal calculation issue

[heady iris]

Oh right, duh. It has to split across uv islands

[foggy pasture]

heya, I was hoping someone could help me out a bit with an issue I'm out of my depth on. I have a cinemachine camera (unrelated to this issue) that's tracking a target, and I have a pixel perfect component on said camera (my own, I ripped out the pixel position rounding component from the 2d one). as you can see in the video, when moving horizontally there's zero jitter, perfect. but when moving in the z axis, there's quite a bit of jitter, even though the code should be rounding it the same

    public Vector3 RoundToPixel(Vector3 position) {
        float unitsPerPixel = 1.0f / pixelsPerUnit;
        if (unitsPerPixel == 0.0f)
            return position;

        Vector3 result;
        result.x = Mathf.Round(position.x / unitsPerPixel) * unitsPerPixel;
        result.y = Mathf.Round(position.y / unitsPerPixel) * unitsPerPixel;
        result.z = Mathf.Round(position.z / unitsPerPixel) * unitsPerPixel;

        return result;
    }

    // Snap camera position to pixels using Camera.worldToCameraMatrix.
    void PixelSnap() {
        Vector3 cameraPosition = m_Camera.transform.position;
        Vector3 roundedCameraPosition = RoundToPixel(cameraPosition);
        Vector3 offset = roundedCameraPosition - cameraPosition;
        offset.z = -offset.z;

        // Get world to local camera matrix without scale
        var invPos = Matrix4x4.TRS(cameraPosition + offset, Quaternion.identity, Vector3.one).inverse;
        var invRot = Matrix4x4.Rotate(m_Camera.transform.rotation).inverse;
        var scaleMatrix = Matrix4x4.Scale(new Vector3(1.0f, 1.0f, -1.0f));

        // Calculate inverse TRS
        m_Camera.worldToCameraMatrix = scaleMatrix * invRot * invPos;
    }

i was hoping someone here might be able to provide insight into what i'm doing wrong here

a thought I've had is that since my outlines are depth and viewnormal based, it might be because the depth values are changing? I'm not sure

[low harbor]

Hi, is there a way so the input to result is such:
0 = (0,0,1)
1 = (1,0,0)
2 = (0,0,-1)
3= (-1,0,0)
Or, return in Vector3Int.forward, right, down, left ?
With sine ? Or...?

[heady iris]

Is the camera looking at the player at an angle? That could be relevant..

[foggy pasture]

it is, at a 30* tilt, which would make sense... adding a pan angle introduces x jitter

[rigid island]

create your own mapping?

    [SerializeField] private Vector3Int[] mapping = new Vector3Int[]
{
    new Vector3Int(0, 0, 1),
    new Vector3Int(1, 0, 0),
    etc
};

var dir = mapping[0] // gets 0,0,1

[heady iris]

I bet you need throw in a sin(30deg) or cos(30deg) factor somewhere

[knotty sun]

that looks like you want a Vector3Int array

[foggy pasture]

yeah I agree, let me mess around with that and I'll try and see, thanks

[heady iris]

I am working on a tool that needs to be able to run a set of "Filters". Some filters require no context at all (e.g. a filter that randomly says "yes" or "no"). Others require context (e.g. a position in the world).

I'm having some trouble figuring out how I'll structure this.

An existing implementation of mine has several abstract classes, such as:

public abstract class Filter {
  public abstract bool Test();
}

public abstract class PositionFilter {
  public abstract bool Test(Vector3 pos);
}

The nuisance here is that I want to be able to put Filters into a List<PositionFilter>. After all, Filter's Test method needs strictly less information than PositionFilter's Test method.

The current fix is to create classes that just hold an instance of a simpler filter:

public abstract class PositionBaseFilter {
  [SerializeField] private Filter filter;

  public bool Test(Vector3 pos) => filter.Test();
}

This works, but I'm wondering if I can do something to avoid all these "holder" classes.

Ideally, each class will ask for a specific set of information, and I'll be allowed to plug in any class that doesn't require information that my code can't provide.

I'm not sure if I can actually do this with C#'s type system.

the dream scenario would result in something like this:

• RandomFilter <-- no requirements, can fit anywhere
• PositionFilter <-- fits anywhere that can provide a "position" input
• PositionAndRotationFilter <-- fits anywhere that can provide both a "position" input and a "rotation" input

[plucky inlet]

Why not have a main class to inherit from or an interface?

[heady iris]

The nuisance is that this is the opposite of a classic inheritance hierarchy

If I made PositionFilter inherit from Filter, it would have this pointless Test method that doesn't work

[plucky inlet]

Why does Filter need the test() then?

[heady iris]

Because Filter represents a filter that can make a decision without any input

[plucky inlet]

So its a DecisionFilter, not a plain filter

[heady iris]

the point is that inheritance makes no sense here -- filters that require more inputs aren't a "subclass" of filters that require fewer inputs

You cannot use a PositionFilter like a Filter. It's actually the exact opposite

a Filter works anywhere that a PositionFilter does

some kind of inverse Liskov substitution principle

This also makes interfaces non-useful

[plucky inlet]

So, what do you want to achieve actually? You want to have a list of different types?

[heady iris]

Suppose I have some code that knows a position

It wants to make a decision based on that position

I want it to be able to run any kind of filter that can operate with only a position input

This includes:

• Filters that require a position
• Filters that require nothing at all

It excludes filters that require any other kind of information

You can do the check at runtime, of course, but that's not very satisfying

[vestal arch]

so i have a few disconnected trains of thought floating around

• this would require something akin to super in java (as in, generic constraints) which c# doesn't seem to have
• perhaps instead of requiring nothing (Filter) vs requiring Position vs requiring Position&Rotation, you could make it unary?
  • a nullary filter would just ignore the input, perhaps the input could be null
  • a unary filter would just have that input, directly?
  • a binary+ filter would have the inputs as a struct

along that latter train of thought; maybe you could have a unified FilterData type? that doesn't seem too great to work with though, static checking would be hard

maybe something via attributes...

[heady iris]

the FilterData idea is roughly what I've done in another place in my game

It's kind of gross. I have an interface for each kind of data that adds a property, and I downcast to those types to see if I can give them the information

Then, when the class does work, it checks if it got all of its information and throws an exception if that didn't happen

Obviously, static type checking goes out the window here

[vestal arch]

oh wait, super does exist, it's in

[heady iris]

Contravariance did come to mind

[heady iris]

But inheritance is fundamentally not compatible with this design. Deriving from a type implies that you can do everything the parent does

And that is not the case here.

[vestal arch]

if the parent is generic, perhaps that could work

[heady iris]

The "current fix" I mentioned here does work. I just have to explicitly implement every possible combination of "more specific filter" and "less specific filter"

[vestal arch]

but that goes back to the FilterData thing

ive been thinking about java.lang.Comparable

[plucky inlet]

I am just trying to wrap my head around what combinations you have and how you distinguish your filters as well as how do you applay them. where do you want to filter them out?

[heady iris]

imagine I have three arbitrary kinds of data: foo, bar, baz

One filter might need to know about foo, bar, and baz to be able to make a decision

Another might need just foo

And a third might need nothing at all

[plucky inlet]

But shouldnt your filters just get the data available and decide on their own?

[heady iris]

Any context where I can execute the first filter -- meaning I provide it with foo, bar, baz -- is also a context where the second filter and third filter are valid

[heady iris]

Right, I'm just hoping I can get some kind of static type checking here

But I think that's not possible

[vestal arch]

a given Comparable<T> can make a decision based on T
T can get more specific
in usage, U super T is used, so Comparable<Object> is assignable to Comparable<String> for example

what about having interfaces for each input data point?

[heady iris]

that's what I did here #archived-code-general message

[vestal arch]

i think you can get static analysis with this though

one sec

[heady iris]

I'm asking a negative question here

It's not "Does it implement X?"

it's "Does it ONLY implement X?"

Having one interface for each type also implies that you can ask questions based on exactly one piece of information

[vestal arch]

isn't that intended though

[heady iris]

No. If a filter demands foo, bar, and baz, you must give it all three of those to make a decision.

[plucky inlet]

but there could also be "does it ONLY implement X AND Y?" which ends in up in a problem of how you ask, at least feels like 😄

[vestal arch]

yeah ok i think i misunderstood the perspective

[heady iris]

public interface INeedsPosition {
  public bool Test(Vector3 pos);
}

public interface INeedsRotation {
  public bool Test(Quaternion rot);
}

public class SomeFilter : INeedsPosition, INeedsRotation {
  // ...
}

This doesn't fit my design.

