archived-code-general

[elfin tree]

had no idea

[crude elm]

Hi,

I'm trying to solve a context steering issue when the desired movement matches obstacle direction i.e. the target is behind the obstacle. Most steering behaviours I've seen doesn't account for that out of the box.

Should this be a combination of movement context and stronger avoidance context, or there is a trick to mapping obstacles direction?

[elfin tree]

What would be the best way to convert a TimeSpan to a delay in a coroutine?

Normally I use WaitForSeconds but TimeSpan.TotalSeconds returns a double, same for milliseconds

I get that TimeSpan by substracting two DateTime, but I could do something else with the two DateTime

[late lion]

You can easily cast a double to a float.

[elfin tree]

Yeah I guess I'm worried about loss of precision

Like how are miliseconds a double

[late lion]

If you intend to use WaitForSeconds, then you can only use a float. A double is more precise than a float. You're not going to get worse precision than float when going from double to float.

[elfin tree]

ohh, i thought it was the other way around, makes sense

ty

[rain crater]

Hey, does anyone have any idea why this function causes Unity to get stuck on reloading domain, when I start a game, call it, and then stop play mode and try to start again?

public async void NATPunch()
    {
        var discoverer = new NatDiscoverer();
        var cts = new CancellationTokenSource(2000);
        var device = await discoverer.DiscoverDeviceAsync(PortMapper.Upnp, cts);

        if (device != null)
        {
            await device.CreatePortMapAsync(new Mapping(Protocol.Tcp, port, port, "game"));
            await device.CreatePortMapAsync(new Mapping(Protocol.Udp, port, port, "game"));
        }
        else
        {
            cts = new CancellationTokenSource(2000);
            device = await discoverer.DiscoverDeviceAsync(PortMapper.Pmp, cts);
            if (device != null)
            {
                await device.CreatePortMapAsync(new Mapping(Protocol.Tcp, port, port, "game"));
                await device.CreatePortMapAsync(new Mapping(Protocol.Udp, port, port, "game"));
            }
        }

        Debug.Log("Function ended"); //This prints
        return;
    }```
I'm using Open.NAT to do port forwarding in my game, and I'm doing stuff according to how it says on their github. The function normally gets to the end (the debug log prints) but then happens what I said. I've heard some stuff about async function not being the best with Unity, so I guess it might be something with it. 
And it seems like this function causes problems, because making it return immediately after calling it makes everything work

(If this is not Unity-specific enough, which it might be, please let me know, i'll go ask somewhere else)

[knotty sun]

you are not Disposing your CancellationTokenSource instances which may be a problem. Also how are you invoking this method?

[rain crater]

just a normal call NATPunch(); from a normal method
I tried to make cts = null, unless you mean the .Dispose() thing (i'll try it now)

tried the Dispose thing, it doesn't help

technically I could just not call this function in the editor, but then it would be nice to fix it, so it doesn't do some weird things in build (just noticed - when I try to stop the game it freezes, like Unity on domain reload)

[glad sundial]

Hello, I have a simple question. How can i do to detect if an active GO is triggered on my disabled GO ?
I tried to create an empty GO and add trigger class but didn't work well ...

[serene stag]

the scripts on disable go are disabled too.

[glad sundial]

yes i know

that's why i create a empty GO

I copy every properties but it didn't work well

[leaden ice]

I have no idea what you mean by this question to be honest

[serene stag]

you are trying to do a collision between enabled go and disabled go and checking ontrigger.

disabled go never detects collisions.

[glad sundial]

I'll add some context. I have a GO (my agent) and I have an item that spawns on 2 platforms randomly. I don't want the item to spawn if the agent is already on the spawner. So the disabled GO is the item and the active GO is the Agent

[glad sundial]

that's why I use an empty GO. Empty GO is like a copy of the item but without mesh. So i can use it to ensure that there is no agent on my position 🐙

my English is very limited. That's why it's hard to understand me. I'm really sorry !!!

[serene stag]

then why make the item go disabled just make the script which is spawning disable when the agent is on the spawner platform simple.

[glad sundial]

No, I can't do that because as there are several spawners, I just have to make sure that they don't spawn on the platform where the item is supposed to spawn.

private void Update()
    {
        if (NbItemSpawned == 0) // No item on stage
        {
            GameObject randomItem = RandomizeSpawn(); // Get random item and random platform
            m_SpawnChecker.SetItem(randomItem); //Setup trigger checker for item

            if (!m_SpawnChecker.get_IsTriggered()) // GO not triggered
            {
                NbItemSpawned++; // Update nb items
                randomItem.SetActive(true); // Active item
            }
        }
    }
...
TriggerClass.cs:

public void SetItem(GameObject item)
    {
        m_IsTriggered = false; // Reset value
        gameObject.transform.SetParent(item.transform.parent); // Set Item parent
        
        // Copy Transform
        gameObject.transform.localPosition = item.transform.localPosition; // Copy localPosition
        gameObject.transform.localRotation = item.transform.localRotation; //Copy localRotation
        gameObject.transform.localScale = item.transform.localScale; //Copy localScale
        
        // Copy Box Collider
        BoxCollider itemCollider = item.GetComponent<BoxCollider>();
        m_TargetCollider.center = itemCollider.center; // Copy center
        m_TargetCollider.size = itemCollider.size; //Copy size
        m_TargetCollider.isTrigger = itemCollider.isTrigger; //Copy trigger state
        
        gameObject.SetActive(true);
    }

Mb what I'm trying to do will help

[vale bridge]

leftMouseClick_prev += Time.deltaTime;
rightMouseClick_prev += Time.deltaTime;
midMouseClick_prev += Time.deltaTime;
jumpPressed_prev += Time.deltaTime;

Is there a better way to do these crude deltatime counter increases

Also was wondering if I should check isGrounded in update or fixedUpdate, I imagine it would be the latter?

[heady iris]

If you have a bunch of variables and you need to increment them all, there's not much more you can do

[heady iris]

Check it at the same rate that your character moves. So, if you're moving with a rigidbody, check it in FixedUpdate (and do it before you try to jump, of course)

[vale bridge]

roger that

[swift falcon]

i have a quad object which has a videoplayer component inside it how do i access it in my code i have tried different approaches but doesnt work

its a simple play and pause function but doesnt work

Using a variable that is inside the class prevents the lambda to get pooled and it creates lambda every time. But why? Here's the compiled version to show you. Look at the M() in right side.

[heady iris]

I'm more surprised that the compiler managed to avoid creating garbage in the first case!

The first anonymous function captures no variables.

It only depends on its arguments

The second anonymous function captures a variable.

[swift falcon]

Oh

So it is because the object may be changed in runtime

[heady iris]

can you not just reference the VideoPlayer component?

[swift falcon]

i tried but i cant as it is not a part of the scene

the quad is not in the scene

[heady iris]

so it's on a prefab, then..?

[swift falcon]

yeah

[swift falcon]

Yes. I really did a lot research about this thing :d Good to share

[heady iris]

then store a reference to the VideoPlayer when you instantiate the prefab.

https://unity.huh.how/references/references-to-prefab-children

[swift falcon]

i will try that

[balmy raptor]

Hi, what would be the best channel to post a question I have about Unity As Library for Android Studio?

[swift falcon]

i am not able to wrap my head around this

the prefab is spawned by XR origin

i followed the instructions

but how do i get to the videoplayer

for instance the video player is a component of a child inside the parent prefab

[fresh hazel]

Hey, so I'm making a project with Unity but I have huge lags and my profiler looks like that
It looks like the major issues are the EditorLoop and the RenderLoop, but I'm not sure what it mean exactly and how I can know what the problem is exactly?

[heady iris]

EditorLoop can be a bit misleading (in both directions)

I've seen situations where EditorLoop is obviously counting time taken to render the scene

but it can also add some very nasty lag spikes that won't actually happen in the built game

Full-screening the game view can help -- this stops unity from having to draw the inspector

which can get really expensive

[old linden]

Hi. I'm wanting to rotate around the Player given a fixed radius above the Player, whilst the player is moving but without parenting. I have the following code which works fine, but when the player moves, the camera increases its radius, instead of maintaining the same distance. What's the best way to do this?

[umbral fox]

Unity2022 LTS ... exiting play mode triggers Awake + Start + OnEnable ... anyone know when/why/how this was changed?

Seems to happen whenever I subclass Graphic (nothing else needed to trigger it)

(or is this something that's always been true with Graphic and I somehow didn't notice before? Seems very unlikely, but I guess possible I didn't know!)

[heady iris]

Graphic has [ExecuteAlways] on it.

this allows for things to update as you work on your UI in edit mode

[umbral fox]

Has this always been the case? I am 99% sure it didn't used to do that (and specifically: I had to add it to subclasses when I wanted that behaviour)

But maybe I'm just losing my mind 🙂

[heady iris]

Unity UI has barely changed in a long time

It may have changed from [ExecuteInEditMode] to [ExecuteAlways], but that just indicates that it's safe to run in prefab isolation

[umbral fox]

That's not true. e.g. 2022 fundamentally broke UnityUI in multipe places :).

Also: UIToolkit team has repeatedly broken UnityUI in each major Unity release for years, to support/hide/workaround their own bugs.

[heady iris]

what?

[umbral fox]

Other than by installing old editors and manually reviewing the Unity source ... any suggestions on how I could verify it's always been like this / whether it changed recently? ... doesn't matter at this point, just want to check if I'm losing my memory 🙂

[umbral fox]

They broke all Custom Drawers (reported as bug, acknowledged - I have it from Unity officially). They broke all custom inspectors (reported as bug - they eventually fixed it). They broke all callbacks on RectTransform (reported as bug, QA refuses to allow it to be fixed, someone at Unity doesn't want to fix that one)

[heady iris]

https://github.com/needle-mirror/com.unity.ugui/blob/a601a2bf30161c47959231b627a8c40f64d69a68/Runtime/UI/Core/Graphic.cs#L78>

Graphic has been [ExecuteAlways] since at least unity 2020, just looking at a random example

[heady iris]

They broke everything and you were the only one who noticed it?

link me to the bug reports on the issue tracker.

I should point out that custom property drawers and custom editors have nothing to do with UGUI

they use either UIToolkit or immediate-mode graphics

[umbral fox]

Sorry, my friend, that's simply not true. UI Toolkit made some changes to how events are delivered that specifically broke all custom property drawers that were written WITHOUT UI Toolkit if you had both present in the same project.

I wasn't impressed, and the final response I got from them was essentially: Yes but we want everyone to use UIToolkit, so you need to replace all your existing property drawers with UIT property dtawers.

(Fine, I was able to do that with my own code - but it meant we had to drop some 3rd party code/assets until they were all using uit too)

[heady iris]

I cannot find anything on the issue tracker.

[umbral fox]

Tried to fish out bug numbers for you, but I didn't keep copies of old emails when projects ended (I asumed I could always go back into fogbugz if I needed to re-read them).

[heady iris]

I also cannot find anyone complaining about this kind of thing, which I'm sure would have been a massive uproar

[umbral fox]

I asked around other studios at the time - most weren't using custom property drawers, so it didn't affect them - or they were heavily customising their editors and had already upgraded everything to UIT.

[winter bramble]

Bike shedding question for anyone who has tried. I want to create a room editor game in which users could load and execute assets during runtime. i.e. import a .zip that includes game "furniture" (really .lua scripts or other interpreted langs I build the interpreter for like js, and then other binary assets .fbx files, audio files, etc.) which can be placed in the game and persisted.

Wonder if Unity would be a good choice for this, or if im better off solving the whole runtime system in my own engine... (obv realize the my own part is reinventing / building a massive amount of what unity provides, I just want to emphasize the runtime part.)

[umbral fox]

(I moved all my editor-drawers code to UIT, and other than being careful waiting for 3rd party code to update, stopped caring)

[umbral fox]

Which is hte part that you think might not be a good fit for Unity? There are plenty of games and assets that broadly speaking do that.

[heady iris]

There are runtime importers for different kinds of assets.

I know that there are assets for runtime model import

Unity can decode audio at runtime...but only when loaded from the web, bafflingly

https://docs.unity3d.com/ScriptReference/Networking.DownloadHandlerAudioClip.html

I don't think it can decode an mp3/ogg that you've stored locally.

[winter bramble]

Good question, I am bit shakey of the common architectures that represent this "realtime 3d runtime asset plugin system" but I suppose the fact that if I want to say execute player added scripts at runtime to interact with "Unity's runtime" there could be some integration issues / limitations / extensions that would make this harder because the whole point of this game is this part, the rendering isnt going to be fancy. Id also like to be able to accomodate shaders at runtime and more fancy player added things... feels like the integration here is the thing I sense might be very hard to do it the Unity way.

[winter bramble]

Is this security related? Kind of like CSP? Like if its just a stream of bytes why do I have to serve it over HTTP?

