archived-code-advanced

[thorny mortar]

yeah right, But I am getting it in zip file because I get many audio clips and pngs by one request to the database instead of getting every file 1 by 1

you mean saving the AudioClip.getdata() samples and load it when needed?

[scenic forge]

Yeah, data in audio clip are PCM already, you can save and reuse them directly.

If you get things in one big zip though and you have no control over the audio format, then I don't see a way other than using an audio decoding library.

[wooden niche]

Also posted the question to JetBrains. They've created a feature request ticket for this: https://youtrack.jetbrains.com/issu...02.806161068.1685517004-1179846385.1685517004

Pls upvote this to make it happen 😉

[dusty wigeon]

For the love of god, do not make Invoke(..., 0);

It seem that you have a dependence between "Piece" and "Board" but some reason you do not make the dependance explicit. This is terrible design.

Personnaly, I would have my piece hook to an event on the Board itself. Do not make things like afterRestart because you gonna have afterAfterAfterRestart at some point.

[bleak citrus]

yes, this is exactly what events are for

[fickle zodiac]

has anyone experienced their signing key for google play games, not working anymore?

[regal olive]

Hi yall I've got some unusual problem with Sprite Shapes. So basically i try to spawn bunch of collectible items using RaycastHit2D, but they end up being on wrong positions. It happens because Sprite Shapes for some reason don't update their colliders when they are off camera. When im observing gameplay through scene mode all Sprite Shapes update their colliders corectly because i see them. The problem Is that i have to spawn them ahead of camera because that would be weird for collectibles to pop into existence in front of player right 😛 ?

[sly grove]

SpriteShapeController has a BakeCollider method you can call

[regal olive]

@sly grove that doesnt work, just solved this with extending render bounds

[sage radish]

Just going through the 2022.2 changelog and struck gold:
https://docs.unity3d.com/2022.3/Documentation/ScriptReference/Physics.ContactEvent.html
Finally a static event alternative to MonoBehavior.OnCollision[Enter/Exit/Stay] (with some limitations)

[austere jewel]

Gotta keep up with the previews sections of the forums to find out about these things earlier https://forum.unity.com/threads/new-contact-dispatch-for-physics.1254033/

[untold moth]

Wait, is the 2022 lts out yet?

[austere jewel]

Yes

[buoyant dawn]

Alright, so I'm trying to make chunk based procedural generation (sorta). basically, I want to have a bunch of premade say 100m x 100m blocks that I can procedurally generate a world out of. So I would just need to go left to right, top to bottom generating these. I want to also make sure that certain blocks only spawn connected to certain other blocks. What would be the best way to go about this?

[bleak citrus]

Sounds a lot like the “wave function collapse” algorithm

Give that a googling

[buoyant dawn]

Alright

[frank peak]

wave function collapse, or you could devise some sort of height map generator + grid builder (using 3 loops, nested to build it).

(assuming you're doing tilemaps or something for the grid builder method)

idk you could probably devise some sort of mesh builder thing too but the specifics elude me

[sand wraith]

Is there any way can have Coroutines in Zenject Dependency injection framework since service classes are not inheriting from Monobehaviour

[glacial wedge]

Hi. this will work? is this legit?

public void Method(){
  CancelInvoke(nameof(Method));
  [code executed only once, no cares about how many invoked calls]
}

[dusty wigeon]

Backtracking, WaveCollapseFunction is backtracking with constraint propagation and lowest entropy heuristic to choose the next "random" node. Start with Backtracking, then add constraints propagation, then modify your heuristic. Backtracking does everything that WaveCollapseFunction does, but it is way easier to understand at first. Their relation is comparable with A* and BreadthFirstSearch.

Also, you might find the variant of WaveCollapseFunction that generate your constraints base on a image/correctly made level. This is not necessary.

[dusty wigeon]

If it does work, this is terrible design. Use a boolean to safe guard your method.

public void Method()
{
  if(methodHasBeenExecuted)
    return;

  methodHasBeenExecuted = true;
}

[bleak citrus]

The intuition is that you pick the tile that is least flexible -- the one that has the fewest possible choices (maybe only one!) -- and pick one of the options.

You then constrain its neighbors based on the tile your chose.