[vestal arch]

ah no that's not what im talking about

[heady iris]

I want this filter to have one method that makes a decision based on a position and a rotation

And that just doesn't fit at all, yeah

[vestal arch]

IHasRotation and such, rather

[heady iris]

Yeah, interfaces work the other way

[vestal arch]

i think im explaining my idea poorly, i'll go back to making my example

[heady iris]

This has convinced me that I can't warp the C# type system into doing this 😛

[plucky inlet]

How about custom "tags" to assign and based on that Init the filter and do its local thing?

[heady iris]

That's roughly what I did here (through interface abuse) #archived-code-general message

[knotty sun]

Sounds to me like you need a base class to make a variable able to hold multiple class types then use reflection to find which Test method is implemented

[plucky inlet]

I also wonder, do you want to run the filters runtime but prefiltered on editor? or do you want to filter them out in runtime?

[heady iris]

I'm setting these up in the inspector

[plucky inlet]

Ah got it.

[heady iris]

Which does mean I can just filter the list of choices based on, say, attributes

That's probably going to be the way forward

[plucky inlet]

So you could just reflect to the methods available and add them to a custom event to call those functions?

[heady iris]

I'm envisioning something like this

[plucky inlet]

Thats what I did to call stuff on scripts from UI and not to have tie up everything. Just select the type in inspector, update the dropdown to add the functinos to call and then just call the singleton of it and fire those methods.

[heady iris]

[Provides(typeof(FooInformation)), SerializeReference] List<Filter> filters;

The inspector would only show me classes that don't require any other kind of information.

Filter would wind up looking like this:

public abstract class Filter {
  public abstract bool Test(FilterData data));
}

and FilterData would contain all the possible things you could give to a filter

So it would be perfectly legal to run a filter without all of the data that it needs

But the inspector would forbid me from assigning such filters in the inspector

[tulip spruce]

How do I fix that Don'tDestroyOnLoad does not keep deavtivated gameobjects?

or how do I work around that

[latent latch]

whatcha mean

[leaden ice]

Don'tDestroyOnLoad does not keep deavtivated gameobjects?

What do you mean by that?

[latent latch]

I don't remember that behaviour

[leaden ice]

that's not a thing

[plucky inlet]

How would that prevent you from assigning?

[latent latch]

Oh I think what you may mean is that because the object is deactived it may not be flagging DDOL

[tulip spruce]

deactivated objects are not shown in "DontDestroyOnLoad" in the UI

[heady iris]

I already use a package to support SerializeReference (it's vertx's package, actually)

[leaden ice]

I don't think it matters if they're active or not. You have to have called DDOL on them to put them there.

Show your code and what's happening exactly

[heady iris]

I would throw out any types that need information I'm not giving

[tulip spruce]

😄

[latent latch]

is the object ever calling start^

[leaden ice]

Start won't run if it's not active

so DDOL is not the problem

[tulip spruce]

oh

[plucky inlet]

Ah got it, okay, makes sense

[tulip spruce]

should it be in Awake?

[leaden ice]

Awake won't run if the object is not active either

The problem is trying to run code on a script that isn't active

[tulip spruce]

do I give up or how do I make it run?

[leaden ice]

What are you trying to do

put the code on an active object is the solution

[tulip spruce]

oh

I did not think about that

big brain

thanks

I am still learning

[leaden ice]

All of us are

[vestal arch]

i can't even warp the java type system into doing this

[leaden ice]

Is there a TL;DR explanation of what you're trying to do?
I've been messing around with Roslyn Source Generators for a few weeks now and am convinced those are the solution to almost every problem at this point 🤣

[plucky inlet]

@leaden ice

It all started from there 😄

[heady iris]

I want to be able to plug less-specific filters into places that support more-specific filters

If I can provide ABC, then filters that need ABC, or AB, or C, or nothing at all should all fit

But filters that need D shouldn't fit

[leaden ice]

Are static/compile time errors a requirement here?

Or would runtime errors be ok

[heady iris]

I'd prefer the former. I can already achieve this (with some headaches) if I just check at runtime

[leaden ice]

So what's the typical use case look like?

Something like... uh...

myThing.Filter(new FilterA());```?

[vestal arch]

(i am still trying btw. i think i broke the java playground...)

[leaden ice]

Where myThing is of some specific type that should accept certain filters and not others?

[heady iris]

The typical form is something like this

wait, typing code on my ipad sucks

it sends the message early like 50% of the time

Suppose I'm spawning prefabs along a line

I want to be able to reject certain placements based on the chosen position

[leaden ice]

This could probably be done with a Roslyn Analyzer btw

[heady iris]

But I should also be able to reject placements based on nothing at all (e.g. a coinflip)

[leaden ice]

which lets you add new rules to the compiler basically

[heady iris]

also, this particular thing is editor-only -- but I've run into identical problems in runtime code

But in both cases, I'm setting everything up in the inspector in advance

so I can get away with a custom editor that enforces the rules

[plucky inlet]

Could you explain a bit more, how you add those filters. or rather, to what? its in inspector, right? so you have some kind of serialised list in inspector and you want to give the list some kind of property, so it knows what iflters to accept?

[leaden ice]

I'm not even sure "type information" is enough to work with.

Knowing the filters operates on a Vector3, for example, doesn't mean the filter works on any Vector3

it might take only positions

or euler angle rotations or something

[plucky inlet]

@heady iris sorry, was meant for you

[heady iris]

Yeah, it'd be more than just the literal type of the argument. That's okay

[vestal arch]

wait i might have a system going in java...

[heady iris]

[Serializable]
public abstract class Scorer : IPolymorphic
{
    public abstract float GetRawScore();
}

[Serializable]
public abstract class EntityScorer : IPolymorphic
{
    public abstract float GetRawScore(Entity entity);
}

[Serializable]
public class EntityConstantScorer : EntityScorer
{
    [SerializeReference] private Scorer scorer;

    public override float GetRawScore(Entity entity)
    {
        return scorer.GetRawScore();
    }
}

Here's an example of what I've done before.

I'll store a List<EntityScorer> on something that needs to compute a score based on an entity

Rather than also having a List<Scorer>, I created EntityConstantScorer

That lets me fit a Scorer into a list of EntityScorer (through a wrapper)

[plucky inlet]

phew, that feels hacky 😄

[tulip spruce]

I did another approach, because I had issues referencing the gameobject in the other gameobject, so what I did was have the other gameobject active but the canvas turned off. then the code will run so DDOL will run and then I enable the canvas back and deactivate the gameovject. and it worked. I am doing a lot wrong. ik

[leaden ice]

I really don't feel like I have enough context to answer if you're doing anything wrong here but it sounds like the code is now on an active object, which is exactly what I suggested.

[tulip spruce]

oh, ok

thanks for the help anyway

I would be helpelss else

[heady iris]

It's tolerable so far because I haven't had a need for too many kinds of scorers

and they don't overlap much in what information they get

The punchline might be that this problem doesn't exist at all 😉

i.e. instead of having filters like:

• ABD
• BD
• ABCD

I just have:

• ABC
• AB
• A

[plucky inlet]

I think we all knew almost form the start of this conversation, but eager to get the perfect solution 😄

[heady iris]

with very little crosstalk

Yeah :p

I've banged into this several times and dealt with it in varying ways

It was very helpful to talk about it! Thanks!

[plucky inlet]

Oh, thank you too. Very challenging and was fun to think about that whole structuring , great insightful conversation 🙂

[heady iris]

If I do run into a situation where I have lots and lots of overlap, I'll just do a runtime check

[plucky inlet]

If you have ABC, AB, A, you could even just use an enum. But as soon as you have a mixture of those, it wont work

[still jungle]

u can use enum bitmask to combine

[plucky inlet]

I just use enum to control roles for example. the higher the enum value, the higher the role and less things are restricted or what not

[heady iris]

This does remind me of "blackboards" in the context of game AI

where you have a shared pool of information and you can't really be sure what will be in it

[vestal arch]

!code

[tawny elk]

Posting code

📃 Large Code Blocks

Use links to services like:
https://paste.mod.gg/, https://hastebin.skyra.pw/, https://paste.ofcode.org/, https://paste.myst.rs/, https://scriptbin.xyz/

📃 Inline Code

Surround code with three backquotes. Not quotation marks.
To format as C#, add cs to the first line:
```cs
// Your code here
```
Add a comment with a line number if there is an error message.

[vestal arch]

this is as far as i got with java (im more familiar with it)
https://paste.ofcode.org/M3ueVR5kQEKKb5qhagaxmE

can't construct multiple-datapoint-filters with this, T & U can't be used as a type like that, unfortunately...

[heady iris]

I bet Typescript can do some bullshit here

(as it always can)

[vestal arch]

of course it can lmao

as everywhere 👍

(maybe i'll try that... (like actually))

[limpid siren]

hello, I'm using this code to force a 4:3 aspect ratio on the game I'm working on
it works well on a 1920x1080 laptop screen, but I got a friend to try the game on steam deck, and the aspect ratio change did not work. The game just filled the screen.
Any idea how I could fix this?