[heady iris]

No; it's just a weird design, as far as I can tell

There are, of course, libraries you can use to decode MP3/OGG

Once you get an array of samples, you can create an AudioClip https://docs.unity3d.com/ScriptReference/AudioClip.Create.html

and then use SetData

Custom user scripting is another beast. I haven't touched that at all yet.

Something like Lua would be reasonable.

[umbral fox]

There's many examples of doing it, from embedding an interpreter (fine on most platforms, but e.g. when doing mobile you need to first check that your preferred interpreter actually compiles into IL2CPP correctly, and that Apple etc will allow it), to implementing your own cut-down language by hand, and many assets that will do either/both/etc for you (can't recommend any off the top of my head, but I've seen and used several, and implemented our own interpreterst from scratch for simple languages, and implemented our own custom langauges for simple in-game scripting).

[winter bramble]

I see. Good to hear it’s not ringing any don’t do this alarms. I suppose what I really want to build is a game that supports mods. I suppose they don’t need to be “injected” at runtime. Is there a cleaner more trivial solution if they don’t need to be added at runtime but instead pulled in from some menu and then restart? Ideally creating and publishing mods could be done in Unity separately and then just exported as a module. I guess I’d love any resources on mod architectures in Unity 🙂

[leaden ice]

You can use UnityWebRequest with local files

[heady iris]

ooh, you're right.

mildly cursed

[unique delta]

is it posible to remove an element from a list by name?

[leaden ice]

Only by iterating over the list until you find it

If you need to do things by name you should look into a Dictionary instead of a List

[heady iris]

well, there is RemoveAll

nums.RemoveAll(item => item % 2 == 0);

this deletes all even numbers from the list

I want to be able to memoize a bunch of classes. That looks like this:

Dictionary<ValueTuple<Foo, Bar>, MyClass> memoized = new();

protected MyClass(Foo foo, Bar bar) { /* constructor goes here */ }

public MyClass Memorize(Foo foo, Bar bar) {
  var tuple = ValueTuple.Create(foo, bar);

  if (memoized.TryGetValue(tuple, out var result)) return result;

  result = new(foo, bar);
  memoized[tuple] = result;

  return result;
}

So, if an instance has already been created with those exact arguments, just return the instance. If not, create a new instance and put it in the dictionary.

This works fine, but it's a decent bit of manual fiddling for each class.

I'm wonder if there's any way I could make this easier.

maybe i can create a separate "Memoizer" class (that would require a public constructor, but oh well)?

[unique delta]

can t i use this : list.RemoveAt(Convert.ToInt32(GameObject.Find("...."); ?

[heady iris]

i mean, sure, if the name of the game object is the index you want to remove from the list..

but that seems very weird

[unique delta]

it s perfect for me

[swift delta]

not to bump off fen's question but what's a good method to ensure a bullet going somewhat fast doesn't phase through colliders? some googling suggested attaching a raycast to the bullet and use its rigidbody as the ray's direction and distance (via magnitude) but I wanted to make sure that'd be a good idea

[lean sail]

does that even work or compile? finding a gameobject, then converting to int, and that magically being the index of in a list?
looks like AI code

[swift delta]

it's worth noting the issue lies in thin colliders, such as floating targets

[heady iris]

If it's a non-kinematic rigidbody, try changing its collision detection mode

[heady iris]

notably, this needs a .name in there somewhere to have any hope of working...

[heady iris]

https://docs.unity3d.com/ScriptReference/Rigidbody-collisionDetectionMode.html

[swift delta]

I remember switching it to other collision detection methods but didn't have much luck

let me check something

[heady iris]

Note that both the bullet and the target need non-default modes

[unique delta]

i just thinked it is working

[heady iris]

i don't know what this is supposed to mean

[lean sail]

then no it definitely does not work

[swift delta]

the target in question has no rigidbody, so what exactly would I need to change on them?

[heady iris]

For best results, set this value to CollisionDetectionMode.ContinuousDynamic for fast moving objects, and for other objects which these need to collide with, set it to CollisionDetectionMode.Continuous

so you'll want to get a rigibody on the target

I believe it can still be kinematic.

[swift delta]

wack. how about interpolation on the bullet?

[heady iris]

Interpolation is purely visual

It has nothing to do with collision detection

[swift delta]

noted

continuous dynamic mode on the bullets themselves changed nothing, so I'll go ahead and plug the kinematic rigidbodies on the targets

no dice

[heady iris]

These two options have a big impact on physics performance. Alternatively, you can use CollisionDetectionMode.ContinuousSpeculative, which is generally cheaper and can also be used on kinematic objects.

ah, so Continuous on the target with a kinematic rigidbody may not work properly

you should check if collisions are working properly at all

slow the bullet down or give it a very large collider

[swift delta]

increasing the hitbox's size helps, so yeah

[heady iris]

ah, ok

[swift delta]

specifically behind it, making it wider

[heady iris]

If the targets don't move, then just making the bullets use the Continuous collision detection mode should be sufficient

as long as the targets don't have a rigidbody on them!

https://docs.unity3d.com/ScriptReference/CollisionDetectionMode.Continuous.html
https://docs.unity3d.com/ScriptReference/CollisionDetectionMode.ContinuousDynamic.html

have a look at the descriptions

[swift delta]

I tried just that, but it didn't yield any results with or without the rigidbodies

on the targets, that is

the bullets' prefab already had its rb settings changes

[heady iris]

show me the bullet inspector

also, how are you moving the bullets? are they non-kinematic rigidbodies that you just set a velocity on?

[swift delta]

yeah, they are instantiated and accelerated on creation

things are being handled on an onTriggerEnter basis and I'm wondering if that's related

the bullet is the one handling said interactions

    void OnTriggerEnter(Collider other)
    {
        if (other.gameObject == owner) return;

        if (other.CompareTag("Player"))
        {
            PlayerOnline playerHit = other.GetComponentInParent<PlayerOnline>();
            Debug.Log("Teste");
            if (IsOwner)
            {
                PlayerOnline ownerPlayer = owner.GetComponentInParent<PlayerOnline>();
                Debug.Log(playerHit.GetTeam().teamTag);
                if (playerHit.GetTeam().teamTag != teamTag.Value)
                {
                    playerHit.Damage(damage);
                }
            }

            if (IsServer)
            {
                Debug.Log("Teste");
                Destroy(gameObject, 1);
            }
        }
        if (((1 << other.gameObject.layer) & groundLayer) != 0)
        {
            col.enabled = false;
            rb.velocity = rb.angularVelocity = Vector3.zero;
            rb.isKinematic = true;
            StartCoroutine("TrailGone");
            if (IsServer)
            {
                Destroy(gameObject, 1);
            }
        }
    }

bit of spaghetti here, sorry. we're still feeling our way around online interactions

[heady iris]

oh, this is using triggers, not collision messages

I am unsure if continuous collision detection works for those.

[swift delta]

let me swap this out for regular non-trigger collisions

[heady iris]

    public static class Memoizer<T, V> where T : struct
    {
        private static readonly Dictionary<T, V> Memoized = new();

        public static V Memoize(T tuple, System.Func<T, V> ctor)
        {
            if (Memoized.TryGetValue(tuple, out var result)) return result;

            result = ctor(tuple);
            Memoized[tuple] = result;

            return result;
        }
    }

        public StateActivity Memoize(ActivityData data, Entity entity, EntityState state)
        {
            return Memoizer<(ActivityData data, Entity entity, EntityState state), StateActivity>.Memoize(
                (data, entity, state),
                static tuple => new StateActivity(tuple.data, tuple.entity, tuple.state));
        }

I think it works! But it's also mildly gross.

i wish it could at least infer the types of T and V

If I didn't need to make the class itself generic (so that I can define a Dictionary<T, V>), I could put the generics on the method itself, and then that could be inferred

there is the evil option of storing a Dictionary<Type, Dictionary<object, object>> and keying it on typeof(T), but that results in all of these value types getting boxed

and the main objective here is to reduce allocations

[swift delta]

so hey things are working now

only problem:

guns aren't supposed to shove enemies into the Nth dimension lmao

should be an easy fix at least. thanks for the help

[heady iris]

Giving bullets a super low mass would be the physically-correct way to solve that

[swift delta]

that does make sense mathematically

that somehow made it worse at 0.0001 mass

[heady iris]

are you constantly setting the velocity of the bullet?

[swift delta]

nope, only when fired

the bullet has no drag either so it should just keep flying at the same speed it was fired at

bullet.rb.AddForce(
    (bulletPoint.transform.forward +
     bulletPoint.transform.right * deviation.x + 
     bulletPoint.transform.up * deviation.y)
     * guns[gunID].bulletSpeed,
     ForceMode.VelocityChange //edited
);
bullet.owner = playerObject;
Destroy(bullet.gameObject, 2);

ah I didn't specify its force mode

let's try that

mhm, that did it

[heady iris]

Ah, right -- you wound up with a bullet that's going really really fast

good to see the collision detection still worked lol

The default mode is ForceMode.Force, which doesn't really make sense for shooting a bullet

you probably had to use a ridiculously high bulletSpeed

(50 times what you expected, for a bullet with a mass of 1)

using ForceMode.Force means that you're applying a constant force

if you apply 1 newton of force in FixedUpdate for 1 second, a 1 kilogram object will accelerate up to 1 meter per seond

so doing it for one frame gave you 1/50th of a second of acceleration

[swift delta]

yeah. velocitychange is ignoring its mass, so that should mitigate the issue

thanks again for the help

[heady iris]

    public static class Memoizer
    {
        private static readonly Dictionary<Type, object> maps = new();

        public static TMemoized Memoize<TArgs, TMemoized>(TArgs tuple, System.Func<TArgs, TMemoized> ctor)
        {
            if (!maps.TryGetValue(typeof(TArgs), out var mapObject))
            {
                mapObject = new Dictionary<TArgs, TMemoized>();
                maps[typeof(TArgs)] = mapObject;
            }

            var map = (Dictionary<TArgs, TMemoized>)mapObject;

            if (map.TryGetValue(tuple, out var result)) return result;

            result = ctor(tuple);
            map[tuple] = result;

            return result;
        }
    }

Now I'm not the biggest of this this, but it does move the type arguments to the method..

it turns out you can do Dictionary<,> to avoid actually specifying a specific type

truly awful

oh, no, you can't: i just couldn't see the error because i'm not used to Rider's look

[simple egret]

Almost compiles, you need to sync the generic type constraints of TArgs to what the Dictionary requires

where TArgs : notnull

[heady iris]

that's not part of the type constraint, is it?

because the code is working :p

you just aren't allowed to pass null at runtime

[simple egret]

You mentioned an "error", I copy-pasted it in an online compiler and it yelled about TArgs not matching the constraints

[heady iris]

ah, no, this was an error with trying to do Dictionary<,>

[simple egret]

Ah, you can only use these in typeof normally

[heady iris]

yeah

[simple egret]

Unbound generics

[heady iris]

Maybe that comes into play if you're using nullable reference types

[simple egret]

Nah it was a hard compiler error, though the compiler is on .NET 8 so it might not exist on Framework

[heady iris]

i barely understand which order these things go in

.NET standard, .NET framework, .NET 2: .NET Harder

anyway, this does appear to be working

gotta do a bit of performance testing

woooooooo! big performance gains!

3x faster, way way way less GC

and I'm at least 70% sure that I can safely memoize these objects

[swift falcon]

Anyone got any idea why the IL2CPP build fails but mono passes? I get this error

'hash_compare': is not a member of 'stdext'

😕

[cosmic rain]

I can only assume that you don't have the correct dependencies installed

[exotic quiver]

I'm facing an issue that the unity editor gets unresponsive when using an UnityWebRequest.Get Request. As soon as i use

Debug.Log("right before sending");
yield return pointsRequest.SendWebRequest();
Debug.Log("right after sending");

the editor gets unresponsive and the last thing i can see in the console is "right before sending". What can cause this issue? The same code was running flawlessly a couple of weeks ago

[lunar lynx]

anyone know why Texture2D.GetPixelData() doesn't return the correct pixel count?

Here's my code:

Graphics.CopyTexture(brightnessRenderTexture,0,0,brightnessTexture2D,0,0);
Debug.Log(brightnessTexture2D.GetPixelData<Color>(0).ToList().Count);

brightnessRenderTexture is my RenderTexture at 64px by 64px. same dimensions for brightnessTexture2D.
64 x 64 = 4096
Debug.Logging both RenderTexture and Texture2D widths and heights confirm this.
I copy the RenderTexture Mipmap 0 to Texture2D Mipmap 0.
I count the number of pixels in brightnessTexture2D via GetPixelData().ToList().Count, but it only returns 1024, which is 32 x 32. Looks like it's reading not from mimap 0, but mipmap 1?