I implemented the simple neighbor-based system. It works pretty well, although I did monumentally screw up the rotation calculations over and over

I want to give the image-based synthesis a try

[glacial wedge]

should I use both?

public void Method()
{
  if(methodHasBeenExecuted)
    return;

  methodHasBeenExecuted = true;
  CancelInvoke(nameof(Method));
}

[dusty wigeon]

No. Why you want to use CancelInvoke ?

[bleak citrus]

CancelInvoke only makes sense if you are also using Invoke

it does not magically stop the function from being called

[glacial wedge]

to clear a bit the stack of methods that will be called?

it's on a pooled object so I want to be sure that a call started from his previous usage doesn't interfere with the current usage

[bleak citrus]

that is absolutely not what it does

https://docs.unity3d.com/ScriptReference/MonoBehaviour.CancelInvoke.html

it cancels all outstanding Invoke calls

this would only be relevant if you tried to call the method several times via Invoke

it also wouldn't stop you from calling the method again in the future

[dusty wigeon]

What you describe is not the lowest entropy heuristic. Choosing the one with smallest domain is different.

Entropy is the measurement of Chaos. It quantify how things are certain. By example, a perfectly balance coin would have maximum entropy while an unbalanced coin would have lower entropy. To use this heuristic, you need to have non uniform distribution of your choice.

It is explained the difference in the following article:
https://www.boristhebrave.com/2020/04/13/wave-function-collapse-explained/

[bleak citrus]

Ah, yes, good point. It's not as simple as "how many choices?"

That would be correct if and only if everything had equal probabilities

Least uncertainty, not least choices.

[dusty wigeon]

In that case, entropy would be equal for everything. Making it a useless heuristic.

[bleak citrus]

well, even if you used binary "can use this tile" or "can't use this tile" values (rather than storing the probability of getting a specific tile), entropy would still vary based on how many valid choices there were

[dusty wigeon]

Pretty sure it won't. I can be wrong, it is a while since I did entropy calculation.

[bleak citrus]

uncertainty declines as you rule out tiles

I've done a whole bunch of it for a cryptography class :p

[dusty wigeon]

So I did.

I would double check what you are saying.

I am not sure.

[bleak citrus]

entropy is the negative sum of p * log(p)

for all possible events

if there is only one possible event, p=1, so 1 * log(1) = 0

[dusty wigeon]

If you have more choice of equal propability, it will give you the same result.

[bleak citrus]

if you have x equally likely events, the entropy is log2(x)

[dusty wigeon]

What is the entropy of 25% 25% 25% 25% ?

[bleak citrus]

you can imagine that as the number of bits you'd need to identify the outcome

[bleak citrus]

2 bits of entropy

[dusty wigeon]

vs
50% 50%

[bleak citrus]

1 bit of entropy

and then 100% is 0 bits of entropy, because the outcome is completely certain

and 33% 33% 33% is...1.584 or so

if a tile has 10 equally likely options, then it's more uncertain than a tile that has 2 equally likely options

[dusty wigeon]

You are right, it does increase. Must have been an other concept that was normalizing the entropy.

[bleak citrus]

There could be some other term that's constant in that case. Maybe something that measures bias.

but yeah, collapsing the lowest-entropy tile means that you're doing least "surprising" thing possible

if you just collapsed random tiles, you'd get tons of little islands that don't mesh with each other

incidentally: making a priority queue to keep track of which tile to collapse next was a HUGE performance gain

in my case, I pushed tiles into the priority queue every time their entropy went down

(and then just ignored duplicates)

[bleak citrus]

the most surprising thing possible would be to just choose every tile uniformly randomly

[dusty wigeon]

If you have uniforme propabability, then you do not really need to calculate the entropy. Which makes it simpler for most people to understand. This is why I usually suggest to look into backtracking; It is the entry point.

[bleak citrus]

ya

although uniform probability can make it hard to generate maps with variably-sized features

if i were generating an island map, for example, i'd make the land-coast and coast-water connections less likely than land-land and water-water

[dusky dust]

Hi, Ive had a issue where in unity my game works, but when i build, there are errors like embedded, wondering if anyone would be kind enough to download the project and see if he / she can fix it

Those are the erros i got in the webgl build

DMs open

[dusty wigeon]

The error System.Linq.Enumerable.First means that it tries to get the first element of something but fails to do so as there is no element correspond. If you could look and see what it is trying to get, you might resolve your issue. Probably something about setup that you did wrong.