[leaden ice]

How about just something like...:

public class Context {
  public Vector3 position;
  public Entity entity;
  public Color color;
}

public abstract class AbstractFilter {
  public abstract bool Filter(Context ctx);
}

public class PositionFilter : AbstractFilter {
  public override bool Filter(Context ctx) {
    return ctx.position.x > 5;
  }
}```
And just let the filters optionally take whatever data they want from the Context object

[vestal arch]

typescript is just 2 programming languages on top of each other with one masquerading as a type system

[tulip spruce]

is it possible to change something in the inspector when the game is running and have those changes be persistent?

[vestal arch]

i believe we did discuss this briefly, it loses the part about static checking what data is available

[heady iris]

I know that Cinemachine does this.

[plucky inlet]

yes, I think there is (was) an option to keep values somehow, but that can mess up a lot of things. you can always right click copy component and paste when in edit mode

[tulip spruce]

Understood.

Thanks

[heady iris]

That's what I'll do if I just need one component

[wheat spruce]

private void AddMovementForce()
{
  Vector2 movementForce = _moveDirection * _speed;

  float angle = Vector3.Angle(transform.up, _wheels.TurnDirection);
  Quaternion rotation = Quaternion.Euler(0, 0, -angle); // take the

  Vector2 finalDirection = rotation * movementForce;
  _rb.AddForce(finalDirection, ForceMode2D.Force);
}```I'm trying to make it so that the direction the object at the front travels becomes influenced by the direction that the back object is facing. I've tried a few different things, and nothing seems to do the right thing

Something more like this

_moveDirection is the way the user aims the front half of the train, but its the back wheels that need to influence some steering

[vestal arch]

@heady iris in ts it Just Works™️ lmao
playground

[plucky inlet]

yeh, typescript, python, just throw in whatever you want and get whatever you want 😄

[heady iris]

Yep! & is the secret sauce there

[heady iris]

One immediate problem

Vector3.Angle is not signed

So you can't tell left from right

[vestal arch]

java kinda has it but just doesn't allow it as broadly 😔

[heady iris]

Consider using Quaternion.FromToRotation

This gives you a rotation from one direction to another

That might misbehave for an exact 180 degree mismatch, though

In that case, Vector3.SignedAngle would be more reliable. It needs a third direction as a reference point

Vector3.SignedAngle(one, two, Vector3.forward);

This would be appropriate.

SignedAngle gives you a number in the [-180..180] range. I can never remember which way is negative

[wheat spruce]

float angle = Vector3.SignedAngle(_moveDirection, _wheels.TurnDirection, Vector3.up); Are these the right properties to use in it? I've tried a number of combinations and none seem correct

[heady iris]

You need forward, not up

The third argument is the axis that you're rotating around

[wheat spruce]

ooooohh right

[heady iris]

in a 2D game, forward points into the screen

You might want to calculate two signed angles relative to some arbitrary reference point

float moveAngle = Vector3.SignedAngle(_moveDirection, Vector3.right, Vector3.forward);
float wheelsAngle = Vector3.SignedAngle(_wheels.TurnDirection, Vector3.right, Vector3.forward);

float skewedAngle = Mathf.LerpAngle(moveAngle, wheelsAngle, 0.5f);

Vector3 resultDirection = Quaternion.AngleAxis(skewedAngle, Vector3.forward);

I think that'll do it.

Note that it'd be dramatically simpler to just do

Vector3.Slerp(_moveDirection, _wheels.TurnDirection, 0.5f);

I just don't know what's going to happen if those two vectors are antiparallel

(i.e. exactly 180 degrees apart)

[wheat spruce]

    float moveAngle = Vector3.SignedAngle(_moveDirection, Vector3.right, Vector3.forward);
    float wheelsAngle = Vector3.SignedAngle(_wheels.TurnDirection, Vector3.right, Vector3.forward);
    float skewedAngle = Mathf.LerpAngle(moveAngle, wheelsAngle, 0.5f);

    Quaternion angleAxis = Quaternion.AngleAxis(skewedAngle, Vector3.forward);
    Vector2 finalDirection = (Vector2)(angleAxis * Vector3.up);```I wasnt able to use the Quaternion how it is at the end, is this right?

[heady iris]

You'd want to multiply with Vector3.right at the end

since that was your reference point

[tulip spruce]

if I have an event trigger to play an audio on PointerClick, and a OnClick Playgame (it is changing the scene) will the audio not play then?

[plucky inlet]

you mean the audio gets destryoed on scene change?

[wheat spruce]

it ends up doing this

[tulip spruce]

does it?

[plucky inlet]

You can test this in your unity editor 😉

[heady iris]

if the audio source goes away, the audio will stop playing

[tulip spruce]

I did and I can't hear anyting

[tulip spruce]

Don't tell me I need to put my buttons in DDOL?

[heady iris]

not the buttons -- but what about a single audio source that's used by all of your buttons?

[tulip spruce]

ah I see

you are good at this

[heady iris]

is it abruptly turning to the right here?

i can't tell what the issue is

[wheat spruce]

I'm not pressing anything, it just propels to the right

are you sure vector.up isnt correct? It would be (0,1,0) and its the Y axis that the front of the train heads towards

[heady iris]

Vector3.forward is a world-space direction here. It's an axis pointing into the screen

you're finding the angle between your vehicle's move direction and a reference point

(which is Vector3.right)

that angle is measured around the forward vector

[heady iris]

that's a completely separate issue though

all this is doing is finding a direction that's halfway between _moveDirection and _wheels.TurnDirection

if _moveDirection is zero it could get a bit weird

[wheat spruce]

    Vector2 movementForce = _moveDirection * _speed;
    Vector3 finalDirection = Vector3.Slerp(movementForce, _wheels.TurnDirection * new Vector2(-1, 0), 0.5f);
    _rb.AddForce(finalDirection, ForceMode2D.Force);```Quick thing to note, I did give the lerp a try. It sort of works, but at the start of the video I try and press A and D to turn the front and it doesnt move

it only turns when the back gets tilted

[heady iris]

what's with the scaling on TurnDirection?

[wheat spruce]

to invert the direction

[heady iris]

that's throwing out the Y component entirely

[wheat spruce]

yeah

[still jungle]

late to the party but this approach may be helpfull

[wheat spruce]

I'll work on it more tomorrow, its past 8pm so I'm going to stop working for the evening

will give me time to reflect on the code, I think in general the steering part of the code doesnt do quite the right thing

(also havent eaten anything today so I'm probably running on empty 😅)

[plain moon]

Why without animation my character is under root object with colliders? Child object(model) have (0,0,0) coords its not offset or smth else. Just weird thing

[leaden ice]

You need to set this to "pivot"

Then you can see the actual positions of the objects when you select them

It will probably show you that the center of the model is offset

this is a problem in the 3D model itself

You would fix it in Blender (or any other 3D modeling software)

[plain moon]

Set by pivot and this happened 🙂 Download model from mixamo

[tulip spruce]

if I spam onenter sound many times all sound in the scene is gradually being muted and then if I stop it gradually comes back, is that supposed to be like that?

[leaden ice]

Yes now you can see where your object actually is

See how the collider is super offset from the model?

You should adjust the Center of your collider so the pivot is starting from the bottom like the model

[plain moon]

Yes, but can I fix this somehow without blender and etc. I would like to use Root Motion anims. Would it be compatible with pivot in center of model?

[leaden ice]

Add an empty parent and offset the model

but if one of the animations is completely offset from the others there's not much you can do about that

It's generally much better IMO to have the object's pivot on the ground rather than the center - makes code much simpler.

[plain moon]

That's how it works now. Root is rigidbody and colliders, first child animators and etc

[leaden ice]

The colliders can always be offset by either:

• simply adjusting the center property in the inspector
• moving them to child objects

you can also just offset the visuals separately

you can do whatever you'd like inside that hierarchy more or less

[tulip spruce]

It seems like when I have the exit to menu button (works by onClick: SceneManger.LoadScene(buildnumber_toMainMenu)) But when I do this the music starts to play again

music that I have in main menu

what would be the best apporach to fix this

[leaden ice]

What do you want to happen instead?

Shouldn't the main menu music play when you start the main menu?