[mossy snow]

this isn't doing what you think it's doing. You're copying data on the GPU, but reading system memory. You probably want to be using https://docs.unity3d.com/ScriptReference/Rendering.AsyncGPUReadback.html or https://docs.unity3d.com/ScriptReference/Texture2D.ReadPixels.html

[cosmic rain]

Share the relevant code. Are you subscribing to a response callback or something?

[exotic quiver]

public IEnumerator MyFancyMethod()
{
    using (UnityWebRequest pointsRequest = UnityWebRequest.Get($"https://somefancyurl"))
    {
        pointsRequest.SetRequestHeader("Authorization", "bearer SomeBearerToken");
        pointsRequest.SetRequestHeader("Accept", "application/json");

        Debug.Log("right before sending");
        yield return pointsRequest.SendWebRequest();
        Debug.Log("right after sending");
        
        
        //... some other code
    }
}

no, just yielding for the response.

[cosmic rain]

Hmmm... Check the log files to see if there are any other messages.
Also, does it crash or freeze?

[exotic quiver]

Unfortunatelly, did already check the logs. no further messages. it just freezes, no crash.

[cosmic rain]

If so, you could try breaking with a debugger. Might need to attach it to a process, instead of using the default attach to unity mechanism.

[exotic quiver]

i'll give it a shot. thanks

[exotic quiver]

now i was able to get unity to crash when attaching a debugger:

    Native Crash Reporting
=================================================================
Got a UNKNOWN while executing native code. This usually indicates
a fatal error in the mono runtime or one of the native libraries 
used by your application.
=================================================================```

not quite helpful.

[cosmic rain]

There should be some info in the log files.

[exotic quiver]

well this is the info from the editor.log from the crash. everything else is giving no further clues, just the stacktrace n loaded modules.

[cosmic rain]

Well, the stack trace is important. What does it say?

[exotic quiver]

it pinpoints exactly to the SendWebRequest

[cosmic rain]

I see. What unity version are you using? Did you search online for similar issues?

[exotic quiver]

2022.3.17, upgraded from 2022.3.8 as there were similar ones in the unity bug tracker, but apparently all are marked as fixed within my version(s)

[cosmic rain]

Maybe it's some old package cache? Did you try deleting the library folder?

[exotic quiver]

oh great idea! gimme a sec

[crystal holly]

anyone got advice for getting vscode working with editorconfig? mine appears to ignore it :\ (couldn't see a channel for questions about IDEs, apologies if this is in the wrong place)

[mighty void]

Hi someone could help me? I have a problem, i'm making a 2d procedural terrain with perlin noise

But i dont know how i can set the block levels, like how put the grass the dirt and that

Cause i dont wanna program It i wanna make It in a for statment

[crystal holly]

choose a block based on height?

[mighty void]

Yes

[crystal holly]

It was a suggestion, not a question 🙂

[mighty void]

I dont know how i can do that because is a procedural world, the height always is changing by perlin noise

[crystal holly]

Presumably you're looping through the values of the perlin noise, so value = max height for that column. So top block gets grass, blocks below get dirt?

[mighty void]

I'm thinking to divide the height of any colum and position objects using that

What I want to do is automate the process to be able to add elements to a list of blocks and be able to position them

[empty elm]

I have a class CustomButton and interface ButtonObserver { onClick() }
The button component needs a reference to the observer but I can't serialize it since its an interface. I could make ButtonObserver a serializable abstract class but that comes with other disadvantages. What are some work arounds?

[rigid island]

if its an observer why does custombutton need to know about observer ?

wouldn't custom button just fire event for button observer to listen to ?

[empty elm]

Yeah you’re right that would work here. Thanks!

Are there any limitations to the event approach? Because the observer pattern demands the ability to add and remove observers so wouldn’t it require that the observed object knew its observers? I suppose that can be done through subscribing and unsubscribing to the event

[rigid island]

pretty much , only the observer needs to reference the button

something like this ?

public class CustomButton : MonoBehaviour
{
    public event Action OnClick;
    public void Clicked() => OnClick?.Invoke();
}
public class Listener : MonoBehaviour
{
    [SerializeField] CustomButton customButton;
    private void OnEnable() => customButton.OnClick += CustomButton_OnClick;
    private void OnDisable() => customButton.OnClick -= CustomButton_OnClick;
    private void CustomButton_OnClick() => Debug.Log("buttonClicked!");
}```

[lunar lynx]

is there a guarantee that materials using RenderTextures wait for RenderTextures to be ready before rendering the material?
for example: I have a Camera that renders to a RenderTexture. That RenderTexture is used in a material, and that material is applied to Model.

What is the order of the render in Unity? Does it go

1. Camera renders to RenderTarget.
2. RenderTarget applied as material input.
3. Render Model with the material.

Or is it possible for the sequence to go:

1. Render Model with the material (but RenderTarget input has not been updated yet)
2. Camera renders to RenderTarget.

[empty elm]

Exactly, thank you

[mighty void]

I'm thinking that i dont know xd

Someone could guide me?

[lean sail]

you may wanna clarify further, because your question is seemingly just "i dont wanna use an if statement" which alone is a silly idea. what specifically is the issue with doing so?

[quasi jewel]

Have any of you guys tried or have an opinion on the free code camp C# course?

I'm taking it right now even though I've been using unity for a couple months and feel like it's great

[mighty void]

I want to create a code to generate terrain with an unestablished list of blocks, my idea is that the blocks can be generated in an order no matter how many blocks you add to the list

The terrain is not generated with a fixed height, but it is generated with a maximum height that cannot exceed

[lean sail]

i meant what specifically is the issue? like you say you dont wanna use an if statement which implies this entire thing could be solved if you use one.

[mighty void]

I don't find a way in which I can place the blocks in an order according to their properties of each one, for example, if the height of the column is 17 then place grass on that block and the following ones (randomly denyro of itself) I generated 4 blocks of earth and the rest are rock, I can't find the logic, it's hard for me to think logically sometimes, it would help if you could guide me, and I would like to do a foreach on the list I have of blocks because I feel like I can access all of their blocks. values, in the case of using if no, how could I adapt it to its infinite or changing values ​​(from the list I mean)

[lean sail]

im confused really on what the issue is really. saying you would like to do a foreach doesnt clear things up because i dont know what blocks you even have, if you have a list of blocks with properties (like conditions of where they should be) then whats stopping you from actually placing these?

[mighty void]

I seed you a dm

[lean sail]

dont, read the server rules. keep it in here

[mighty void]

Ok

I have a list of blocks, it is not really defined, I have plans to create more blocks and add them to the list, and with the same script generate other worlds, distinguish

I'm not really sure how to replace the blocks using a for breaking down the properties, for example the value of the grass generation, its minimum value is 1, but if a mountain has a peak of height 17, how do I change that block 17 and I generate 3 blocks further down the ground or different levels of land depending on the column

[mighty void]

The coordinates always make me nauseous because I have little imagination.

[lean sail]

you'll need to show the code and what you're specifically stuck on, because if you have all these conditions already coded then i dont see what the issue is

or really just look into what minecraft does since this seems to be a clone

[mighty void]

I can't generate the blocks in the order I require. If I want to generate random blocks depending on the height, how do I do it? I mean, how do I do something like this?

It's something similar but in 2D

Like this

But when i try with a for always the The result is always that blocks like grass are placed at coordinate 0

and the land is generated in the same way following that pattern, how do I use the height of each column to generate the blocks in the correct way?

[rigid island]

https://screenshot.help

[mighty void]

:T i take a photo beacuse i send in WhatsApp :T

[spring creek]

Then send it a different way 🤷‍♂️

[vale bridge]

Accidentally stumbled upon the Enum type when I wanted to declare an enum, apparently there's two different "enums" depending on if you capitalize the e or not?

Enum vs enum what's the difference?

[rigid island]

Enum is original class

just like string is String

[vale bridge]

I actually have no idea what the difference is

Ohhh

[rigid island]

enum is for example when you want to declare new enum type,
Enum contains static methods you can use

[rigid island]

https://learn.microsoft.com/en-us/dotnet/fundamentals/runtime-libraries/system-enum

[vale bridge]

I'm reading this post about it as well https://stackoverflow.com/questions/7074/what-is-the-difference-between-string-and-string-in-c

so similar to the string thing enum is the C# alias for the original Enum class? Something like that?

[rigid island]

keyword for creation vs the actual implementation basically

[vale bridge]

Can I say that string and enum are both created for the purpose of easier "creation" of their original types

[rigid island]

in a way kinda

but they kinda serve bit different purposes

[vale bridge]

I see

[lean sail]

its just an alias

[chilly surge]

There is also a semantic difference, when you use string it explicitly means System.String, it can't be anything else; whereas if you use String, it could be System.String, it could be MyAwesomeProject.String, or any other class named String.

In practice this rarely ever matters, but it's just good to use string over String, it makes your intent very clear.

[rigid island]

also you can directly access string methods like with string.Parse
With enum the enum keyword is for creation , you can acces static methods with Enum class

var list = (Somethin[])Enum.GetValues(typeof(Somethin));//valid
var list = (Somethin[])enum.GetValues(typeof(Somethin));//nope```

[edgy ether]

so, is it possible to transfer a component and it's values from one gameobject to another via script?

ex: a polygon collider's values

[lean sail]

have you tried AddComponent with a reference to the component u want to copy?

[edgy ether]

no? but i will now

thx

[heady canopy]

Hi, I downloaded for the first time and I can't open a new project

[grave bloom]

how do i set notification icon with unity mobile nofitication's unified API ? it is really driving me crazy

[mossy snow]

you might want to spend a little more time clicking around the documentation

[spring basin]

Hello had a doubt about an email I got.

Does this mean, I have to add levelplay to apps which serve Unity Ads exclusively as well?

[sleek bough]

@spring basin This is a code channel

[spring basin]

could you please redirect me to the channel where I can ask this question? Thank you

[sleek bough]

This is a support issue anyway. Create a ticket on website. Did you even try to look this up? People probably discussing it on a forum somewhere.

[spring basin]

didnt really find a thread on the forum for this specific one ._.

ill look again

[sleek bough]

Closest thing for long form questions/conversations here is #1157336089242112090 , create a thread there.

[spring basin]

thank you

[lean seal]

hi, I wanted to make a thing so that when you click on the object it goes down some distance and I did it, but there is a problem
if the parent object is rotating and the child object goes down with update: transform.position = transform.up * -20 * Time.deltaTime; then everything works correctly
but if you try to do the same with a coroutine. then the object goes to the place where it would be if the parent object were not rotating

public float fallDistance = 100f; // The distance the image will fall
public float fallDuration = 0.5f; // The duration of the fall
private IEnumerator CustomAnimation()
{
    Vector3 originalPosiition = transform.position;
    Vector3 fallPosition = originalPosiition + (transform.up * (-fallDistance));

    float elapsedTime = 0;
    while (elapsedTime < fallDuration)
    {
        transform.position = Vector3.Lerp(originalPosiition, fallPosition, elapsedTime / fallDuration);
        elapsedTime += Time.deltaTime;
        yield return null;
    }
}

Maybe i do something wrong?

[knotty sun]

move your fallposition calculation inside the while, presumably the transform.up is changing due to the rotation

[fervent furnace]

transform.up will be changed if parent is rotating
and you use this in your "original" calculation

[muted patrol]

@vale wharf check your inbox!

[vale bridge]

Is this an accepetable/good way of setting up controls? I'm curious how people have set up controls before the new unity input system:
https://pastebin.com/4bmvHt0Q
https://pastebin.com/4zDeUjgE

[slate imp]

Hi, i've got a strange behaviour from unity :
in the following, the remove method is called by an event, and in the if(gameObject != null) Destroy(gameObject) i have an error of the object is already destroyed :

    /// <summary>
    /// Removes the word interactable
    /// </summary>
    private void Remove() {
        
        if (pairedWord != null) pairedWord.OnWordTouched -= PairedWordHandler;
        gameState.OnGameOver -= OnGameOver;
        if (gameObject != null) 
            Destroy(gameObject);
    }

the error :

MissingReferenceException: The object of type 'WordInteractable' has been destroyed but you are still trying to access it.
Your script should either check if it is null or you should not destroy the object.
WordInteractable.Remove () (at Assets/_Summer2024/WordSaber/Scripts/SaberInteractions/WordInteractable.cs:46)
WordInteractable.OnGameOver (System.Object sender, System.EventArgs e) (at Assets/_Summer2024/WordSaber/Scripts/SaberInteractions/WordInteractable.cs:51)
GameStateSO`1[T].GameOver () (at Assets/_Summer2024/Shared/Scripts/So/GameStateSO.cs:27)

[knotty sun]

if it was the gameObject it was complaining about why does the error message say object of type 'WordInteractable' ?