[dusky dust]

How do i look what is it trying to get? @dusty wigeon

[dusty wigeon]

By reading the code if it is possible.

hexmone.framework.js:10... You might want to check all call of the call stack as the first seem to only be the dump

If not, you might want to contact the support.
Alternatively, you could restart from scratch and redo the correct step.

[dusky dust]

thx

[soft moon]

does anybody know how to override Delete key in TMP_InputField so that it does not do anything?

[dusty wigeon]

Why would you do that ?

[soft moon]

I am making typing game and Delete key should just do nothing

but if cannot be caught in onValidateInput

inputField.onValidateInput += ValidateInput;

[dusty wigeon]

You could modify the behaviour of the package by modifying it. If there is really no otherway

[soft moon]

what do you mean?

[dusty wigeon]

Better than that

You can just inherit.

[soft moon]

yes, that's exactly what I am trying to do

[soft moon]

using TMPro;
using UnityEngine;

public class CustomInputField : TMP_InputField
{
    protected override void LateUpdate()
    {
        if (Input.GetKeyDown(KeyCode.Delete))
        {
            print("KeyCode.Delete");

            return;
        }

        base.LateUpdate();
    }
}

it does not work though

I do not know why 🙄

[dusty wigeon]

Yeah, I was looking at the following.

But is seem it is not virtual.

[soft moon]

yep, so..

[dusty wigeon]

Just take the package and put it in your asset folder.

Modify it there.

[soft moon]

I do not get it

[dusty wigeon]

Take the content of the package, and drop it in your own project.

[soft moon]

so I should kinda copypaste the TMP_InputField code to e.g. CustomInputField?

[dusty wigeon]

https://support.unity.com/hc/en-us/articles/9113460764052-How-can-I-modify-built-in-packages-#:~:text=You can modify built-in,to another location [2].

[dusty wigeon]

Not really.

You copy that in your own project:

Then you modify KeyPressed directly or make it virtual.

[soft moon]

yeah, I cannot find TMP_InputField there

never mind, I have found it already

[soft moon]

should I open it with Notepad for example and modify it there?

[empty shell]

So I have a problem where I am suppose to have a object that is not null as far as the inspector goes, but the console still gives me a difficult time.

Basically, I finish my game, press the restart button when I finish my level. After that I finish the level again and an object that was set-up in the inspector says it's considered null.

Here is how I restart my level :

    public void RestartLevel()
    {
        SceneManager.LoadScene(SceneManager.GetActiveScene().buildIndex);
    }

[dusty wigeon]

No ? Use whatever IDE you normaly use.

[tired fog]

Hi!

[dusty wigeon]

What is the actual error ? What is the object ?

[soft moon]

I have modified this field I guess ?? I still cannot modify built-in packages. There is written:

After you move the package, you must remove it from theLibrary/PackageCache/ folder.
but.. what package should I move and remove ??

[tired fog]

I have N native arrays that I want to pass to a job, all the same length and size, I cannot do NativeArray<NativeArray<float>>. So, what would be a way to do it? the length of it might be really long, so, in case I need to flatten it, how could I do that in a job?

[empty shell]

Making a thread right now with everything

[dusty wigeon]

Take your time, you gonna find a solution. You have everything in hand. I cannot help you more.

[soft moon]

You should load it by its name if you mean this. In your code you do: "LoadScene with name index"

public void RestartLevel()
{
    SceneManager.LoadScene(SceneManager.GetActiveScene().name);
}

[soft moon]

ok, thanks anyway

[tiny pewter]

UnsafeList<UnsafeList<>>
it may work, i havent tested it

[dusty wigeon]

Can you not do NativeArray<float> ?

[tiny pewter]

but if there is any data race, you need search how to implement a lock

[tired fog]

The data is already as NativeArray<float> but I have N of those. How can I mix all NativeArray<float> into a single one? (without slowly doing it on the cpu by iterating each value)

[tiny pewter]

or float**

[dusty wigeon]

By using a single array that is N x NativeArray<float>.size

[tired fog]

yeah, but how do you copy the data from one to the other, without doing it slowly on the CPU

[dusty wigeon]

Then you can use array[i * width + j].