[tulip spruce]

yes but since I put DDOL on the music

oh

nvm

I need to put another music in the game instead which is not DDOL

[leaden ice]

if you have DDOL on the music playing object you should make the code ensure there is only one copy of it

the pattern is a "DDOL singleton"

If you want to simply have one in each scene, you wouldn't use DDOL at all

[tulip spruce]

Yeah I figured that would be best

I don't do DDOL if I not have too

[heady iris]

so I've got an interesting idea that I'm implementing now

it's a bit like this

public class SimplexNoiseFilter : Filter
{
    [SerializeField, SerializeReference, SerializeReferenceDropdown]
    private PositionSource positionSource;

    // ... more stuff
 }

PositionSource is another abstract class that tells you how to get a position

One position source just grabs the "position" member from a FilterData object that gets thrown into the filter's Test method

Another gets a position from a TransformSource!

...one kind of which, in turn, can look to see if you provided a "root" member in the FilterData

this also leads to a hilarious function signature

public abstract bool TryTest(FilterData data, out bool result);

[heady iris]

hang on, this is just dollar store dependency injection

and more layers of indirection than a main-story fetch quest

[tame summit]

Hey, I'm not sure where to ask this but im trying to make a car with the wheel colliders and it has low acceleration but fast top speed and im looking to make it a little more realistic

I know very little about wheel colliders but I would love to learn more about it

[dusk apex]

Maybe lookup the Internet for some tutorials on it. After doing so and getting stuck or having questions for clarity, ask #💻┃unity-talk
If it's a general programming question, you can ask here.

[tame summit]

ok, thank you. no harm ment

[keen turret]

Hey, I'm trying to get my simulation from c# to a compute shader, it's an implementation of an XPBD solver, distance constraint of the edges and volume constraint of the tetrahedrons, but it looks bad and jitters. Even ensuring no parallel execution so there's no race condition with the numthreads set to 1 (I don't know if this really ensures that it doesn't run in parallel, because it's the first time i write a compute shader in my life). I also tried to compute the rest length and rest volume of the constraints in the gpu to make sure it's not caused due to a difference in floating point accuracy between the cpu and gpu, nothing works, the simulation is not totally broken, it's jittery, This is an example of the distance constraint in c# and on the compute shader:

public void Solve(NativeArray<Particle> particles, float deltaTime)
{
    float alpha = compliance / (deltaTime * deltaTime);

    Particle pA = particles[indexA];
    Particle pB = particles[indexB];
    float w0 = pA.InverseMass;
    float w1 = pB.InverseMass;
    float w = w0 + w1;
    if (w == 0f)
        return;

    Vector3 grad = pA.Position - pB.Position;
    float len = grad.magnitude;
    if (len == 0f)
        return;

    grad = grad / len;
    float C = len - restLength;
    float s = -C / (w + alpha);
    pA.Position += s * w0 * grad;
    pB.Position += -s * w1 * grad;
    particles[indexA] = pA;
    particles[indexB] = pB;
}

the compute shader:

[numthreads(1, 1, 1)]
void SolveDistances(uint3 dispatchThreadID : SV_DispatchThreadID)
{
    DistanceConstraint constraint = distanceConstraints[dispatchThreadID.x];
    
    Particle pA = particles[constraint.indexA];
    Particle pB = particles[constraint.indexB];
    
    float w0 = pA.InverseMass;
    float w1 = pB.InverseMass;
    float w = w0 + w1;
    if (w == 0.0f)
        return;

    float3 grad = pA.Position - pB.Position;
    float len = length(grad);
    if (len == 0.0f)
        return;

    float3 norm = grad / len;

    float C = len - constraint.restLength;
    
    float alpha = constraint.compliance / (deltaTime * deltaTime);
    float s = -C / (w + alpha);

    pA.Position += s * w0 * norm;
    pB.Position += -s * w1 * norm;
    
    particles[constraint.indexA] = pA;
    particles[constraint.indexB] = pB;
}

[keen turret]

the rest of the code is here: https://paste.mod.gg/jqvmdmgahcgb

[cosmic rain]

(Compute) shaders will run in parallel no matter what you do. It's their nature and what makes them fast.
Numthreads just defines the number of threads in a thread group.
As for the jitter, can you record a video?
How are you executing the shader and getting the results? Are you sure it's caused by the compute shader?

[keen turret]

Solved, the problem was exactly that, instead of trying to dispatch the shader one time, I dispatch it for every parallelizable group of constraints that i generated using the graph coloring algorithm so no race condition occurs, thank you

[pine condor]

Does anyone know how to spawn a player prefab like with DefaultPlayerPrefab at runtime in unity transport unity 6?

[leaden ice]

in unity transport unity 6
wdym by this? Netcode for GameObjects?

https://docs-multiplayer.unity3d.com/netcode/current/basics/playerobjects/
https://docs-multiplayer.unity3d.com/netcode/current/basics/object-spawning/

[pine condor]

uh... how do i get the clientid?

[leaden ice]

the clientID of what

[pine condor]

whatever defaultplayerprefab uses idk

[leaden ice]

I don't understand the question.

[pine condor]

it's asking for a client id

[leaden ice]

Player objects get spawned in as soon as the client connects

[leaden ice]

what is?

What are you trying to accomplish here?

[pine condor]

To manually spawn an object as PlayerObject, use the following method:

GetComponent<NetworkObject>().SpawnAsPlayerObject(clientId);

[leaden ice]

Just use your own client id (assuming this is from the client)

[pine condor]

where do i find that?

[leaden ice]

On your NetworkManager of course

https://docs.unity3d.com/Packages/com.unity.netcode.gameobjects@2.2/api/Unity.Netcode.NetworkManager.html#Unity_Netcode_NetworkManager_LocalClientId

In the future: #archived-networking

[desert blade]

How does one get around UnityWebRequest.PostWwwForm(url, JsonUtility.ToJson(payload)) spitting out ascii on the receiving end? Server keeps coming back to me with %7b which is ascii for {

WWWForm() doesn't work though cause that needs to be paired with a content header of x-www-url-encoded .

I find it weird that unity's default post is a x-www-url-encoded but their service api docs only accepts application/json

[leaden ice]

Server keeps coming back to me with %7b
Not sure I understand what you mean exactly here - you mean your code on the server side is receiving that data? Or the server is responding with that?

Also what are you trying to do exactly?

[desert blade]

requesting allocations with multiplay requires me to use the TokenExchange API with a payload of projectID and environmentID.

created a class ExchangePayload with the previously mentioned requirements as variables. plugged it into a JsonUtility.ToJson() and slapped that as the form for the UnityWebRequest.Post(url, payload)

Debugging the JsonUtility.ToJson() showed me that it was converting the class to json just fine. Sent it over to the server only to get back a unable to parse body. %7b... is not a valid json

[leaden ice]

Well - pretty sure PostWwwForm is going to URLEncode the data at the very least

Are you sure the server is expecting form data?

It's weird that you would be using both HTTP Form data AND json 🤔

does this API have documentation?

[desert blade]

https://services.docs.unity.com/docs/service-account-auth/index.html

[desert blade]

its expecting application/json

unless i've butchered this somehow

[leaden ice]

then you shouldn't be using PostWwwForm

[rigid island]

just make regular UnityWebRequest Post with that payload

[desert blade]

hmm its screaming at me because of the obsolete function

[leaden ice]

https://discussions.unity.com/t/posting-json-through-unitywebrequest/667630/3

SOmething like this should do

[rigid island]

which one ? UnityWebRequest is not obsolete at all

www class is

[leaden ice]

just that overload is obsolete

[desert blade]

wOt

[leaden ice]

try like in the link I shared

[rigid island]

you have to set the headers like the link ^

here is the Post that works if you want to do it all in one line
https://docs.unity3d.com/6000.0/Documentation/ScriptReference/Networking.UnityWebRequest.Post.html

its 3 args

[desert blade]

thanks a lot guys for helping out @leaden ice @rigid island

[azure frost]

My bro and I are having trouble with the Kinematic Character Controller again.