[maiden heath]

is there a way to make unity search for a specific file name and use it to set a value?

i mean in my assets

[slate imp]

I don't know, the line 46 of WordInteractable.cs is the if (gameObject != null)

[vale bridge]

I know if you put your desired file in the Resources folder you can search it from there

[maiden heath]

Whats a resources folder

[slate imp]

https://docs.unity3d.com/Manual/UnderstandingPerformanceResourcesFolder.html

[maiden heath]

thanks

[slate imp]

I recommend you to use addressables, its a better approach : https://docs.unity3d.com/Packages/com.unity.addressables@1.19/manual/index.html

[maiden heath]

interesting

thank you

[slate imp]

Maybe it could be due to having the gameobject already destroyed but i dont know why the if would make an error

[knotty sun]

show code for WordInteractable.OnGameOver

[cosmic rain]

Is that remove method called from an event somewhere?

[slate imp]

its litterally calling Remove :

    private void OnGameOver(object sender, EventArgs e) {
        Remove();
    }

[slate imp]

here is the entire class https://pastecord.com/ekydyxytuf.cs

the method is private

[cosmic rain]

Make sure you unsubscribe properly. And don't subscribe more than once.

[knotty sun]

OK, I think you are trying to Destroy the holder of the event which is being processed

[slate imp]

i would agree, but it would have an error when subscribe or unsub more than once, i've already had errors linked to more than one sub and unsub, and it makes an error right away

[slate imp]

i don't quite understand

[cosmic rain]

What error does it make?🤨

[slate imp]

it says that you can't subscribe twice to an event and can't unsubscribe to an even that you are not sub

[cosmic rain]

Hmmm... Is that a unity event or something?

I don't think regular C# delegate events throw an error like that.

[slate imp]

Csharp events : public void event EventHandler

[knotty sun]

if the event which calls OnGameOver belongs to WordInteractable then when you Destroy the gameobject you are also destroying the source of the event call

[cosmic rain]

Ah, it's an EventHandler

[slate imp]

ohhhh maybe yes, i'll look that

[knotty sun]

It's like trying to remove the chair you are standing on if you get my meaning

[slate imp]

yep got the meaning, i think i've got the source of the issue :

In the full log, at the begining we have line 34 WordInteractable that send an event to say he's touched, then destroy himself, but when he's destroyed, the player change and the health of the player goes to 0, which makes the gameOver and re-tries to Remove

MissingReferenceException: The object of type 'WordInteractable' has been destroyed but you are still trying to access it.
Your script should either check if it is null or you should not destroy the object.
WordInteractable.Remove () (at Assets/_Summer2024/WordSaber/Scripts/SaberInteractions/WordInteractable.cs:46)
WordInteractable.OnGameOver (System.Object sender, System.EventArgs e) (at Assets/_Summer2024/WordSaber/Scripts/SaberInteractions/WordInteractable.cs:53)
GameStateSO`1[T].GameOver () (at Assets/_Summer2024/Shared/Scripts/So/GameStateSO.cs:27)
WordSaber.GameModeManager.EndGame () (at Assets/_Summer2024/WordSaber/Scripts/GameModeManager/GameModeManager.cs:47)
WordSaber.SurvivalManager.EndGame () (at Assets/_Summer2024/WordSaber/Scripts/GameModeManager/SurvivalManager.cs:28)
WordSaber.SurvivalManager.WordHit (System.Object sender, System.Boolean isCorrectWord) (at Assets/_Summer2024/WordSaber/Scripts/GameModeManager/SurvivalManager.cs:37)
WordSaber.WordsManager.<Spawn>b__18_1 (System.Object sender, System.Boolean _) (at Assets/_Summer2024/WordSaber/Scripts/WordsManager.cs:102)
WordInteractable.WasHit (System.Boolean isRightHand) (at Assets/_Summer2024/WordSaber/Scripts/SaberInteractions/WordInteractable.cs:34)
SaberInteraction.OnTriggerEnter (UnityEngine.Collider other) (at Assets/_Summer2024/WordSaber/Scripts/SaberInteractions/SaberInteraction.cs:12)

But i though having the if gameObject != null should prevent this issue

[knotty sun]

no, because gameobject is not null at that point

you need to move the Destroy(gameObject) back to the start of the call chain

Although if it is GameOver why even bother?

[slate imp]

I try this way

[slate imp]

because it throw an error and stop the game

[knotty sun]

no, I mean why bother destroying, a scene Load will clean up anyway

[slate imp]

moving it at the beggining is the same

[slate imp]

Its a VR game, the scene is the same, some buttons appears to restart the game or go back to the main menu, the Words disapear ( the words are entities of the game, its a game about writing words)

so they disapear when they are touched or the game is ended

[mellow sigil]

It might work if you check for if (this != null) instead

[slate imp]

i have a warning saying this is always != null

[knotty sun]

how can this be null and yet still be executing?

[mellow sigil]

I think what's happening is that the gameobject is destroyed but the method is called by an event so accessing gameObject throws the error inside this method instead of when calling the function

[slate imp]

by events, i would agree that this could be "partially" null

[mellow sigil]

Unity objects override the null check so destroyed object == null

[knotty sun]

the problem is he is trying to Destroy the object which is executing the current code

[mellow sigil]

No, the problem is trying to execute the current code on a destroyed object

[slate imp]

it worked ^^ thanks

events are strange

[knotty sun]

thet just means the Destroy never happens.
Use the debugger to check

[mellow sigil]

The object is already destroyed

[lament mural]

Hey guys! I'm currently running a spherecast to check if a position would be intersecting a collider if my player would teleport there. What I realised is that spherecast doesn't seem to detect a collider if it starts inside the collider. Is there a way around this?
I'm using a spherecast so that I know the player's body does fit in the area.

[mellow sigil]

Physics.CheckSphere

[lament mural]

Thank you for your reply @mellow sigil! I also found Physics.OverlapSphere() which does essentially the same but it turns out that yours fits my needs better.

[lean seal]

    Vector3 originalPosiition = transform.position;
    Vector3 fallPosition;
    float elapsedTime = 0;
    while (elapsedTime < fallDuration)
    {
        fallPosition = originalPosiition + (transform.up * (-fallDistance));
        transform.position = Vector3.Lerp(originalPosiition, fallPosition, elapsedTime / fallDuration);
        elapsedTime += Time.deltaTime;
        yield return null;
    }

like this?

[gusty aurora]

Is there a way to return a yield instruction ?

ie. if(bool) return "yield return null"
else return noyield ?

[knotty sun]

what would be the point of returning noyield. Just don't return

[gusty aurora]

I can't return nothing lol

[knotty sun]

so dont return at all

[leaden ice]

yield break; ends the coroutine immediately. maybe you want that? But you absolutely can just.. not return anything.

[lean seal]

yield return null

[gusty aurora]

Doesn't work. It skips a frame no matter if true or false

public static IEnumerator CheckFixedTimeStep()
{
if (Time.inFixedTimeStep)
{
yield return null;
}
}

[lean seal]

public static IEnumerator CheckFixedTimeStep()
{
    if (Time.inFixedTimeStep)
    {
        yield return null;
    }
  yield return null;
}

u getting error bc if if skipped u r not retuning anything

so u need else with return or return after if

if im not wrong

[knotty sun]

public static IEnumerator CheckFixedTimeStep()
{
    while (Time.inFixedTimeStep)
    {
        yield return null;
    }
}

[gusty aurora]

Doesn't work. Still skips a frame when Time.inFixedTimeStep is false.

print(Time.frameCount);
yield return CheckFixedTimeStep();
print(Time.frameCount);

=> 101 then 102 even when from Update

[vale bridge]

Bump

[knotty sun]

yes.
yield return StartCoroutine(CheckFixedTimeStep());
I presumed you knew how to start a coroutine

[gusty aurora]

Doesn't work, exactly same issue as above

[hexed pecan]

Maybe you want WaitForFixedUpdate..?
What is the point of this coroutine?

[gusty aurora]

delaying the beginning of the coroutine for one frame if and ONLY if it was started from fixed update

[knotty sun]

so why the separate coroutine what is wrong with

print(Time.frameCount);
if (Time.inFixedTimeStep)
    {
        yield return null;
    }
print(Time.frameCount);

Although I'm not convinced inFixedTimeStep applies to Coroutines

[gusty aurora]

And now how to move that to a separate method ?

It's the one I posted on my very first message, try to move it to a separate method and you'll see its behaviour changes

[leaden ice]

No this is not correct

[leaden ice]

Yielding another coroutine will always result in an extra frame

[maiden heath]

why does OnTriggerExit2D not show up in intellisense

[gusty aurora]

Seems so. Is that a bug ? Yielding an empty coroutine still results in an extra frame

[leaden ice]

Not a bug, just a consequence of the coroutine system

[cosmic rain]

Shamed why? Is it against your religion or something?

Even if it is, it's your personal matter of wether to use it or not.
I say if it works for you why not.🤷‍♂️
That being said, it might not always work for you, so learning the basics is still a must imho.

AI is still not at the level where it can write anything slightly complex for you. It can help you learn the concepts and write it on your own though.

[abstract nimbus]

has anyone bothered with calculating the distance between two points in any number of dimensions?

[late lion]

It's the same math for any number of dimensions.

[abstract nimbus]

I tried this but something seems to be odd with some tests

    public class VectorN
    {
        public double[] dimensions;

        public VectorN(int dimensions)
        {
            this.dimensions = new double[dimensions];
        }
        public VectorN(double[] dimensions)
        {
            this.dimensions = dimensions;
        }

        public static double Distance(VectorN a, VectorN b)
        {
            if (a.dimensions.Length != b.dimensions.Length)
                throw new Exception("The vectors must have the same amount of dimensions!");

            double sum = 0;
            for (int i = 0; i < a.dimensions.Length; i++)
            {
                sum += Math.Pow(a.dimensions[i] - b.dimensions[i], 2);
            }

            return Math.Sqrt(sum);
        }
    }

[heady iris]

Geometry can behave in non-intuitive ways in higher dimensions.

For example, the volume ratio between hyperspheres and hypercubes crashes to near-zero

What's the problem that you are seeing?

[abstract nimbus]

I can't explain lol

[heady iris]

so you have a problem, but you can't articulate what the problem is...?

[abstract nimbus]

idk how to prove that my Distance function works either

well the problem is a long story 😛

[cosmic rain]

You could at the very least make sure that it works with 2 and 3 dimensions by comparing to Vector2/3 distance or making some math on the paper.

If it works for these 2, it likely works for any dimension.

[heady iris]

If your problem is that your distances feel "too long", consider that [0.5, 0.5] has a magnitude of 0.707, whilst [0.5]_10 (so, a 10-dimensional vector where every value is 0.5) has a magnitude of 1.58

[abstract nimbus]

I will compare Unitys 3D distance with my 3d distance then

[heady iris]

This is why the hypersphere has such a tiny volume. It takes up very little space in the hypercube it's embedded in.

[abstract nimbus]

it does work for the 1st, 2nd and 3rd dimensions at least

maybe my problem isn't with the distance calculation then

[heady iris]

well, you'd better start telling it if you're still having problems!

[abstract nimbus]

I'm mainly experimenting so if it sounds crazy you have been warned

this is an XOR gate in 2 dimensions. Instead of a traditional neural network. This method uses dimensions as inputs and the output is the total distance between the black dots position and all the white dots. The size of the circles reduces the distance

so I'm trying to achieve this in any amoutn of dimensions

[heady iris]

And what's the issue?

Is the distance larger than you expected?

[abstract nimbus]

the total distance seems to be always positive when I increase the dimensions

and it struggels to learn at higher dimensions

in this case I tried with 12 dimensions

[heady iris]

You have encountered the CURSE OF DIMENSIONALITY 👻

https://en.wikipedia.org/wiki/Curse_of_dimensionality

[abstract nimbus]

lol

and I thought I came up with a really innovative way to make a neural network 😂

[heady iris]

If you add dimensions but keep the number of points constant, your density falls off a cliff

[heady iris]

and you're encountering this exact problem

The volume of a high-dimensional hypersphere with radius 1 is way smaller than the volume of a hypercube that's 2 units across

[abstract nimbus]

🤔

[heady iris]

It is non-intuitive, because you're trying to think about the only thing you can really comprehend: a 2-sphere inside a 3-cube

(so, a sphere in a cube)

[cosmic rain]

What about that setup involves "learning"?

[abstract nimbus]

it's just a mutation algorithm

[heady iris]

I presume there's some process that minimizes the distance

whoops, wrong reply

[heady iris]

(i hate how spheres are off by one)

[abstract nimbus]

and yes, that's the selection criteria for my mutation algorithm I guess

[heady iris]

If you keep cranking up the dimensionality, then points randomly positioned inside of a hypercube get further and further apart.

[abstract nimbus]

but they are not cubes tho? They are spheres cuz they have a radius

[heady iris]

I'm talking about randomly placing points

[cosmic rain]

Isn't that what neural networks do already with back propagation?