[tired fog]

yeah, but doing that on the CPU is slow, and width might be really big, so it would be a huge bottleneck. ideally I would do that on a job, but I can't do that

[tiny pewter]

@tired fog try float** and pass a int* array which represent the length of each column and a int for row number

[dusty wigeon]

You start with a single array...

[empty shell]

So I have a problem where I am suppose

[tired fog]

could you explain that further?

[tired fog]

?

[tiny pewter]

float** is pointer to pointer and it can represent a 2d array
or you are not familiar with pointer operations?

[tired fog]

do you mean a job for each copy? @dusty wigeon

[dusty wigeon]

I do not know a lot about job system, but when I do Compute Shader, I use a single array.

[tiny pewter]

if you really dont want to flatten it you may need it or unsafeList<unsafeList>

[tired fog]

Since the length is fixed for all arrays, I know exactly where each array starts and ends without the need of a pointer. But not sure how the float** would work

[tired fog]

Mmm, yeah, okay, well i think executing n jobs that copy data of a n array to a global array might work

[dusty wigeon]

You do not make any copy

You work with the same array all the way.

[tiny pewter]

since a array is a pointer (the array name is the address) and you want to store all the address of all 1d array in a single array so you need a array of pointer ie pointer to pointer

[tiny pewter]

so T**

[tired fog]

Ah okay, yeah no. The diferent arrays are results of diferent systems, I cannot make it from scratch as a single array

[tiny pewter]

i think you can limit the high and low index in each job

[tired fog]

and I can do that with the jobs system? I didn't know I could have an array containing the pointer of the other arrays without including the other arrays at the job

[dusty wigeon]

Ok, then I do not know. I did not have this issue myself and I did not really work with the Job System.

[scenic forge]

Your array can just contain pointers to arrays produced by individual systems, and access them that way without needing any copying.

[tired fog]

I'm not sure how to create that pointer array

[tiny pewter]

you can treat pointer as T

i dont even know if i can pass a T** to job, let me try it

[tired fog]

something like

        NativeArray<float*> arrayPointers = new NativeArray<float*>(...);

is not allowed

[tiny pewter]

!code

[thorn flint]

Posting code

📃 Large Code Blocks
Large code blocks should be posted as links to services like:
https://gdl.space/, https://paste.ofcode.org/, https://hatebin.com/
https://paste.myst.rs/, https://hastebin.com/

📃 Inline Code
Surround code with three backquotes. Not quotation marks.
To get C# formatting the first line should only contain cs or csharp.
Add a comment with a line number if there is an error message.
```cs
// Your code here
```
Do not share screenshots of code unless requested.

[scenic forge]

Why isn't it allowed?

[tired fog]

It's just isn't, it throws errors

[tiny pewter]

https://www.toptal.com/developers/hastebin

[scenic forge]

What errors?

[tiny pewter]

yes i test it you cannot have a pointer as type parameter

so i use malloc

[tired fog]

I'm not following

[tiny pewter]

sorry i didnt share my code

let me share it again

https://gdl.space/cusotududa.cpp

[scenic forge]

I have my own UnsafeArray<T> which really is just a wrapper for T* using UnsafeUtility.Malloc and UnsafeUtility.Free

Which then can be used in NativeArray just fine.

I don't see why you wouldn't be able to use NativeArray<T*>.

Have you looked at what the errors say? Do you have allow unsafe enabled?

[tiny pewter]

same error
i am so lazy to write a new unsafe container

[scenic forge]

Huh

Well yeah just wrap it in your own struct, it's only a few LoC anyways.

Ah yeah seems like you have to wrap it and then slap on https://docs.unity3d.com/ScriptReference/Unity.Collections.LowLevel.Unsafe.NativeDisableUnsafePtrRestrictionAttribute.html

[tired fog]

Mmm, okay, let's see

How can I get then the pointer of a NativeArray? XD

Or do I need to use UnsafeList

[scenic forge]

^

[tiny pewter]

nativearray is a struct, so NativeArray<T>*
if you want get the array (address) you need to use unsafeutility

[scenic forge]

Also, couldn't you already use UnsafeList inside NativeArray?

[tiny pewter]

no

you cannot put unsafelist in native list or vice verse
only unsafe list in unsafe list is supported

[acoustic umbra]

Hey can I repost a question that that was "way out of scope for #💻┃code-beginner" here?