if(InputMod.GetButtonDown(CrouchInput)) {
                characterInputs.Crouch = (!characterInputs.Crouch);
                Debug.LogError("We press crouch and are now " + (characterInputs.Crouch ? "crouching" : "standing"));
            }```


```C#
// Crouching input
                        if (inputs.Crouch)
                        {
                            Debug.LogError("Be crouching");
                            _shouldBeCrouching = true;

                            if (!_isCrouching)
                            {
                                _isCrouching = true;
                                Motor.SetCapsuleDimensions(0.5f, CrouchedCapsuleHeight, CrouchedCapsuleHeight * 0.5f);
                                MeshRoot.localScale = new Vector3(1f, 0.5f, 1f);
                                
                            }
                        }
                        else if (!inputs.Crouch)
                        {
                            Debug.LogError("All rise");
                            _shouldBeCrouching = false;

                        }```
Segments taken from two scripts, and we're trying to make a "toggle crouch". What keeps setting "_shouldBeCrouching" to false, causing the character to only remain crouched for one frame?

[latent latch]

if (inputs.Crouch)
{
    Debug.LogError("Be crouching");
    _shouldBeCrouching = true;

    if (!_isCrouching)
    {
        _isCrouching = true;
        Motor.SetCapsuleDimensions(0.5f, CrouchedCapsuleHeight, CrouchedCapsuleHeight * 0.5f);
        MeshRoot.localScale = new Vector3(1f, 0.5f, 1f);
        
    }
}```
Why wouldnt you be crouching if input.Crouch here is to be true? Feels like you've one too many flags going on

[azure frost]

@arctic sparrow

[arctic sparrow]

public void SetInputs(ref PlayerCharacterInputs inputs)

Here's the problem. I have never used ref in any of my code, but I can only theorize that it's responsible for resetting the crouch bool

[azure frost]

What's your take on this, @latent latch ?

[latent latch]

ref keyword is used to send value types into methods as references, so unless PlayerCharacterInputs is a struct then it's probably not needed... but you should probably understand what yall are doing with it anyway

[arctic sparrow]

It is indeed a struct.

[latent latch]

As for the code in general, it would probably be more helpful if it was posted via paste site cause I can't figure the flow of it with what's there.

May want to look into a statemachine and eliminate a lot of those flags

[thin aurora]

Pls don't use ref, ever

If you don't know what it does, it's not a good idea to use it

[latent latch]

ref is fine if you're passing structs into methods

otherwise just make classes

[thin aurora]

If you have to pass structs by reference and you don't know what ref is then that's a good candidate for refactoring in general

It also ruins the idea of a struct

[azure frost]

https://hastebin.skyra.pw/ulamobamuy.csharp
ExamplePlayer

https://hastebin.skyra.pw/elefacuged.csharp
ExampleCharacterController

[thin aurora]

Why do you use ref here, though?

Ah, more code

[azure frost]

That's the way KinematicCharacterController had it.

[thin aurora]

You don't mutate the value, why did you add ref here?

You can just pass it by value

[arctic sparrow]

That's how it was originally. Seek is still learning how this all works

[thin aurora]

Then I would just remove it

If your other script requires mutation of this object then I suggest you get a reference to the class directly and mutate it from there

There's very little reason to actually use ref

Not because it's "bad". ref is very specific and generally you should just use another approach

I think Mao generally gave a good answer too. You probably just want to use a class here.

[latent latch]

Ok, so this is probably what's happening. Every frame it's making a new Input struct

and because it's being read every frame, Crouch is being acted upon even if the value isn't being set explicitly by the player

[quartz folio]

the same thing would happen whether it was a class or a struct because it's just being newed every frame

[latent latch]

right, so the bool value is by default false and that's what's being read into the input

[quartz folio]

store the state of the crouch at the ExamplePlayer class, and set a toggled version of that to the input struct

[arctic sparrow]

So if we only initialized the class/struct once, then we could have the variables be persistent?

[quartz folio]

The design of the example input is stateless. You have to add your own state, just set the crouch input based on a bool you yourself control at the class level

[latent latch]

KC is pretty good, but such a pain to get it working with cinemachine cause there's a lot of it is hardcoded for the basic main camera

[azure frost]

Okay, it is fixed!

[visual pollen]

Hey, wanted to ask for some tips or ideas on something im trying to do.

My current goal is to have a slider where I can drag, and if it reaches max or min value, it can loop around. Visually the slider will be radial in nature. Any ideas how I should go about the looping aspect? I dont need direct code, just ideas or opinions on how to tackle the issue

[fiery spruce]

Looping as in it loops back to min if it goes past max? If so simply value %= max; value = Mathf.Max(value, min)

[thin aurora]

https://docs.unity3d.com/6000.0/Documentation/ScriptReference/Mathf.Repeat.html

[fiery spruce]

Didn’t realize that’s a thing, cool

[visual pollen]

Aight thanks! Ill test both out! Also agreed, never knew Mathf.Repeat was a thing! Cheers you lot

[copper glade]

I'm developing a 2D game, primarily focused on UI, and I'm looking for a better way to save and load game states. Currently, I have a lot of UI elements and game objects that need to be saved. My current approach is to manually serialize each script's data to JSON and then load it back when the game resumes. This works, but it's very cumbersome and requires manual updates to the saving and loading logic every time I change my UI or game logic.

My game consists of several mini-games, so I end up having separate saving/loading systems for each one. I'm hoping to find a solution that's more automated and less error-prone. I'm working offline and only need to save locally.

Is there a package or resource you would recommend for saving and loading complex 2D UI scenes? I'm looking for a solution that avoids me having to write all the saving and loading logic manually.

Any help would be appreciated!

[plucky inlet]

You could generate a unique ID when first adding the script and then save its ID and state to a static global list and save that list as JSON.

And when then loaded, the script will just onenable check itself inside that list and get its values from there

[knotty sun]

Sounds like a job for my Save for Unity asset

[steady bobcat]

every time

[copper glade]

which one?

[knotty sun]

without knowing exactly what you are doing I'm guessing the Core asset and the UI Add-on will do the job
You might want to look at the Core+ bundle as it has a 50% New Release discount atm

[azure frost]

So, in the Kinematic Character Controller, where in the scripts is the value that determines which direction the player is facing?

[arctic sparrow]

Seek, I literally said to follow the movement values from the inputs class.

[wintry crescent]

What are some recommendations from people here, for serialized dictionary packages?

I need a good one, and there's plenty, but I don't know how to choose

and last time I tried to find one myself, I couldn't find a good one

ideally not paid, or at least cheap

[thin aurora]

As in a package that allows you to serialize a dictionary?

I assume this is binary serialization or just being able to show them in the inspector?

[wintry crescent]

show them in inspector, specifically

[trim schooner]

IIRC Odin does it, and that's the best

[thin aurora]

I can imagine this is a pretty common problem which has a free solution

[wintry crescent]

isn't odin insanely expensive

[trim schooner]

dunno, go look

[thin aurora]

First Google search: https://assetstore.unity.com/packages/tools/utilities/serialized-dictionary-243052

[wintry crescent]

I don't want a first result from google search

[leaden solstice]

I wouldn’t really recommend Odin for it

[wintry crescent]

last I tried that it ended up shit

and had to remove it

I want a recommendation

can't count on google for that

not these days for sure

[thin aurora]

Lol

[wintry crescent]

regarding that ? emoji - google went to shit

[thin aurora]

At least try it before you judge

[wintry crescent]

I tried a bunch a while ago

[thin aurora]

Did you try this one?

[wintry crescent]

and got tired searching for a solution that didn't require modifications to work and wasn't crap

I dunno it was a while ago, I am asking for a recommendation from someone who uses a serialized dictionary package

so that I don't have to try 20 of them

[thin aurora]

This package averages nearly 5 stars. I would just try it

[leaden solstice]

Odin requires whole different serialization for it, means impossible to get rid of it

[wintry crescent]

I'm not asking for a lot, please I just want a recommendation from someone who actually uses one

[azure frost]

https://hastebin.skyra.pw/ahivivagig.csharp
Okay, Chirz and I are having trouble making a "Recenter Camera" input.

[wintry crescent]

please dont recommend me random things you find on google

please

if you don't use one, please

I am not looking for the top 5 google results right now

because I already tried a bunch and they were crap

so I'm asking for someone who knows what they're talking about

to recommend me something

[leaden solstice]

What is your usecase and definition of “good one”

[wintry crescent]

it's subjective, and I'm asking for a subjective recommendation

[leaden solstice]

Just answer?

[wintry crescent]

holy crap friends this is insanely unproductive I am just asking for someone who is already using one to say what they're using and if it's any good