[abstract nimbus]

I'm just experimenting different ways lol

[heady iris]

that's one way to train a network, yeah

but there's no rule that says you can't just randomly wiggle around

or that dogs can't play basketball

[cosmic rain]

Yeah. What I'm trying to say is that the simpler options were probably already explored by many researchers before.

[abstract nimbus]

are you saying that:
a sphere in higher dimensions with radius x is bigger than a sphere in lower dimensions with radius x?

by size I mean distance between the spheres origo to its edge

[heady iris]

In fact: no!

They peak at a certain point and then drop to near-zero

At n=12, you're on the downslope

a edge-length-2 hypercube has a volume of 2^n, whilst a radius-1 hypersphere (which fits neatly into the cube)'s volume has a factorial in the denominator

which we call a "Big Oof"

[abstract nimbus]

wdym by factorial in the denominator

[abstract nimbus]

I know what factorial is but I'm confused in this case

[heady iris]

gamma extends the factorial function to non-integers

this function crashes to zero very quickly

High-dimensionality spheres have basically zero volume.

[abstract nimbus]

so I just increase the radius for higher dimensions

[heady iris]

you have to increase it a lot

[maiden heath]

how do i get rid of the box icon? im using aseprite importer

[heady iris]

That means that this object is a prefab produced by an importer.

If you unpack the prefab, it becomes a normal object

(which, of course, breaks the prefab connection)

[maiden heath]

how do i unpack it? right click?

[heady iris]

Why do you want to do this?

[heady iris]

https://en.wikipedia.org/wiki/Curse_of_dimensionality

see the "Distance function" section

[maiden heath]

isnt the box icon supposed to not be there

[heady iris]

it explains the exact problem you're encountering

[heady iris]

...no?

[maiden heath]

oh that was the impression i got.. i thought the box icon acted like a zip folder

[heady iris]

In this sense when points are uniformly generated in a high-dimensional hypercube, almost all points are much farther than r units away from the centre.

[heady iris]

If you unpack the prefab, then changes to sprites will still work -- since the object is referencing those imported sprite assets. However, any changes to the prefab will not be reflected.

[maiden heath]

oh i see

[heady iris]

this is more obvious with 3D models -- if you add another object in blender, the unpacked prefabs won't get the new object

[abstract nimbus]

the math doesn't give me much but I think I get the point you are trying to explain

[heady iris]

I dunno if this actually matters for the prefab created by the Aseprite importer. If it just makes an object with a sprite renderer and an animator, then there's nothing really stored in the prefab itself

it just references an Animator Controller and some sprite assets

in that case, it doesn't matter if it's unpacked or not

[abstract nimbus]

but I also just noticed that my algorithm struggles to learn the more "neurons" it has so increasing the dimensions means increasing the neurons

[heady iris]

"WeaponManagerController" is a mouthful

both "manager" and "controller" imply that they operate on many objects (especially "manager")

[heady iris]

To maintain the same density of points, you have to double the number of points every time you add a dimension

"GunPickupAndThrow" 's job is to pick up and throw weapons. The name should indicate this.

I remember suggesting "WeaponHandler" for that reason

[abstract nimbus]

why am I not allowed to say "hmm"?

[heady iris]

the server blocks a lot of no-content messages

[abstract nimbus]

lol

[heady iris]

I do get whacked by the "hmm" filter from time to time

[abstract nimbus]

not working either 😂

it's too smart

[heady iris]

I noticed you were creating points with values like [1, -1, 1] in that code you briefly posted

Those points are positioned on the corners of the 3-cube

[abstract nimbus]

yes cuz I'm converting the values with Tanh

so they are in range -1 to 1

[heady iris]

The distance from the n-cube's corners to the surface of the [n-1]-sphere gets larger and larger as you add dimensions

[abstract nimbus]

[n-1]-sphere meaning the edge of the sphere?

[heady iris]

a 2-sphere is the 3D sphere you are familiar with

hence n-1 when comparing it to an n-cube

a 3-cube is a 3D cube

go figure (:

[heady iris]

You very quickly hit a point where a unit sphere on a corner and a unit sphere at the center don't overlap

[abstract nimbus]

man time to explode my brain

@heady iris are you talking about this distance?

[heady iris]

Correct.

[abstract nimbus]

so that increases in higher dimensions

[heady iris]

That distance grows as you add more dimensions.

[abstract nimbus]

yeah it makes sense when you compare the distance in 2d vs 3d

[heady iris]

The part of a sphere closest to a corner will have coordinates that all have the same magnitude (either positive or negative)

adding more dimensions means that more numbers have to have a sum-of-squares value of 1, so the value of each number has to go down

[abstract nimbus]

so should I do something like radius = Pow(radius, dimensions) 😂

[heady iris]

sqrt(d * x^2) = 1

therefore, x trends to zero as d increases

[heady iris]

I suspect you'll have trouble getting meaningful results if all of your points have a very large radius

[plush musk]

Hiho, I'm trying on a PropertyDrawer for RectOffset that is used for HorizontalLayout padding.
I got it to draw like I want it to, but changes to the fields are not updated to the object. Any idea?

code: https://gdl.space/eqamopuwoy.cs

[heady iris]

link doesn't work

[abstract nimbus]

but what if I normalize them so they have the same relationship but stay between -1 and 1

[heady iris]

I'm guessing you forgot to call SerializedObject.ApplyModifiedProperties();, but that's a shot in the dark right now :p

[heady iris]

I don't know what that would mean

[wise pumice]

Hey guys, I was following along to a tutorial, but I'm having a couple of weird bugs.

The code is here - https://pastecode.io/s/3xjbhpp2

I'm having an issue where my AI like, stutters or something when arriving to his destination and I can't figure out why

I took a little gyazo video to show it and will list below.

One other thing I am having a problem with is, where we do this section in the synchronize

``bool shouldMove = Velocity.magnitude > 0.5f && Agent.remainingDistance > Agent.stoppingDistance;

    Animator.SetBool("move", shouldMove);
    Animator.SetFloat("locomotion", Velocity.Magnitude);``

My AI just starts sort of teleporting a bunch, I've fixed it for now by changing the last line to
shouldMove ? 1 : 0 (In the animator, the locomotion number was rapidly changing and was showing 1.5 at times, which shouldn't happen because we are getting magnitude?)

1st video - https://gyazo.com/a71b835c6f4b86fc81c64c8457e1362d
2nd video - https://gyazo.com/56e9d231184dcc130ade2735f9acd9c7

[abstract nimbus]

btw one funny thing I did was sending my code files to GPT and asking it how I would train it and then it gave me a Gradient Decent algorithm that worked with some tweaks here and there lol

so now I can use that instead of mutation

[heady iris]

Do you have root motion enabled on the animator?

If your animations contain root motion, they could be causing you to move around suddenly

[wise pumice]

Yeah the tutorial is for use with root motion nav agents

[plush musk]

I updated the link.

[heady iris]

Your code isn't using PropertyFields. It's just directly modifying the fields on a C# object.

I'm not sure what the correct move is here, actually

I'd ask this in #↕️┃editor-extensions

[plush musk]

okay, thanks

[heady iris]

The desired velocity is proportional to the distance to the agent's next position. That sounds wrong...

perhaps you wanted to normalize it?

I'd get rid of the smoothing, too, so that it's easier to spot the issue

by the way, you can just use transform.InverseTransformVector to turn a world-space vector into a local-space vector

[wise pumice]

I'm pretty fresh to it, so not really too sure what to actually do in these examples at the moment 😅

So in regards to the InverseTransform you mentioned, and normalizing. How would I make these changes?

I assumed the normalize you mentioned was for the shouldMove bool, but I'm not too sure atm

Still trying to come to grasps with it

[abstract nimbus]

@heady iris another way to solve the issue is by making the space more dense. Maybe that's what you tried to say before?

        public static SpatialNeuron GenerateRandomNeuron(int dimensions)
        {
            double radius = NeuralNetwork.GenerateRandomNumber(-1, 1);
            var neuron = new SpatialNeuron(new VectorN(dimensions), radius);
            for (int i = 0; i < dimensions; i++)
                neuron.position.dimensions[i] = NeuralNetwork.GenerateRandomNumber(-0.001, 0.001);

            return neuron;
        }

this is how I generate those spheres in x dimensional space. I also gave some spheres negative radius

I hardcoded 0.001 to test with 12 dimensions

[heady iris]

Vector3 worldSpaceVector = /* something */
Vector3 localSpaceVector = transform.InverseTransformVector(worldSpaceVector);

but the Vector method isn't quite appropriate here

this should be a direction

You shouldn't move twice as fast when you're twice as far away

So we should:

• normalize the vector
• use transform.InverseTransformDirection

Vector3 worldSpaceVector = /* something */
Vector3 worldSpaceDirection = worldSpaceVector.normalized;
Vector3 localSpaceDirection = transform.InverseTransformDirection(worldSpaceDirection);

The Direction method doesn't change the length of the vector you give it

the Vector method does (if you are scaled)

A normalized vector has a length of 1.

[wise pumice]

So, I understand the idea you are mentioning there, but I'm just confused as to where exactly I add/edit this in to the script I pasted?

Still wrapping my head around the tutorial so not sure what's what yet in this logic 😂

[heady iris]

you have a comment in your code that indicates exactly where you convert from world to local space

(:

note that after you do this, you'll still need to create a new Vector2

Vector3 localDirection = /* ... */
Vector2 deltaPosition = new Vector2(localDirection.x, localDirection.z);

(local Y is up-down)

X is right-left and Z is forward-backward

which is why you were doing this:

        float dx = Vector3.Dot(transform.right, worldDeltaPosition);
        float dy = Vector3.Dot(transform.forward, worldDeltaPosition);
        Vector2 deltaPosition = new Vector2(dx, dy);

[maiden heath]

i can use scriptable objects to make a dialogue system right? because im not sure how they work

[heady iris]

that implies your bullet points in its right direction

The LookRotation code would work for a bullet that points in its up direction

I've had a similar issue in 3D

If you shoot bullets from the gun's actual barrel, then you have to rotate the gun to point at what the crosshair is hovering over.

It's inconsistent up close, actually

That's where you need a very large rotation

At long ranges, it barely matters

[wise pumice]

Gotcha, thanks for that 🙂
One more thing which I'm not sure if you can help me with.

I like the random wander that the **DoMoveToRandomPoint" does, but I can't actually understand it at all. And currently, the AI will wander either short or massive distances on my plane.

I'd like to change it to be within a certain radius that it generates a point instead, but this one is completely different from the usual way I see

[heady iris]

oh yeah, I've seen this logic before (and used it)

The nav mesh is a...mesh. It's made of triangles.

This code is actually kind of wrong. It's picking a random vertex, then picking a vertex that's either one ahead or one behind it

The idea is to pick a random triangle and pick a random point inside of it

But this might pick a vertex from one triangle and a vertex from another triangle

which could be really far apart (or even give you a point that's out-of-bounds)

When implemented correctly, this code still doesn't work that well

[wise pumice]

Makes sense so far yeah

[heady iris]

It's biased towards denser parts of the mesh

you can one many triangles for a complex area and very few triangles for a simple area

An easy way to sample a random point would be to use NavMesh.SamplePosition

https://docs.unity3d.com/ScriptReference/AI.NavMesh.SamplePosition.html

In fact, the example is for picking a random point on the nav mesh!

Vector3 randomPoint = center + Random.insideUnitSphere * range;
NavMeshHit hit;
if (NavMesh.SamplePosition(randomPoint, out hit, 1.0f, NavMesh.AllAreas))
{
    result = hit.position;
    return true;
}

This picks a random point within range meters, then tries to snap to the nav mesh (moving at most 1 meter)

The larger the radius, the more expensive it gets to do the query

[wise pumice]

Okay, so just working out how to edit it to be like that atm, how exactly would center be defined in this case though?

Oh nvm actually lol

[heady iris]

you'd start from your current position

[wise pumice]

I was reading your section above and not the link

So I was missing some clues

[heady iris]

I have methods that look like this:

public IEnumerable<Foo> GetFoos() {
  yield return whatever;
  // ...
}

I want to add methods that look like this:

public void FillFoos(List<Foo> list) {
  list.Add(whatever);
  // ...
}

The latter will be used in situations where I can just fill up a list and then iterate over that, thus avoiding a bunch of allocations.

I'd rather not write the same logic twice. Is there a better way to avoid repeating myself than to replace the first method with this?

public IEnumerable<Foo> GetFoos() {
  var list = ListPool<Foo>.Get();
  FillFoos(list);
  foreach (var foo in list) {
    yield return foo;
  }
  ListPool<Foo>.Release(list);
}

[quasi edge]

Does anyone know how to edit the position of a textmesh pro gameobject in script?
So far i've tried:

myText.getComponent<rectTransform>().position += new Vector3(0, 1 0);
// and 
myText.transform.position += new Vector3(0, 1 0)

Both of them don't cause an error but they don't work

[heady iris]

Both of these are doing the same thing: myText.transform gives you the RectTransform

If this is on a "Screen Space - Overlay" canvas, then 1 unit might only be one pixel on your screen

Did you check the object's inspector to see if it's moving?