[scenic forge]

https://docs.unity3d.com/Packages/com.unity.collections@2.1/manual/collections-overview.html

It's best practice to use Native collections over the Unsafe equivalents. However, it's sometimes necessary to use the Unsafe equivalents because Native collection types can't contain other Native collections. This is because of how Unity implements the Native safety checks. For example, if you want to get a list of lists, you can use either NativeList<UnsafeList<T>> or UnsafeList<UnsafeList<T>>, but you can't use NativeList<NativeList<T>>.

I remember the reason I wrote my own UnsafeArray was because UnsafeList didn't exist back then, the wild west.

[tired fog]

Mmm

Any way to transform quickly a nativearray into a unsafeList?

[scenic forge]

Just use unsafe list to begin with so you don't need to transform at all.

[tiny pewter]

oh it is allowed to have a native<unsafelist>...

[tired fog]

Mmm, not feasable

[scenic forge]

Why not? UnsafeList should have practically identical API surface as NativeList/NativeArray.

[tiny pewter]

to me unsafelist is easier to use since it has a Ptr property and you dont need to call https://docs.unity3d.com/ScriptReference/Unity.Collections.LowLevel.Unsafe.NativeArrayUnsafeUtility.GetUnsafePtr.html

to get the pointer

[dreamy barn]

Hey does anybody know how I can set camera borders for this endless runner and how I can fix this background issue. Seems like it always gets generated again

[tired fog]

I need to change around 163 fields

[scenic forge]

Sounds like simple find and replace.

[tired fog]

I wish

[scenic forge]

I mean, that's what you need to pay for to get performance, not sure what else to say.

[tired fog]

And i just saw that unsafe list doesn't have the reinterpret method?

[tiny pewter]

idk if it is valid for you just declare a empty unsafe list and point the ptr to the address of native array

check if the Ptr property can be set

you can set it

[tired fog]

but how do you get the address of a native array?

[scenic forge]

^

[tired fog]

Aha, perfect, got it working now! Nice!

[tired fog]

Ah, but I cannot get the unsafe ptr of a native array if the job is not fully completed?

Like

NativeArray<float> aray = new NativeArray<float>(...);
JobHandle ajob = something.Schedule(aray, dependancy = default);
UnsafeList<float> unsafeList = new UnsafeList<float>(){
  Ptr = (float*)NativeArrayUnsafeUtility.GetUnsafePtr(aray),
  .
  .
};
JobHandle secondJob = something.Schedule(unsafeList, dependency = ajob)

[tiny pewter]

i remember there is a error like you cant read any data from the native array if the job havent finished yet

[tired fog]

It normally allows it when using the correct sequence of dependancies, but nativearray unsfe bla bla bla, doesn't allow to set the dependancy to after what should it be caled

[upbeat path]

which Allocator did you use?

[tiny pewter]

try to get the address right after the nativearray is created

[tired fog]

tempjob

[tired fog]

Oka, lets see

[upbeat path]

thats why, try persistent

[tired fog]

doesn't make a difference

[tired fog]

That did the trick

[regal olive]

How can I end a coroutine without having code execution wait until the end of frame?

I have Coroutine A, which waits for the results of Coroutine B. Coroutine B is a recursive method; it calls itself onto child objects. Naturally, recursive methods will be terminated when done with their children. However, instead of waiting until the end of frame directly, I want the parent calls to Coroutine B to still finish executing, and then Coroutine A to finish before the end of frame.

Right now, after the termination of an instance of Coroutine B, it instantly forces a frame update.

How can I prevent this? I.e., terminate a single Coroutine instance without getting a frame update?

IEnumerator CoroutineB(some params) {
   if (some check) {
       for (int i = 0; i < 8; i++) {
         //some code
         yield return StartCoroutine(CoroutineB(params again));
       }
   } //<- if the condition fails, and the coroutine comes to the end of scope, and it terminates, it also forces an end of the frame. This is bad as CoroutineB could be computing elements that shouldn't be done once per frame, but many per                   frame. What should be happening is that the parent caller should continue executing, finish, go to its parent, etc until all recursive calls are terminated and finished, at which point the next frame should start.
}

The reason I'm using Coroutines for this is because I have an indexer implemented that forces a (wanted) frame update after a given number of iterations. I'm traversing an Octree, and it's the easiest to split its traversal with this indexer and a Coroutine being able to break and continue next frame.