that's really not much to ask for

[thick terrace]

i'm pretty sure 99% of people are using that first one off google lol

[fleet gorge]

I make my own serialised dictionary

[leaden solstice]

So hard to say what is your detailed usecase and what you want

[wintry crescent]

yea is it any good

[leaden solstice]

Lol

[wintry crescent]

which ones have you tried?

[thin aurora]

Generally Google results by relevancy so the first few results are the best ones

[wintry crescent]

I'm looking for people who tried them I said it like 10 times now

[fleet gorge]

I already tried the top results, I ended up making my own

Doesn't take very long

[leaden solstice]

Just stuff them in List

[wintry crescent]

that's exactly my point!!! top results aren't the best

[leaden solstice]

Load as Dictionary at some point

[fleet gorge]

There's no other libraries

Coming from someone who's already tried it

Just diy

[thin aurora]

Which is incredibly unspecific

What's wrong with this one? https://assetstore.unity.com/packages/tools/utilities/serialized-dictionary-243052

You instantly shoot down suggestions, but generally these work fine

[wintry crescent]

this is the type of thing I'm looking for
someone who also looked for it and went through the pain
Ideally I'd find a premade package but if it's not possible then I guess I gotta make one myself although I wanted to avoid it

[trim schooner]

There are loads of reviews on that one for you to get a wider range of opinions on it too

[fleet gorge]

yea I'll just share my experience

[wintry crescent]

do you understand why I came here
I am not looking for a suggestion
I can google, I know how to do that
I'm looking for a recommendation from someone who already uses one
that's different

you don't have to understand my thought process

[fleet gorge]

my experience is that all the packages do too much

it's not worth asking around

everyone makes their own

[azure frost]

Can somebody help with this?

[trim schooner]

I'd wager more people use a premade one than make their own

[trim schooner]

see below !ask

[tawny elk]

:thinking: Asking Questions

:mag: Search the internet for your question!
:book: Use the API Scripting Reference and User Manual and this troubleshooting site for commonly posted issues.
:wrench: Attempt to debug your issue.
:thought_balloon: Find an appropriate channel by reading the name and description in #🔎┃find-a-channel
:grey_question: And don't ask to ask, ask a full question illustrating with screenshots if needed.
-# For more posting guidelines, go to #854851968446365696

[fleet gorge]

How may we help?

[leaden solstice]

I have no idea who the hell is Chirz

[plucky inlet]

What works for others might not work for you, just saying. so even a recommendation from someone else can give you personally a bad result in your system. So you HAVE TO try it anyways. No matter your thought process or why you came here, its the way you have to add features (package in this case), test, try, get a result, adapt

[azure frost]

Well, Chirz (@arctic sparrow ) was trying to use the player's current orientation to recenter the camera with. But it didn't seem to work no matter what we set the input to.

[fleet gorge]

Which function does this currently?

What do you have so far

[azure frost]

I posted the script here.

[fleet gorge]

I don't see any code for changing the orientation of the camera

[plucky inlet]

I guess, the missing part is inside this method... CharacterCamera.UpdateWithInput(Time.deltaTime, scrollInput, lookInputVector);

[azure frost]

https://hastebin.skyra.pw/ekibaqosih.pgsql

Here's ExampleCharacterCamera

[plucky inlet]

Id say, add a bunch of debug.logs first and see, if all your if clauses are actually working as they should

[arctic sparrow]

if(InputMod.GetButtonDown(CenterCamInput)) characterInputs.CameraRotation = CharacterCamera.Transform.rotation = Character.transform.rotation;

Wait, I think we might be looking at it the wrong way. It's in the HandleCharacterInput function instead of the HandleCameraInput function

[plucky inlet]

And what are those random brackets?

[leaden solstice]

The scope

I guess

[plucky inlet]

Phew, hurts me personally, but that might just be my preference to not to 😄

[fleet gorge]

you guess?

who's the one who wrote this code

on a scale of 1-10 how much do you understand the code you wrote

[plucky inlet]

its from the question from chirz and seekingthesky

[leaden solstice]

Not me, apparently

[fleet gorge]

ngl it looks like chatgpt code

[plucky inlet]

Not sure chatgpt would write uppercase camelcase mixtures, but could be

[trim schooner]

you're confusing people here - cathei isn't involved with the asking of the question, just offered their interpretation as to why the { is randomly there

[leaden solstice]

It’s sometimes useful in C++, not really in C#

[fleet gorge]

ah

[azure frost]

Whoever made the Kinematic Character Controller is the one who wrote most of this code.

[plucky inlet]

whatever, so, you were talking about you might looking at the wrong method? Did you start debugging with logs already? Not sure what @arctic sparrow meant exactly about

[arctic sparrow]

Well, we had that code segment in the wrong function. Right now, we're getting somewhere, but something in the camera code is causing it to snap back to where it was before centering the camera.

[plucky inlet]

Did you start to comment out all camera affecting parts to see, when that behaviour starts to happen?

[arctic sparrow]

Still working on it

OK! I found a fix!

if(InputMod.GetButtonDown(CenterCamInput)) CharacterCamera.PlanarDirection = Character.transform.forward;

"CharacterCamera.PlanarDirection" was one of the values leading into determining the camera's rotation.

[quartz folio]

It's using the C# convention, so I would hope that it did

[plucky inlet]

So the code was uppercase for public and _lowercase for private, thats not the c# convention, right?

[quartz folio]

Pascal for public, underscore camel for private, yes

[plucky inlet]

Guess I have been too much into unity scripts then 😄

[plucky inlet]

you got anything to read up for me, just wondering why all scripts including unity built ins are not that way

[quartz folio]

Because Unity is very old

https://learn.microsoft.com/en-us/dotnet/csharp/fundamentals/coding-style/identifier-names#naming-conventions

[plucky inlet]

Haha, funny, I thought that _ thing was very old as it remembers me of very very old scripts I wrote back then

[quartz folio]

All new packages use the correct conventions, but the built-in code does not because it's a legacy codebase with that old style

[plucky inlet]

thanks for enlighting me here! 🙂 guess I will start the next project with those conventions correctly then

[thin aurora]

Isn't the convention in Unity like this because they used to support a javascript-like language that had these conventions?

Which is obviously long gone, but now this remains

[placid summit]

ugh _ is an odd convention when IDEs now color code for private public

[quartz folio]

They also use Hungarian notation

Which has nothing to do with JavaScript-likes. It's just a really old codebase with very old decisions

[thin aurora]

Never seen that happen, ever

[quartz folio]

You get used to it real quick, so imo as long as it's consistent it really doesn't matter

[placid summit]

yeah tbh if I work in a team I go with whatever but I do not feel the need for _ or s _ myself ... cant write underscor

[sleek bough]

There's another good reason to use underscore for private. When you work with a large managing class, and it's hard to track all private class fields, when you start typing with _ you get them to popup easily.

[quartz folio]

I used to use Unity's conventions, then I switched, and imo it's nicer to use C#'s because then you feel more comfortable when working externally where that convention is practically saturated

[sleek bough]

But personally I don't like using it for aesthetics and trying avoiding classes get so large that I need that :D

[thin aurora]

Is this some VS setting that's disabled by default?

[placid summit]

ah yes - because otherwise I probably would do it!!

[quartz folio]

Some people don't have class and struct type names highlighted differently and that freaks me out

[thin aurora]

I might turn that on then. I think the conventions are still very useful for the reasons listed, but it's definitely an added step in readability

[placid summit]

What I do not like is Unity's old camelCase properties which have long been wrong

Is there a convention of not always using properties in structs like with Vector3?

[thin aurora]

VS supports analyzers using .editorconfig files. I believe it also has conventions for that

Though it might suggest properties rather than fields, which is the opposite in your case

[quartz folio]

Structs don't necessarily need to protect input because usually they're not doing work, so a property would be unnecessary overhead

[placid summit]

Yes though not sure any overhead remains after unity optimises it, at least with il2cpp

[quartz folio]

In the past Unity does very little method inlining

The Vector structs don't care if you assign X as 1 or 29, it's just data, what good would a property do

Any float is valid, without context it's just raw data

[quartz folio]

https://docs.unity3d.com/6000.0/Documentation/Manual/UnderstandingPerformanceSpecialOptimizations.html

Reducing method call overhead

[placid summit]

There is no reason to use a Property in many class cases but it is standard practice afaik. Was not sure if differes for structs

[quartz folio]

It's only standard to avoid making breaking API changes

If you've got nobody consuming an API, then it's a practice ready to be dropped

[placid summit]

thats an interesting read. I thought list.Count caching was not required...

[quartz folio]

They haven't updated that doc in a long time so I would expect certain things to have changed