[quasi edge]

yeah, that was the problem

thanks

it didn't occur to me that it would be in pixels instead of units

[heady iris]

well, those are roughly equivalent here

[quasi edge]

I was using the same constant I had used for other gameobjects so thats why it barely moved

[heady iris]

you might have noticed that an overlay canvas looks huge in the scene view

Assuming there's no scaling going on, your position on an overlay canvas is your position on the screen, measured in pixels

so if you're 500 pixels from the left, your world position X is 500

[quasi edge]

oh i see

so in the scene view 1 unit is actually 1 piexel

that makes sense

[small oxide]

What is the best way to optimize the performance of a game? (cullings)
I tried with chunks, but I couldn't do anything

[spring creek]

It varies completely game to game.
Step one is always to profile though

[small oxide]

The game could be considered an open world, it has procedural generation, and I have the idea of ​​making the map relatively large

[spring creek]

Chunking is relevant to that.
Step one is still to profile

[small oxide]

What do you mean with profile?

[dusk apex]

The profiler

[spring creek]

You should look up a guide on how to use it. The hierarchy view is going to be the most helpful

[small oxide]

Okay, ty

[mighty void]

I am creating a procedural panera terrain using perlin noise, the problem I have is that I want the terrain to be generated in capable, there is no maximum height at which the mountains are generated, but there is a height which it cannot exceed, the problem I have a list with blocks which I will add blocks as I create them, the problem is that I don't know how to ensure that the blocks are generated correctly, because in this way blocks are generated on top of each other, any questions or comment or help would help me

[grim gyro]

Has anyone encountered this error?😶‍🌫️
The funny thing is that it occurs only in the web build. It doesn't happen on other platforms or in the editor

[gusty aurora]

Waow, I studied the full demonstration of this formula last year, didn't think it would ever be useful 😮

[maiden stone]

Hello, I heard some people saying that scriptable objects can be problematic in some cases. Does this mean I shouldn't use them for my game? and in what instances can it be problematic?
One more question. Im trying to make an audio handler. For this, should I use a scriptable object or singleton?

[heady iris]

that's incredibly vague

you can cause problems for yourself if you try to store data in scriptable object assets, though

If you load a new scene and nothing is referencing the object, it gets unloaded

any changes you made to it will be gone the next time it loads

I use ScriptableObjects (almost) entirely for read-only data

[late lion]

The way I like to think about ScriptableObject is that it's how you create your own asset types, just like how MonoBehaviour is how you create your own component types. More descriptive names for them would be something like CustomAsset and CustomComponent.

When you think about it like that, it sounds silly to say you shouldn't use ScriptableObject. It's the only way to create custom asset types, which can be very useful.

[maiden stone]

Okay, so then would it be better to use a singleton rather than a scriptable object to store / hold audio's temporarily for my audio handler?

[late lion]

Some of Unity's built-in asset types, like Material, are designed to be mutable (for better or for worse). It's not terrible, but can be confusing.

[late lion]

I don't understand what you mean by this.

to store hold audio's temporarily for my audio handler

[maiden stone]

My apologies. It was meant to be store / hold.

[heady iris]

Yeah, "CustomAsset" is a much better description

[heady iris]

Explain what this actually means.

What will you be doing with this object/singleton?

[maiden stone]

I am trying to create a system where I can use to play audio. This audio will be created as its own instance in game and put at a specific vector 3, based on where the method is called from. It can be called by any script. Audio, as well can play in mono instead of stereo. (Probably by either changing some settings or just setting the audio position constantly to the player position)

I'm thinking, I could have an array, or some data holding to store my audio/music (currently I have music being held in a scriptable object, and being sectioned in a property drawer in a specific order that it is supposed to play.) I'm trying to make a system, in which can handle multiple sounds, as well as music. I already have the script for reading these scriptable objects nearly created. When read, returns a sound file, which I would then like to import into my playaudo() method, and in turn, play the audio with any given parameters

My question; should I use a singleton, scriptable object, or script connected to a game object to handle this?

I hope this clears things up?

(I very well could be over thinking this)

[frozen ridge]

i'm trying to implement a building placement system,

If i have BuildingSO and Building classes, BuildingSO holds name, prefab, cost and Building has BuildingSO, does this create a circular dependency since BuildingSO needs the prefab and Building is on the prefab and needs BuildingSO

[rigid island]

how ? why does BuildingSO know about Building

[frozen ridge]

buildingSO has prefab, on the prefab there is Building class

i'm confused as well

[spring creek]

You don't want the class on the actual prefab

It should be on an empty just for placing, then you instantiate the specific building prefab when placed

[frozen ridge]

but why, shouldn't every building has a building class

oh wait

should i create an empty parent and make the prefab its child?

and then the parent will have building class

[spring creek]

Upon instantiation make it a child, sure.

That sounds reasonable

The prefab for the SPECIFIC building should just be the model and any specific behaviour that is unique to it

The parent you want to be generic. Building would be on EVERY building, so it belongs there with the parent

[frozen ridge]

let me think i don't think i understand

is there a better way to do it? I feel like i'm creating a problem that shouldn't exist

[rigid island]

isnt the prefab just a visual? whats it got to do with building class

[frozen ridge]

But i need the prefab to instantiate a building

[rigid island]

Im talking about the one on the SO

don't you have something like this

public class BuildingSO : ScriptableObject
{
    public string buildingName;
    public Transform visualPrefab;
    public int cost;
}```

[frozen ridge]

When initiating the building itself, i get the game object from the SO

[rigid island]

so the example I wrote basically?

[frozen ridge]

yeah just a GameObject instead of Transform

[rigid island]

alright. so I fail to see anything circular about this

BuildingSO has no idea about the Building class

in other words you're probably worrying about the wrong thing

[frozen ridge]

yeah i think i'm a little dumb tonight sorry

[timid brook]

I can't figure out why sticking to planets doesn't work...

I'm trying to make a outer wilds inspired character controller that walks on planets like in Mario Galaxy.
Everything was working fine up until I started rotating the planet along it's axis. The player being parented to the planet moves alongside it, however now the player doesn't rotate anymore, how can I fix this?

Video of non rotating planet: (player rotating works!)

Video of rotating planet: (player rotating doesn't work)

This is the code that attracts the player to the planet, what I think the problem is, is that body.rotation isn't getting set when the planet is rotating.

using System.Collections;
using System.Collections.Generic;
using UnityEngine;

public class GravityAttractor : MonoBehaviour
{
    public float gravity = -9.8f;

    private Quaternion targetRotation;
    private float rotationSpeed = 5f;

    public void Attract(Rigidbody body) 
    {
        Vector3 gravityUp = (body.position - transform.position).normalized;
        Vector3 localUp = body.transform.up;
        body.AddForce(gravityUp * gravity);

        targetRotation = Quaternion.FromToRotation(body.transform.up, gravityUp) * body.rotation;
        if (body.constraints != RigidbodyConstraints.FreezeRotation) return;

        body.rotation = Quaternion.Slerp(body.rotation, targetRotation, Time.deltaTime * rotationSpeed);
    }  
}

[frozen ridge]

does the player fall over when trying to rotate left or right?

[timid brook]

no

I have setup constraints

It's weird, when the planet isn't rotating it works fine, but after i start rotating it around it's axis... It's like body.rotation = Quaternion.Slerp(body.rotation, targetRotation, Time.deltaTime * rotationSpeed); this doesn't set the rotation

and i have tried without the slerp aswell, to no awail

[frozen ridge]

i think its either that or your targetRotation gets messed up because the planet is rotating

Since the player is parented to the planet its rotation or other values could be different from what you expect

[timid brook]

i think so too, but i don't know how to fix it

so i added a test float that increases after time to test what really happens```
using System.Collections;
using System.Collections.Generic;
using UnityEngine;

public class GravityAttractor : MonoBehaviour
{
public float gravity = -9.8f;

private Quaternion targetRotation;
private float rotationSpeed = 5f;
float test = 0;

public void Attract(Rigidbody body) 
{
    Vector3 gravityUp = (body.position - transform.position).normalized;
    Vector3 localUp = body.transform.up;
    body.AddForce(gravityUp * gravity);


    targetRotation = Quaternion.FromToRotation(body.transform.up, gravityUp) * body.transform.rotation;
    Debug.Log(targetRotation);
    if (body.constraints != RigidbodyConstraints.FreezeRotation) return;

    //body.MoveRotation(Quaternion.Slerp(body.rotation, targetRotation, Time.deltaTime * rotationSpeed));

    test += Time.deltaTime;
    body.rotation = Quaternion.Euler(new Vector3(0,test));

}  

}

[languid hound]

Does anyone see something wrong with this?

using System.Collections;
using System.Collections.Generic;
using UnityEngine;

public class CollisionTest : MonoBehaviour
{
    public Transform target;
    public Vector3 lastPos;
    public float sphereScale;
    public LayerMask layersToCollideWith;

    public void Update()
    {
        transform.position = target.position;

        if (PathCheck(out RaycastHit hitInfo))
        {
            Debug.Log("ok");
            transform.position = hitInfo.point + hitInfo.normal * sphereScale;
        }

        lastPos = transform.position;
    }

    public bool PathCheck(out RaycastHit hitInfo)
    {
        if (Physics.SphereCast(lastPos, sphereScale, (transform.position - lastPos).normalized, out hitInfo, (transform.position - lastPos).magnitude, layersToCollideWith))
        {
            return true;
        }

        return false;
    }
}

I'm trying to sweep from the last to the current position to detect if it's penetrating something. Just as a test I'm trying to push something out

[timid brook]

this is how it should work increasing the y rotation every frame (this is with planet rotation off)

this is with it on, it's weird how it jumps about

[timid brook]

sorry for writing over you

[timid brook]

you can draw the sphere cast using gizmos, maybe that should help you

[hexed pecan]

Are you having problems? It looks fine to me, assuming that transform.position = target.position; is there for a reason.

[languid hound]

Yeah the transform is meant to follow a target since this is just a test

[languid hound]

I did try something similar and it generally looked fine which is why I'm confused

[hexed pecan]

Actually you have this:cs transform.position = hitInfo.point + hitInfo.normal * sphereScale;
Depending on the behaviour you are after, you might want something like:cs transform.position += hitInfo.distance * directionNormalized;

Your current way of doing it kind of pushes it out in the normal's direction

Which may end up being further than you expected

If I'm thinking clear right now 🤔

[languid hound]

Yeah the idea is it goes to the point it hit and then gets offset by the hit normal by the sphere's size which should in theory push it out completely in the right direction

Okay now I'm confused wtf

using System.Collections;
using System.Collections.Generic;
using UnityEngine;

public class CollisionTest : MonoBehaviour
{
    public Transform target;
    public Vector3 lastPos;
    public float sphereScale;
    public LayerMask layersToCollideWith;

    public void Update()
    {
        if (Physics.SphereCast(transform.position, sphereScale, transform.right, out RaycastHit hitInfo, sphereScale, layersToCollideWith))
        {
            Debug.Log("!!!");
        }

        Debug.DrawRay(transform.position, transform.right, Color.yellow);
    }
}

I tried rewriting it from the start but for some reason it's saying that it hits the cube but only when it's further than the sphereScale (a unit away than it should be able to reach)

[simple egret]

It's normal, you're passing the scale (radius) into the distance parameter. Think of sphere cast as a thicker raycast, it still has a distance. The cast extends sphereScale units towards the right

[languid hound]

I'm not though

[simple egret]

maxDistance param

[languid hound]

Wait

[simple egret]

Consider using CheckSphere() if you don't need info on what you hit (just whether you hit or not), or OverlapSphere() if you do

[languid hound]

I need precise information on the rayhit which OverlapSphere can't give afaik

I'm still confused because this is using OverlapSphere as maxdistance
Even when the ray intersects it just doesn't log until it doesn't intersect

[simple egret]

Switch to 3D view, make sure the object isn't rotated around Y. Seems like it is, the yellow debug line is shorter than 1 unit from this perspective

Also temporarily get rid of the layer mask parameter to rule out a bad mask

[languid hound]

Even without the layermask param it still doesn't log until the line stops intersecting

Sorry wireframe was the clearest way to show this to me. This is a shaded view

[hexed pecan]

Your line is 1 unit long, it doesn't correspond to the spherecast. It only shows the direction, not length

[languid hound]

But sphereScale is also only 1

[hexed pecan]

So is it intentional that you spherecast with the sphere's radius as the max distance?

[languid hound]

Yeah

[heady iris]

Doesn't the spherecast ignore any colliders it starts out inside of?

[hexed pecan]

Yes

[heady iris]

Also, you're using a new enough version of the editor to turn visualize queries with the physics debugger

That may be illustrative.