Actually, with each new recursive call, this indexer is incremented, and there are yield return null statements when the indexer reaches a said value.

[dull raft]

@regal olive I'm not certain that I'm understanding you correctly, but I think you could just manually iterate through the IEnumerator instead of running it with StartCoroutine(), as yield in a coroutine-operated IEnumerator will always suspend execution of the coroutine for some time. That is, I don't think your one-frame delay is the result of the coroutine ending - rather it's a result of yield return StartCoroutine(CoroutineB(params again)); - it's Unity's coroutine execution loop receiving a value from yield.

I was recently infatuated with the idea of allowing a coroutine to execute flat-out and throttling it by yielding a frame only when necessary. I wrote a little test code, but haven't actually had an application for it so haven't really tested it to any great degree, but this is what I came up with... it might provide some ideas https://gdl.space/fodataqiji.cs

In short, as best I could figure out, there is no means to yield a value in a Unity-executed coroutine and have it not suspend for a bit. If you need continuous execution, then I'm pretty sure you have to manually iterate through the IEnumerator yourself. If you do that from another IEnumerator which you do execute with StartCoroutine(), then you can still use yield instructions of whatever variety, but have full control over which are actually passed out to Unity

[sly grove]

The easiest way to do this is probably to just not use recursion

There's nothing you can do with recursion that you can't do with a Stack

And without sub coroutines being started you will have full control of when to yield

[velvet rock]

I have the fallowing code that is grabbing the mesh from each instance of a component(SuperTextMesh), combining them, and making a material/texture to assign to them. I am facing a problem though that the generated mesh UV's don't match the generated texture? Anyone know how to fix this?
https://gdl.space/ezewucaray.cs

[velvet rock]

I fixed the UVs, I needed to grab the UVs off of every mesh I was combining. But now have a problem with the transform being off position.

combine[i].transform = superTextMeshes[i].transform.localToWorldMatrix;

I tried doing * worldToLocalMatrix, but to no luck. An issue might be the parenting of objects? RootObject/ComponentObject/HolderObject/TargetMeshObject

The new object is being added as a child of RootObject, but maybe the target mesh being that low is causing local to world space conversion issues?

Ah yeah if the root object is at (0,0,0) then the generted mesh is at the right spot.

Looks like the transform is being applied twice, but if I don't assign to combine.transform then no transform is being applied?

Instead of messing with the Matrix4x4 I just am moving the object to world space origin and then moving back after the new mesh is generated XD.

which ironically I think is what the * worldToLocalSpace solution was supposed to be doing basically.

[slender quartz]

using the cloudsave service, is there a way to search for all ids with the key "username"? if it helps im trying to use the friends service and add someone by their username saved in cloudsave, so others dont need to type out a huge id unity created to add someone

[regal iron]

Hey, is there any other way to get device's rotation degree ? Like the tilt control in many racing game (Asphalt, Mario Kart, etc)
I tried to use Gyro, but when i tilted the device forward, the output degree also changed. What I want is to control the MC by rotating the device. Like when you play racing game and use rotate to steer.

[regal iron]

this is what I refers to "tilt forward"

[sly grove]

Gyroscope is the way. No idea what you mean by "the output degree changed" that's kind of the point.

[regal iron]

what i need is to only get these value

so, tilting device backward (to yourself) and forward (away from yourself) wouldn't do anything on the input

[sly grove]

Vector3.SignedAngle

With vectors transformed by the gyro input

[regal iron]

this is what I have tried

                    //Quaternion referenceRotation = Quaternion.Euler(0, 0, -90);
                    Quaternion deviceRotation = DeviceRotate.Get(); //this is the Gyro input
                    Quaternion eliminationXY = Quaternion.Inverse(
                        Quaternion.FromToRotation(referenceRotation * Vector3.forward,
                                                  deviceRotation * Vector3.forward)
                    );
                    Quaternion rotationZ = eliminationXY * deviceRotation;
                   
                    float roll = Mathf.Ceil(rotationZ.eulerAngles.z);```

ah, SignedAngle? Let me try

[dawn zephyr]

Hey guys got a question of implementation, say I wanna use the same scene and only reload it upon completion and also increment an integer, based on that integer certain scripts' parameters will get set. How would you guys implement it? I'm currently thinking about a public static with a private setter, though that results in many scripts referencing it, would u guys consider it a problem and try to centralize it in one place? if so where? tyy ❤️

[dusty wigeon]

I would have a GameManager that has the responsability to reload the scene and would increment the integer.