Also this is a micro optimisation in most cases

[placid summit]

indeed

[chilly surge]

A field's getter and setter accessibilities are tied together. If you want to have public getter and private setter for example, then property is the only way.

Interface can contain properties but not fields. This means that if you need to latter add an interface to an existing class, you don't need to refactor anything.

It's unfortunate that Unity's serialization doesn't play very well with properties, that property vs field is even a choice we have to consider.

[placid summit]

yeah this is a problem since Properties are more common that fields these days!

you can serialize the field but will have odd name if not full property

[cyan ivy]

What

[wheat spruce]

I've got this code that lets you swing the back object around the front.

private void OrbitWheels()
{
  _orbitSpeed += (_aimDirection != Vector2.zero ? 1 : -1) * _speedRate * Time.fixedDeltaTime;
  _orbitSpeed = Mathf.Clamp(_orbitSpeed, 0f, _maxOrbitSpeed);

  Vector2 sideForce = _aimDirection;
  Vector2 force = transform.TransformDirection(sideForce) * _orbitSpeed;
  _rb.AddForce(force, ForceMode2D.Force);
}```The issue is the orbit direction starts to become confusing when you face other directions instead of moving up. At the start Left/Right makes sense, but at the end when you move downwards, Left/Right become inverted

_aimDirection is set with the arrow keys, so it really should be able to use the correct key based on the direction you are heading

[maiden fractal]

Doesn't TransformDirection there treat sideForce as local space direction and make it world space though? What you seem to want it just use the input vector as a world space vector as is

[late lion]

Are you trying to mimic the physics of a trailer?

[wheat spruce]

oh yeah, that was it!

[wheat spruce]

Sort of. But its actually going to be a train carriage thats driving "offroad"

well, offrail

[copper rose]

what do you think guys for a first person game, its better to manage the camera as children of player or both separate gameobjects? i cannot get working smoothly the gamepad, it wors fine mouse only, and maybe its a better approach to add camera as children ?

[atomic kindle]

my gs how do I read a playerpref registry file

I tried google

let me try again

not as in find it as in read the value of it

[wintry crescent]

is there a difference between transform as RectTransform or (RectTransform)transform?

[quick token]

do you want the camera to move with the player at all times? if not, try a position constraint. if you do though, just parent it. kind of depends on what kind of game you're making

[wintry crescent]

some context:

[quick token]

they're the same thing

the second one is a typecast

oh wait, i misread

[wintry crescent]

for a first person game, in the past I often did player as child of camera, but keeping them separate is how I'd do it now

[maiden fractal]

Read from playerprefs? I highly doubt you found nothing

[atomic kindle]

i used hexadecimal

I just yoloed and it worked

[maiden fractal]

what?

[atomic kindle]

anyways I have a bigger problem

I set this to "hey:

"hey"

and this outputs null and saveslot 1

[maiden fractal]

plain text username and password

[atomic kindle]

sorry I dont understand?

[wintry crescent]

it's not a server xd

[quick token]

https://stackoverflow.com/questions/132445/direct-casting-vs-as-operator

string s = (string)o; // 1
Throws InvalidCastException if o is not a string. Otherwise, assigns o to s, even if o is null.

string s = o as string; // 2
Assigns null to s if o is not a string or if o is null. For this reason, you cannot use it with value types (the operator could never return null in that case). Otherwise, assigns o to s.

string s = o.ToString(); // 3
Causes a NullReferenceException if o is null. Assigns whatever o.ToString() returns to s, no matter what type o is.

Use 1 for most conversions - it's simple and straightforward. I tend to almost never use 2 since if something is not the right type, I usually expect an exception to occur. I have only seen a need for this return-null type of functionality with badly designed libraries which use error codes (e.g. return null = error, instead of using exceptions).

3 is not a cast and is just a method invocation. Use it for when you need the string representation of a non-string object.
(i think this explains it)

[maiden fractal]

I know, doesn't make huge deal anyways

[atomic kindle]

for proof that its hey in the registry

[wintry crescent]

hopefully at least

[wintry crescent]

show us your setstring code

[atomic kindle]

oh shooot I sent both getstrings

hold on

[wintry crescent]

and you're absolutely 100% sure that Save.selectedSaveSlot is the same

like dead sure

[atomic kindle]

its a static variable that is not edited anywhere else

public void Submit(string type)
{
user = username.text.ToString();
pass = password.text.ToString();

    if(type == "login")
    {
        Debug.Log(PlayerPrefs.GetString(Save.selectedSaveSlot + "_Username"));
        Debug.Log(PlayerPrefs.GetString(Save.selectedSaveSlot + "_Password"));
        if(PlayerPrefs.GetString(Save.selectedSaveSlot + "_Username") == name && PlayerPrefs.GetString(Save.selectedSaveSlot + "_Password") == pass)
            {
                SceneManager.LoadScene(4);
            } 
            else
            {
                Debug.Log("Wrong password or username");
            }
        
    }
    if(type == "signup")
    {
        PlayerPrefs.SetString(Save.selectedSaveSlot + "_Username", user);
        PlayerPrefs.SetString(Save.selectedSaveSlot + "_Password", pass);
        Instantiate(signUpConfirm, mainCanvas.transform);
        Invoke("ConfirmSignup", 2.0f);

    }
    
    
}

[quick token]

!code

[tawny elk]

Posting code

📃 Large Code Blocks

Use links to services like:
https://paste.mod.gg/, https://hastebin.skyra.pw/, https://paste.ofcode.org/, https://paste.myst.rs/, https://scriptbin.xyz/

📃 Inline Code

Surround code with three backquotes. Not quotation marks.
To format as C#, add cs to the first line:
```cs
// Your code here
```
Add a comment with a line number if there is an error message.

[plucky inlet]

!code

[tawny elk]

Posting code

📃 Large Code Blocks

Use links to services like:
https://paste.mod.gg/, https://hastebin.skyra.pw/, https://paste.ofcode.org/, https://paste.myst.rs/, https://scriptbin.xyz/

📃 Inline Code

Surround code with three backquotes. Not quotation marks.
To format as C#, add cs to the first line:
```cs
// Your code here
```
Add a comment with a line number if there is an error message.

[quick token]

hah, beat you to it

[plucky inlet]

woops 😄

[atomic kindle]

public void Submit(string type)
    {
        user = username.text.ToString();
        pass = password.text.ToString(); 

        if(type == "login")
        {
            Debug.Log(PlayerPrefs.GetString(Save.selectedSaveSlot + "_Username"));
            Debug.Log(PlayerPrefs.GetString(Save.selectedSaveSlot + "_Password"));
            if(PlayerPrefs.GetString(Save.selectedSaveSlot + "_Username") == name && PlayerPrefs.GetString(Save.selectedSaveSlot + "_Password") == pass)
                {
                    SceneManager.LoadScene(4);
                } 
                else
                {
                    Debug.Log("Wrong password or username");
                }
            
        }
        if(type == "signup")
        {
            PlayerPrefs.SetString(Save.selectedSaveSlot + "_Username", user);
            PlayerPrefs.SetString(Save.selectedSaveSlot + "_Password", pass);
            Instantiate(signUpConfirm, mainCanvas.transform);
            Invoke("ConfirmSignup", 2.0f);

        }
        
        
    }

[plucky inlet]

can you store your playerprefs get in a string before logging it?

[atomic kindle]

I signed up first

so its saved

to hey

[plucky inlet]

string yourName = PlayerPrefs.GetString(Save.selectedSaveSlot + "_Username");
Debug.Log(yourName);

[wintry crescent]

not gonna lie, this should work

like if you do signup and login now, this should work

[thin aurora]

That's something you can check by diving into the definition of RectTransform. I assume it has an implicit conversion from transform or this code would throw

Alternatively transform is already a RectTransform here. I believe this is in 2d?

So transform's actual type is a base type for both 2d and 3d

[wintry crescent]

transform does not have an implicit conversion to RectTransform, unfortunately

but I think you don't understand what I mean here, I wanted to know the differences between those two ways to cast transform to recttransform, my question was answered

[atomic kindle]

comes out as this

[wintry crescent]

they're effectively the same for my purpose

[atomic kindle]

thats why I'm so confused Im not the ebst at this but the logic should be there

[wintry crescent]

try giving it a default value perhaps? to check if it's saved as a space or sth, or if it's reading an unsaved value

[thin aurora]

Can I ask the reason why you are doing this casting? You probably want to pattern match

[atomic kindle]

give what a default value?

yourName?

[wintry crescent]

mate you have no clue what you're talking about, from the sounds of it

[thin aurora]

I would appreciate it if you stop being so passive aggressive constantly

Pattern matching is another way to match against a type, just like casting

[atomic kindle]

OK guys, I think I got it figured out

[wintry crescent]

yeah just for testing, do something like this
PlayerPrefs.SetString(Save.selectedSaveSlot + "_Username", user);
Debug.Log(PlayerPrefs.GetString(Save.selectedSaveSlot + "_Username", "nothing was saved"));

[wintry crescent]

lemmie guess user variable was empty xd

[atomic kindle]

turns out this was not the issue

its the if statement that doesnt show true

for some reason

[plucky inlet]

I think, playerprefs is not "done" when logging it, so you get null/nothing

[gilded monolith]