[languid hound]

Ooh I haven't heard about that

[hexed pecan]

If the radius is 1 then it definitely starts inside that box

[languid hound]

Sorry should clarify that I moved it. It starts on the other side of that box 5 units away

[hexed pecan]

Okay, and your max distance is 1 so it wouldn't hit it

[languid hound]

Yeah that's why I moved it

[languid hound]

Okay so this helped a TON

I'll just clean some of this up and then mention what was wronf

Okay yeah no the only issue I had was that sphereScale needed to be divided by two

I thought a SphereCast with a radius of 1 would be the same as a sphere in unity

[heady iris]

presumably you mixed up your diameter and your radius

[languid hound]

Yep rip lol

[heady iris]

yeah, unity spheres and cubes have diameters of 1

(well, cubes have edge lengths, not diameters, but same idea)

[frozen ridge]

How would you guys go about a building placement system, no matter what I think I can't imagine a way that I dont use a manager. Do you guys think its bad practice to use managers

[heady iris]

"managers" are perfectly fine

you're obviously going to have something that's in charge of how buildings are getting placed

[frozen ridge]

It's just that whatever I read they make so much emphasis on encapsulation, I feel like I break it when I use a manager instead of handling it in the object, but as you say how am I even supposed to place buildings without a manager

[late lion]

You can still maintain encapsulation. The manager doesn't need to be able to access the internal state of the buildings in order to place them. Encapsulation doesn't prevent cross-object interaction, it facilitates it.

[heady iris]

Exactly.

[frozen ridge]

But I do access the internal state, at the very least I set the building's state to placed if I was able to place the building

[heady iris]

You'd tell the building that it got placed down.

you're asking "how do I write methods?" at this point :p

[frozen ridge]

I know what I ask feels like a basic and fundamental knowledge and maybe it isn't even that important but I guess it's better that I ask lol

[heady iris]

You might want to read the C# Learn docs about object-oriented programming

https://learn.microsoft.com/en-us/dotnet/csharp/fundamentals/tutorials/oop
https://learn.microsoft.com/en-us/dotnet/csharp/fundamentals/tutorials/inheritance

[frozen ridge]

I feel like I can implement complicated things reasonably well but I have problems when I try to use them together, like I can implement pathfinding algorithms pretty well and modify them however I like but when I try to come up with a system that can move units using different pathfinding algorithms and keeping it decoupled I have problems

[frozen ridge]

Maybe my problems are more fundamental tho I'll take a look thanks

[shut pagoda]

Could anyone help me make a roll mechanic (like, roll a few units in a direction with editable stats and the roll decreases your collision by a bit)? I tried making it myself, looking for it online even using chatGPT but I can't for the love of god make it work (2D)
this is my current code https://gdl.space/eguqutipav.cs

[wispy creek]

Good night, an error occurred with the player movement, the player flies with/among the camera orientation, i will be sending all the scripts and relevant photos below:

Player Scripts:

[leaden ice]

you've attached the "Camera Controller" script to the player object

that doesn't seem right

[wispy creek]

using UnityEngine;

public class PlayerMovement : MonoBehaviour
{
    public float moveSpeed;

    public Transform orientation;

    private Rigidbody rb;

    void Start()
    {
        rb = GetComponent<Rigidbody>();
        rb.freezeRotation = true;
    }

    void FixedUpdate()
    {
        Move();
    }   

    void Move()
    {
        float moveX = Input.GetAxis("Horizontal");
        float moveZ = Input.GetAxis("Vertical");

        Vector3 moveDirection = transform.right * moveX + transform.forward * moveZ;
        rb.MovePosition(rb.position + moveDirection * moveSpeed * Time.fixedDeltaTime);
    }
}

[wispy creek]

exactly, i had it solved before, but i honestly forgot if i had a emptyObject for the camera controller (player's child)

using System.Collections;
using System.Collections.Generic;
using UnityEngine;

public class CameraController : MonoBehaviour
{

    public float sensX;
    public float sensY;

    public Transform orientation;

    float xRotation;
    float yRotation;

    void Start()
    {
        Cursor.lockState = CursorLockMode.Locked;
        Cursor.visible = false;
    }

    void Update()
    {
        //get mouse input
        float mouseX = Input.GetAxisRaw("Mouse X") * Time.deltaTime * sensX;
        float mouseY = Input.GetAxisRaw("Mouse Y") * Time.deltaTime * sensY;

        yRotation += mouseX;

        xRotation -= mouseY;
        xRotation = Mathf.Clamp(xRotation, -90f, 90f);

        //rotate cam and orientation
        transform.rotation = Quaternion.Euler(xRotation, yRotation, 0);
        orientation.rotation = Quaternion.Euler(0, xRotation, 0);
    }

}

[leaden ice]

your script is rotating the actual character object here

so of course when you move forward it will go in the air

when you're facing in the air

[wispy creek]

how am i solving it?

[leaden ice]

Your code looks like it's intended for a first person game

[wispy creek]

do i create a child "Player Camera" for "Player" with the camcontroller??

[leaden ice]

you're making a third person game

that script is not suitable at all

[wispy creek]

it swapps between both cameras

[leaden ice]

you're using the FPS script in TPS mode

don't do that

[leaden ice]

your script should be movingt the camera

not rotating the player

[wispy creek]

ohh

[wispy creek]

i see

[wispy creek]

i meant gameObject*

now the camera doesn't follow the mouse

[wispy creek]

but isn't there a script where i can get both fps and tds working?

[leaden ice]

You can write one

[wispy creek]

i'm working on that, since i need to do a little research about the sync

[mighty void]

Someone could help me with a terrain generation?

[hexed pecan]

Ask your question

[mighty void]

I want to do something like this

But when i try this happens

Like layers

[hexed pecan]

You should probably check if(y < localY && y > localY - thickness)

Where thickness is how many blocks the layer should be in the Y dimension

If you want something like in the first screenshot, that is

[mighty void]

And in that way i can make the correct terrain? I'm going to try

And I have another question,maybe i have to try to make first the foreach and inside make the for (int y) ?

[hexed pecan]

What is BlockData? Is it one vertical slice of the terrain?

Or one block?

In any case you seem to be a lot of unnecessary iteration here

For every cell you check every blockdata?

Maybe the blocks should be in a 2D array, or they should know their own position in the terrain

You could simplify the loops a lot

[mighty void]

IS a block class, contain values like the Sprite, the Minimum level yo spawn, the height level and things like that

Some bool values

[hexed pecan]

You are currently creating a block for every blockdata

Maybe associate a depth value for each blockdata if you want like different types on different layers

[mighty void]

Yes

[mighty void]

And I Would use that instead of thickness?

[hexed pecan]

Yeah, maybe something like if(y > localY - data.maxDepth && y < localY - data.minDepth)

Where min/maxDepth are the depth or distance from the ground level

[mighty void]

That would me work with múltiple blocks? Like adding grass Stone Clay infinite blocks?

[hexed pecan]

That's the point, you add different depth values to different BlockDatas

[mighty void]

I dont have a fixed listo of blocks

And how put in? Like i am doing in the method createblock?

[hexed pecan]

Replacing the current if(y < localY) is the idea

This is the simplest way I can think of - there are improved alternatives

[mighty void]

How?

[mighty void]

Don't really works

Look

[mild goblet]

question to the hive mind. Has anyone created a movement/collision detection system without physics (with the intention of better performance) ?

[spring creek]

I guess how do you define physics? Changes in position and collisions are physics. Do you mean not using the physics ENGINE?

[leaden ice]

I think collision detection is inherently "physics"

but uh - sure I've done lots of grid based games and stuff that didn't involve the physics engine

[shell scarab]

How do I wait for a list request to finish if I can't use coroutines in the unity editor :(

they could have at least made it have an event thats called when finishing...

[somber nacelle]

you could always use coroutines in the editor

also #↕️┃editor-extensions for editor related coding

[shell scarab]

i guess, it feels clunky that I have to install a package to wait for something asynchronously in their base API.

[leaden ice]

wdym by "list request"?

[somber nacelle]

i assume they are referring to this

[rancid frost]

Any ideas on how to represent this in unity?

A standard 2d array wont work due to the values(the biomes) not being designated points
I was thinking of creating a Bounds and having the biomes contained within said bounds.
Any other ideas?

[vivid halo]

Piecewise functions might be a suitable yet inelegant solution.

if (temp <= -5) biome = Biomes.Tundra;`
if (temp <= -1)
{
  if (annualPercipitation < 5) biome = Biomes.Desert;
  else if (annualPercipitation < 20) biome = Biomes.TemperateShrubland;
  else if (annualPercipitation < 130) biome = Biomes.BorealForest;
  else if (annualPercipitation < 150) biome = Biomes.Tundra;
}
// This is probably neater with switch statements.

Pair it with some unit tests that test the diagram thoroughly and you can be sure it is working nearly as intended. If you can access the source data methodology, you might be able to find functions that represent the outputs, but it's not trivial to determine the functions and try to create an elegant (x, y) => biome mapping just based on the image alone.

I am working on a CameraController that has met requirements up until now: https://pastebin.com/H9M98ybh

The issue is that the current implementation does not work when the player is parented to another object, such as a rotating room, because the _currentYaw and _currentPitch private variables are not compatible with being rotated from an external source, and I am unsure how to approach implementation or refactoring, specifically regarding how to rotate the camera swivel externally in a way that is compatible with being externally rotated.

[wooden lintel]

i have a building that can be upgraded to a much bigger, nicer looking building. so i'm trying to bake the lighting with the upgraded building enabled and the original building disabled. then i make duplicates of the lightmap textures and isolate them in a unique folder and bake again with the opposite, upgraded building off and original building on.
when the player upgrades the building it's supposed to switch from the original building lightmaps to the upgraded building lightmaps. but for some reason the upgraded building lightmaps don't look like how it does when i bake the scene with it enabled.
i've been using chatgpt to help me figure out how to do it in the first place but regardless of cgpt's help i can't figure out what's happening. everything seems to work and be linked correctly when the building is upgraded but it just doesn't look like how it should.
how can i troubleshoot this or is there a better way to accomplish what i'm doing? also I'll link the code i'm using

https://hastebin.com/share/nojizuwoyo.csharp

oh and all my lights are set to Mixed mode and I'm baking with Subtractive mode

[pastel patio]

Hi guys, anyone know how to accurately get whether something is exposed? Even just a little bit, but I need to know whether they're completely behind cover

I need it for my 2d game where opponents behind objects will be hidden in network

May be a bit inaccurate in terms of opponents still being detectable a bit out of range, but it'd be bad if they're undetectable within view

[fleet gorge]

raycastall

sort the raycastall array based on distance, foreach object in the raycast, if the object is cover break the loop

but if it does hit an entity then damage them

[gaunt holly]

Hey, someone know how to make roads on Terrain like roads of GTA please ? It’s a texture who add ? or other ? thank you ping me please !

[fleet gorge]

roads on terrain?

using unity's in built terrain system? the most you can do in unity built in terrain system is paint on textures

youll need to use the 3d models of your road

@gaunt holly

[pastel patio]

Haha could just use a simple raycast to check if 1 point is obstructed

[fleet gorge]

recently i was thinking about how to calculate wallbang

[pastel patio]

Or use the Out parameter to get info on that what's been hit

[fleet gorge]

in competitive tac shooters bullets can penetrate walls

im thinking about how to calculate that

[pastel patio]

Dot product

At least, if the condition is based on the angle your bullet has hit the wall with

However what I was asking is essentially if any point of an opponent is not obstructed

In my case specifically, can say that I wanna know if the target is inside the 2dlight cast by the player

However light doesn't have this function

[knotty sun]

why not use a boxcastall from the player to the enemy. if the size of the box is the same as the enemy then it should pick any part of it exposed

[pastel patio]

Doesn't box cast get blocked by the closest hit object instead?

[knotty sun]

I said BoxcastALL

[pastel patio]

xD But box cast all means all targets in the boxcast's path

[knotty sun]

yes, which is what you want is it not?

[pastel patio]

I mean, I suspect that it does not account for obstruction

[knotty sun]

then it is a simple bounds check to see what is obsuring what

[pastel patio]

Instead gives all the box cast results as if the rest doesn't exist

[pastel patio]

What is a bounds check in this context?

[knotty sun]

colliders have a bounds property. I suggest you go and read the Bounds docs

[pastel patio]

Haha well, apparent this bounds system is straight

Basically like taking the min and max of it

While the calculation in this case includes opponents being at 45 deg direction from you etc

[maiden heath]

is there a way to code my own node based dialogue system graph thingy?