The GameManager is a Prefab being instanced on the first scene and then being mark as DontDestroyOnLoad.

(For a small project)

[dawn zephyr]

And where would all the logic of the other classes' parameters be? currently I have it on the start method, and also where would i instantiate that gamemanager? unless you meant to make a singleton i cant think of a way for it

on the start method of each class that is dependant on that Level parameter**

[dusty wigeon]

Usually, you have an entry scene. This scene is the setup scene.

You want to instantiate your object because you want to be able to reuse the setup code everywhere.

[dawn zephyr]

oh right idk how i didnt think of it

[dusty wigeon]

And the GameManager would be a Singleton as well.

Except that you control his instantiation.

Otherwise, you might have suprise lag spike.

[dawn zephyr]

would u make it a singleton even though all i need from it is the level parameter?

i have no use of anything else in the class

[dusty wigeon]

For now, you do not have any other use.

[dawn zephyr]

i see alright

then what about every class that is dependant on the level? do u think its fine for each one of them just do their thing with it on start?

[dusty wigeon]

On Awake/Start of the level.

[dawn zephyr]

yea but say i have 6 classes depending on that level, would all of them on start do their thing with the level? would u not put it all in one script in some way?

[dusty wigeon]

If they depend on the level, I expect them to be instanciated in the level.

[dusty wigeon]

Otherwise, it depends on the what. Does it make sense for them to be in the same class ?

In thoery, you want to respect the Single Responsability Principle

Thus, having multiple object that holds the responsability of one composant.

[dawn zephyr]

so keeping as is in the start method is alright?

[dusty wigeon]

Normally, but it can be wrong depending on the situation.

[dawn zephyr]

i thought maybe ill need to use DI for it in some way

[dusty wigeon]

It is hard to know without knowing everything.

DI is not necessary.

I hate DI, most library are to complex for what you need.

[dawn zephyr]

same.. a standard in my local industry though

[dusty wigeon]

You do not mean Video Game do you ?

[dawn zephyr]

in my country's local unity industry zenject is a standard

[dusty wigeon]

I see.

A lot of people works with that. I'm not fan.

But, if it is standard, you should use it.

Uniformity is better.

[dawn zephyr]

its a fluent api with clear naming though for someone who never used it ( such as me ) these names sound like ambiguities and big confusion

wdym uniformity?

[dusty wigeon]

You should conform yourself with what other people do so you all talk the same language.

[dawn zephyr]

oh yea gotcha

[slender quartz]

would anyone know about doing this?

[dawn zephyr]

anyone familiar with unitask knows what im doing wrong here? i want to cancel the delay and start it from the beginning everytime im calling this function

[compact ingot]

You need to get a token from a new token source. If a token source is canceled it cannot be reused. Also using a continuation in a task is a little bit weird. You’d just put the continuation behavior on the next line after the await

[dawn zephyr]

so everytime i call the delay ill have to somehow set the token to that? also, where would i use continuation if u could expand? im very new to unitask and documentation.. has a big learning curve to say the least

[scenic forge]

Do you need StartFrightTimer to accept a cancellation token?

[dawn zephyr]

if i understood right, does that look right or atleast better?

[scenic forge]

Yep, was just typing that.

[dawn zephyr]

haha okay so when i fill in a cancellationtoken in these functions like delay when i cancel the token it cancels the function?

[scenic forge]

Yes, it cancels the task that's using the token.

[dawn zephyr]

wow okay

also would i need to dispose of it? i saw ppl saying to always dispose or?

[scenic forge]

Yeah

Also you should use TimeSpan rather than hardcoding _frightenedTimer * 1000.

[dawn zephyr]

ooo awesome

what does dispose do actually? is it just the IDisposable?

oh yea nvm

[elder pagoda]

guys, I remember seeing in some video theres some kind of tag that I can give a MonoBehaviour like [FormerName="ASDF"] so that Unity can automatically find the script after I refactor and rename the file, and I cant remember exactly what it was. Anyone know?