does anyone know how to lock cursor to a certain position?

[atomic kindle]

No I set the saveslot value when I press the save file I want to enter, lol I was an idiot and worked from the logit scene itself so octia was right its empty

[plucky inlet]

You can lock it, but have to calculate your offset yourtself, as you cant control the input from the user.

[wintry crescent]

certain position being center of the screen?

Cursor.lockState = CursorLockMode.Locked;

otherwise? don't.

[wintry crescent]

yay!

[atomic kindle]

 string yourName = PlayerPrefs.GetString(Save.selectedSaveSlot + "_Username");
            Debug.Log(user);
            Debug.Log("prefs: " + yourName);
            string yourPass = PlayerPrefs.GetString(Save.selectedSaveSlot + "_Username");
            Debug.Log(user);
            Debug.Log("prefs: " + yourPass);
            
            if(PlayerPrefs.GetString(Save.selectedSaveSlot + "_Username") == name && PlayerPrefs.GetString(Save.selectedSaveSlot + "_Password") == pass)

[atomic kindle]

I'd be happy too if I didnt see this:

WHY????

THEY'RE LITERALLY THE SAME

[plucky inlet]

Equals == true?

[wintry crescent]

that's why we hate strings lmao

[plucky inlet]

Hidden unicode? 😄

[atomic kindle]

how would I ever confirm this

[wintry crescent]

debug the lengths of each string

[atomic kindle]

    public void Submit(string type)
    {
        user = username.text.ToString();
        pass = password.text.ToString(); 

        if(type == "login")
        {
            string yourName = PlayerPrefs.GetString(Save.selectedSaveSlot + "_Username");
            Debug.Log(user);
            Debug.Log("prefs: " + yourName);
            string yourPass = PlayerPrefs.GetString(Save.selectedSaveSlot + "_Username");
            Debug.Log(user);
            Debug.Log("prefs: " + yourPass);
            
            if(yourName == name && yourPass == pass)
                {
                    SceneManager.LoadScene(4);
                } 
                else
                {
                    Debug.Log("Wrong password or username");
                }
            
        }
        if(type == "signup")
        {
            PlayerPrefs.SetString(Save.selectedSaveSlot + "_Username", user);
            PlayerPrefs.SetString(Save.selectedSaveSlot + "_Password", pass);
            Instantiate(signUpConfirm, mainCanvas.transform);
            Invoke("ConfirmSignup", 2.0f);

        }
        
        
    }

it literally both uses user

so it doesnt make sense

its like asking if user == user and getting false

[wintry crescent]

sanity check...
try
if((yourName == name) && (yourPass == pass))

[atomic kindle]

if this works istg

[wintry crescent]

yeah XD

[atomic kindle]

No sir

[thin aurora]

You should not be comparing strings like this. Strings in C# have build in systems that increase performance by reusing them. The reason yours are not used is because they are not reusing the same string instance most likely

So instead of if (a == b), use if (a.Equals(b))

[atomic kindle]

OK

brb

[wintry crescent]

good idea actually yeah

worth a try

[thin aurora]

Also pointing out the obvious, but if a person applies a single character with different casing, it will also fail. name != Name. I assume you understand this

[wintry crescent]

I'd consider that intentional if I was building this, tbh

for password obviously, but even for the name

[thick terrace]

those two are equivalent for String

[gilded monolith]

can you show example?

[thin aurora]

That was for Sehongbear. If their application doesn't case about casing in usernames (which is quite often), then they should make sure to compare a string whilst ignoring the casing

[atomic kindle]

Im restarted

[wintry crescent]

did it work?

[atomic kindle]

i did name instead of pass

[wintry crescent]

oh lmao

[atomic kindle]

through toLower() and all that?

[thin aurora]

No, if you provide the StringComparison enum through Equals() it just changes the way comparisons is handled in general

[atomic kindle]

I'll just to lower them both like this

or is that not good

[thin aurora]

This isn't just by defining how casing works. There are certain cultures where in one case a string might mean the came, but in others it does not

Like weird characters with symbols above them. I don't have a direct example

[knotty sun]

you should use ToLowerInvarient

[atomic kindle]

OK

[thin aurora]

You already use Equals here. Just add a second parameter StringComparison.OrdinalIgnoreCase

No need for ToLower or anything like that

if(yourName.Equals(user, StringComparison.OrdinalIgnoreCase) ...

[atomic kindle]

OK

[thin aurora]

This is generally how string comparison is done

[atomic kindle]

I'll always use .equals from now, thank you sir for teaching me a lifelong lesson

genuinely forgot about allat equals stuff

[thin aurora]

It's not applicable literally everywhere. You might want to use ToLowerInvariant or ToUpperInvariant in cases for normalization or something

Fun fact, you should actually use ToUpperInvariant because ToLowerInvariant and the other non invariant versions have unexpected quirks to them

[atomic kindle]

OK

wait but I did ignorecase not toupper

[thin aurora]

That's better suited for this use case, yes

[thick terrace]

ToUpper/ToLower involve copying the characters into a newly allocated string, passing in the the comparison type changes the comparison logic without needing to do that

[atomic kindle]

    private void CheckLetterInput()
    {
        foreach(Key key in keys)
        {

            if (Input.GetKeyDown(key.ID))
            {
                PressKey();
            }

            if (Input.GetKeyUp(key.ID))
            {
                ReleaseKey();
            }
        }  
    }
    private void PressKey()
    {
        foreach(Key key in keys)
        {
            if (key.CheckKey())

                key.IsPressed();
        }
    }

    private void ReleaseKey()
    {

        foreach (Key key in keys)
        {
            if (key.CheckKey())

                key.IsReleased();
        }
    }

    public void IsPressed()
    {
        gameObject.GetComponent<SpriteRenderer>().sprite = keyPressedAlt;
    }

    public void IsReleased()
    {
        Debug.Log("hi");
        gameObject.GetComponent<SpriteRenderer>().sprite = keyPressed;
    }

anyone know why this doesnt work?

"hi" doesnt show up in the console

meaning that key.CheckKey isn't true for release key while it is true for presskey??

hard to understand

[thin aurora]

Time for clarity then

Start logging the values, see if they return what you expect

https://unity.huh.how/debugging

If you think the issue is key.CheckKey, then debug what the value returns

First thing you always do is pinpoint the exact part of the code with the issue

[atomic kindle]

understood sir

the foreach part works

ok

using this

the first checkkey for presskey is true then 25 falses

but all checkkeys are falses for releasekey

[knotty sun]

would it not have been sensible to log which key you were checking?

[atomic kindle]

oh to print all keys?

ok

OK. I pressed the w key which was right and the presskey got it

however, release key always seems to start form Q

actually nvm

it just doesnt work

checkkey returns true for presskey

but doesnt for release key

even if the same P is entered

[thin aurora]

Keep on rubber ducking

[knotty sun]

and type in full sentences

[atomic kindle]

This is checkkey btw, for clarification

[thin aurora]

Looks good to me chief

[atomic kindle]

This is so weird. They both are the exact same methods, so it doesn't make sense that one is returning false while the other is returning true.

[thin aurora]

I don't get why you overcomplicate it so much

Is Key your own class?

[atomic kindle]

Yes

[thin aurora]

Why do you have that?

[atomic kindle]

I have 26 gameobjects to create a keyboard

[thin aurora]

Oh, this is a game keyboard?

[atomic kindle]

Yes

[thin aurora]

Understandable

[atomic kindle]

is pressed and is released are supposed to be to change the sprite of the key, pressed down and greyed out or not

Theres literally nothing more to uncover. Its clear that everything is working, but that CheckKey is not returning true for ReleaseKey for some reason, but the why is unknown

I personally think its a problem with getkeyupo

but I don't know if theres anything I can use other than GetKeyUp for this

private void CheckLetterInput()
    {
        foreach(Key key in keys)
        {

            if (Input.GetKeyDown(key.ID))
            {
                PressKey();
                Debug.Log("hi");
            }

            if (Input.GetKeyUp(key.ID))
            {
                ReleaseKey();
                Debug.Log("hi1");
            }
        }  
    }

private void PressKey()
    {
        foreach(Key key in keys)
        {
            Debug.Log(key);
            Debug.Log(key.CheckKey());
            if (key.CheckKey())

                key.IsPressed();
        }
    }

    private void ReleaseKey()
    {
        foreach (Key key in keys)
        {
            Debug.Log(key + "1");
            Debug.Log(key.CheckKey() + "1");
            if (key.CheckKey())

                key.IsReleased();
        }
    }

public bool CheckKey()
    {

        return Input.inputString.Equals(ID, System.StringComparison.OrdinalIgnoreCase);
    }

I GOT IT

The inputstring will be empty when I let the key go

duhhhhhh so then if checkkey is called during the time when the key is released, then ofcourse the inputstring will be nothing!

so I need an input for checkkey

Thanks for letting me rubberduck yall

also do I need reload domain opn it takes so long every time. When can I disable it?

Ok new problem

this works but removing the debugs makes it not work anymore, specifically the debuglog on the top

[knotty sun]

do you not know how to write an if statement?

[atomic kindle]

Holdon this doesnt work?

[knotty sun]

think about it, what is the difference in logic with and without the Debug statement