[fleet gorge]

yes its possible, lots of work but many companies have it already

theres a few tutorials on it actually

https://youtu.be/nvELzBYMK1U?si=KmayjEzIIo3f1UCf

first video i found on it

[maiden heath]

oooh thank you!

this seems really cool

easier than having a shit ton of if statements for my dialogue

[fleet gorge]

for myself i was lazy to do the whole thing so instead I used SOs

QuestManager can hold a selectedQuest.
Quest has a list of Objectives.
Objectives include Dialogue, Location(go to an area to advance), Interact, Puzzle etc

to start you can create a Quest, and add to the list of objectives within the quest. create a Dialogue scriptable object and drag it into the list of objectives. in the Dialogue scriptable object, you can add the lines that you want in a unity list.

I realised that its not very flexible though so I might combine both solutions eventually

[magic harness]

Hello guys, im having trouble with my game which uses mouse click to move, where the clicks pass through the UI. i tried using IsPointerOverGameObject on my floor(which has a NavMesh) but it returns true even if it is not over any Ui Object. Debugging raycast its really only catching the plane.

Can anyone help?

[slate imp]

Hi there, I've been making a game in VR, I'm looking for around 60 fps for a good experience, but as you see it's around 45, the part to look is in the center, on the left is when i was in a menu and on the right i was looking at the sky.

I have two main questions :

• What is PlayerEndOfFrame that take some time in the next frame
• What is Gfx.WaitForRenderThread that take a Huge time

Well, i know what globaly is WaitForRenderThread, but i dont know how to have a better profiling of this, neither to how to reduce the WaitForRenderThread

[dry violet]

It seems to me that it is because of the vsync that the headset has, reducing the render load will possibly improve it

[cosmic rain]

You can have a look at the render thread

[maiden heath]

is graphview api deprecated?

[cosmic rain]

Use the hierarchy mode, as it's easier to see what takes time.

[slate imp]

the one that is completly empty ? because their is not one else

[cosmic rain]

Yes, that's why you should use the hierarchy mode.

[slate imp]

i dont have vsync enabled

[slate imp]

this is the hierarchy mode of the render thread

[cosmic rain]

Seems like it's mainly waiting for the GPU to finish it's work.

You could try looking at the GPU module if it works on your hardware

[slate imp]

i dont know how its work

[dry violet]

If you do the build in the glasses, does it also happen? it could be because of the connection

[cosmic rain]

The profiler has different modules. On your screenshot you have the CPU and rendering modules open. You can add and enable other, like the GPU one.

[daring cove]

Hey,

I need help with presenting the players joystick input as a point in a circle.

The joystick input is a Vector3. (pitch, yaw, roll)

I need to translate the input in the circle so that
When yaw is 100%
the point is at the right of the circle.
When pitch is 100%
the point is at bottom of the circle.
When roll is 100%
the point is in a diagonal position on the circle.

I have solved for x and y position.

point.x = root.x + joystickInput.x * circleRadius;
point.y = root.y + joystickInput.y * circleRadius;

the input axis can be negative, but always between -1 and 1.
the point is in 2D space.

The zero position would be the middle of the circle

[slate imp]

If i build i also have low FPS

[gaunt holly]

Sorry, I saw your answer late. I don’t know, I suppose to create basic terrain and after build roads on it but with unity system or add 3d model I created. What is the best to have the best similitude with GTA ?

[dry violet]

3d model

[slate imp]

i can't use GPU profiling with Vulkan

[gaunt holly]

Okay, so I must create portion of road in 3d and add it on my terrain ?

[daring cove]

3d model interpolated with a spline is how they did it in GTA I think. In my own experience making roads by placing models manually is the pain of a lifetime. So look in to ways of automatically making this process.

There are most likely great tools on the asset store for this

[dry violet]

yes like a kit, There are addons in blender to make roads

[gaunt holly]

okay thank you Predator and Sami L

do you have video to help me about this ??

[cosmic rain]

Then look up platform specific profiling tools

[cosmic rain]

Other than that, you just need to optimize the GPU processing time.

[slate imp]

i'm doing some research, i'm not sure where i can find some, i'll look

[slate imp]

yep that's the issue

[cosmic rain]

Try looking for vulkan specific tools.

[slate imp]

yep thanks

[daring cove]

Found this. Its a unity forum article.
https://forum.unity.com/threads/best-approach-to-creating-dynamic-roads-that-follow-a-path-of-data-provided-shaders.860272/

[dry violet]

https://youtu.be/vgqtYgrVTRQ

[dry violet]

https://assetstore.unity.com/packages/3d/characters/easyroads3d-free-v3-987

[maiden heath]

can i reverse an animation in unity using animator?

[dry violet]

speed -1

[maiden heath]

animator.speed?

[dry violet]

animation.speed

[maiden heath]

is there a way to do it from the animator

[dry violet]

No idea

[maiden heath]

oh

ok well thanks anyways

i set animator.speed = -1 but it didnt work and i got a warning

[dry violet]

you have to change the specific animation you want to invert

[maiden heath]

okay

[gaunt holly]

Thank you @daring cove @dry violet

[lofty summit]

I have a Spline object (from inside a SpriteShape), is there a way to get points inside the spline like "traveling" it? Not just the control points, but intermediate points inside the spline 🤔
GetPosition only asks for an integer and returns the control point, I need something like... evaluate?

spline 2.5 had evaluate but 2.6 doesnt? 😐
https://docs.unity3d.com/Packages/com.unity.splines@2.5/api/UnityEngine.Splines.SplineUtility.Evaluate.html

[trim schooner]

Do GetPosition and GetPositionTangentNormal replace parts of it?

[lofty summit]

Oh, I think that my problem is that Spline is not a UnityEngine.Splines.Spline but a UnityEngine.U2D.Spline 😐

yep, different splines 🤦‍♀️

yay, fun 🥲

Ok, I can make an in memory UnityEngine.Splines.Spline from the control points from UnityEngine.U2D.Spline

[lapis acorn]

does Unity has an option to auto initialize a prefab on game start?

In Godot for example there are "autoloads". Does Unity has something similar? I want whenever i start playing a scene, that specific gameobjects and scripts are added to it as DontDestroyOnLoad

[mellow sigil]

Just put them in the scene?

[lapis acorn]

I have several managers. When i want to test some things isolated, I do not want to add all these dependencies just to test a small thing. they should be automatically added whenever i need them

[heady iris]

I handle this with a SingletonPrefab class.

https://gdl.space/omawetizel.cpp

The Safe property is used in places where I don't want to create an instance (like while the game is ending)

Otherwise it's straightforward enough: this loads a prefab from a Resources folder

My GameController class bootstraps itself like this:

        [RuntimeInitializeOnLoadMethod(RuntimeInitializeLoadType.BeforeSceneLoad)]
        private static void OnLoad()
        {
            Instance.Initialize();
        }

[lapis acorn]

interesting 😄

[heady iris]

It then loads other singleton prefabs, like the menu UI and the input controller

[vale wharf]

I think you're referring to a pattern called "bootstrapping", where you load required scripts/assets when the scene loads instead of having to manually place those assets in the editor every time you make a new scene.

For my game, I do this:

public static class App
{
    [RuntimeInitializeOnLoadMethod(RuntimeInitializeLoadType.AfterSceneLoad)]
    public static void Bootstrap()
    {
        bool shouldBootstrap = GameObject.Find("NoBootstrap") == null; //Don't bootstrap scenes with a NoBootstrap GameObject
        Vector3 bootstrapPos = GameObject.Find("BootstrapPos")?.transform.position ?? default; //Allow for defining a specific pos for the bootstrap prefab to be spawned at
        GameObject target = GameObject.Find("Bootstrap"); //Check if bootstrap prefab is already loaded

        if (target == null && shouldBootstrap)
        {
            target = UnityEngine.Object.Instantiate(Resources.Load("Bootstrap")) as GameObject;
            target.transform.position = bootstrapPos;
            if (target == null)
                throw new ApplicationException("No app prefab found in resources!");
            //Don't destroy bootstrap prefab when a new scene loads
            UnityEngine.Object.DontDestroyOnLoad((UnityEngine.Object)target);
        }
    }
}

you can then implement the singleton pattern on specific scripts that are loaded in the bootstrap

though be warned, I think this will cause the prefab to be instantiated AFTER Awake() and OnEnable() is called on monobehaviours in the scene, but before Start() is called.

so for example, if one the objects in the bootstrap prefab is the main camera, anything trying to get the/find main camera from Awake or OnEnable will find nothing (null).

[hallow valve]

Do you guys know why it's giving me this errors when I build the project?

This is the code:

using System.Collections;
using System.Collections.Generic;
using UnityEngine;
using UnityEditor;
using TMPro;

[CustomEditor(typeof(FieldOfView))]
public class FieldOfViewEditor : Editor
{
    void OnSceneGUI()
    {
        FieldOfView fov = (FieldOfView)target;
        Handles.color = Color.white;
        Handles.DrawWireArc(fov.transform.position, Vector3.up, Vector3.forward, 360, fov.viewRadius);
        Vector3 viewAngleA = fov.DirFromAngle(-fov.viewAngle / 2, false);
        Vector3 viewAngleB = fov.DirFromAngle(fov.viewAngle / 2, false);

        Handles.DrawLine(fov.transform.position, fov.transform.position + viewAngleA * fov.viewRadius);
        Handles.DrawLine(fov.transform.position, fov.transform.position + viewAngleB * fov.viewRadius);

        Handles.color = Color.red;
        foreach (Transform visibleTarget in fov.visibleTargets)
        {
            Handles.DrawLine(fov.transform.position, visibleTarget.position);
        }
    }
}

[hallow valve]

on vscode I have no errors

[runic island]

Hey I'm getting some weird errors that aren't making sense to me. Maybe someone with more experience can point out what I'm doing wrong? Been working on this since yesterday. I got majority of the issues worked out but I'm at a loss for these last three.

The code is from a 2023 tutorial so maybe some things in TextMeshPro changed since then but it's just not clicking for me, I've been working on it since yesterday to no avail. https://pastebin.com/J6P8Jb1Q

I put the error codes in the comments along the code.

Here's the entire class so far if it makes a difference. https://pastebin.com/g8zxPuVW

[knotty sun]

just use TMP_Text which is applicable to both UI and World TMP Text objects

[naive swallow]

You cannot include anything in the UnityEditor namespace in a build. It needs to be excluded either by putting it in an Editor folder or using Assembly definitions

[hallow valve]

ok ok, thx

[runic island]

I see.. Since my game is 2D, do I even need to use TextMeshPro at all? If I understand the API it's used for text in a 3D space, and all my text is going to be on the canvas so I should be good to just use the TextMeshProUGUI.. I think? If that's the case then I can simply get rid of the second constructor and everything should still work. I believe I do want TextMeshProUGUI for the addition event handling and UI integration over TMP_Text.

If any of this is way off-base please correct me. I'm mostly trying to get an understanding of how the typewriter effect is done for text panels as well as click and pointer interactions which is why I'm not using an addon.

[knotty sun]

why do you need TextMeshProUGUI? TMP_Text is the standard way of using TMP Text variables

and yes, get rid of the overload

[runic island]

why do you need TextMeshProUGUI? TMP_

[mental solar]

Hi, I'm having issues with the new Input System.
I'm rewriting all legacy OnMouse...() buttons to use the new Pointer events, but some buttons work and some don't, how do I debug this?
I've added this code to a script which is attached to the Main Camera game object

[heady iris]

OnPointerDown is not part of the new input system

https://docs.unity3d.com/Packages/com.unity.ugui@1.0/manual/EventSystem.html

That's part of the Unity UI package.

are you trying to use this for non-UI stuff?

explain what you're doing

[mental solar]

trying to port a game to consoles, mostly
the developer put multiple canvases and colliders in a scene, the buttons are a pair of Sprite + Text + Box Collider 2D + some script with OnMouseDown
and with OnMouseDown the "buttons" work fine, but with the new Input System UI Module and OnPointerDown some buttons work (mostly those that are at the top of the screen) and most of them don't

[small pollen]

Can somebody help? the thing is, i have a "mute" button that when clicked, plays a sleeping animation and mutes the audio source. but when i do click it, all the gameobjects play a sleeping animation and mute

even though i want to be able to individually mute them

[naive swallow]

Show code

[small pollen]

lemme grab it rq

not sure which to give so im jus givin em all

its shitty i know

(i use Spine animation)

And also

when the timeline requests the gameobject to play an animation it plays that animation instead of the mute animation

[heady iris]

None of those are UI components. They do not make sense on a canvas.

OnMouseDown isn't a UI thing either -- it works on physics colliders (which is why it was working previously)

[mental solar]

here's how a typical button in this game looks like:

I guess the "best" solution here would be to read Mouse.current directly, then raycast manually and dispatch events to gameobjects with a custom system

but maybe there's a better solution which I just don't know - I really don't want to re-do these buttons because this isn't my game and isn't my code

[naive swallow]

Is there a "proper" way to modify package code without just copying the whole package into Assets from the package cache?

[heady iris]

Copy it into Packages/

Unity will recognize that it's an embedded package and get rid of the copy in Library/PackageCache

and it will still understand that you have the package in your project