[flint sage]

PreviouslySerializedAs but iirc it's only fields and not classes

[elder pagoda]

wouldnt it be nice if there's a Refactor entry in the context menu for scripts 😄

[dusty wigeon]

You can rename files, there is no issue with that.

Reference to script are base on guid, not name.

[tender gust]

hey

how can I get the angle between 3 points?

the middle point being the center of the angle

[bleak citrus]

Vector3.Angle computes the angle between two vectors.

compute two vectors from A-B and C-B

where B is the center point and A/C are the other two points

[tender gust]

well basically what I'm trying to do is: I have a velocity direction for my character, and his rotation is independent from velocity, I want to get the angle between direction/heading and front rotation of character

[bleak citrus]

so you already hvae two vectors

the velocity vector and transform.forward

[tender gust]

something doesn't seem to work though, the code out of context may help

        var velocity = PlayerController.RigidBody.velocity.normalized;

        var charTransform = PlayerController.CharacterTransform;
        var targetDir = velocity - charTransform.TransformPoint(charTransform.position);
        float angle = Vector3.Angle(targetDir, charTransform.forward);

the angle is almost always the same in whichever direction I point to whilst moving

[bleak citrus]

this is wrong in several ways

[tender gust]

I assumed so

[bleak citrus]

you're trying to subtract a position from a velocity

this makes no sense

charTransform.TransformPoint converts a local-space position to a world-space position

the position is already world space

you also shouldn't be using positions at all

[bleak citrus]

again..

that's all you need

[tender gust]

ok so then Vector3.Angle(velocity, character.position)?

[bleak citrus]

no

I did not say to use position

I said to use transform.forward

do you not understand what that means?

[plucky sparrow]

Quick questions does anyone here having any past knowledge on working on a C# chatbot?

[tender gust]

right sorry

@bleak citrus thanks it works, just got confused for a sec

[tender gust]

you mean a discord bot? try out discord.net library, they have good docs

[plucky sparrow]

Not a discord bot

I am working on vr project where the npcs are all (in short) chatbots

so you can have organic conversations that don't repeat

[tender gust]

you do know that chatGPT has billions of dollars of investment

[plucky sparrow]

yes but all the software that work with it cost money

[tender gust]

you need a chatgpt API subscription, and it aint cheap

[plucky sparrow]

Exactly my reasoning

[tender gust]

or you can download a free low quality model

then run it on your own servers

[plucky sparrow]

I got it to understand (voice) english perfectly and also spanish but

[tender gust]

but you will need a GPU rental probably

[plush hare]

If it was this simple, customer service jobs wouldn't exist

[tender gust]

which is > 500$ a month

[tall ferry]

unless you have some massive data source, there really isnt a way to make your own

[tender gust]

there are some free models you can run on github

but they are low quality

[plucky sparrow]

I will die trying then XD

[plucky sparrow]

Chatgpt is required to be connected to wifi

I wanted to make a rpg game that is fully organic in nature

[tender gust]

you cannot expect your users to run a model locally

[plush hare]

unlikely that you're going to write a local chatGPT that people download on their computers

[tall ferry]

Especially with rpg dialogue

[plush hare]

there's a good reason it's wifi connected

and it's probably querying the supercomputers in whatever datacenters they live in

[plucky sparrow]

hmm ok interesting but I am still going to try

[tender gust]

for 1 user I estimate you need 15% of a RTX 3080 during a conversation using a big model

[tall ferry]

you can make your game feel more "organic" in other ways. How many people are really going to sit there talking to your npcs

[bleak citrus]

i would start by making a game

[plucky sparrow]

already have but thanks for the heads up

thanks for the info though best of wishes on your own projects

[tender gust]

if your players want to speak to a chatbot, they will use bing or gpt

or bard

whichever

just make a game

that skyrim mod was just a demo, it's actually shit

@plucky sparrow

[tall ferry]

One thing you should really consider if you do get this working, you will seriously need to limit whats given as a response. Even chatgpt still has its issues with restricting responses, jailbreaks come out everyday and people have gotten it to say some very bad things

I imagine u could run into trouble if you dont restrict it

[tender gust]

someone trained gpt on 4chan, the results were quite peculiar

[plucky sparrow]

did not know that skyrim had such npcs but that is interesting

[karmic carbon]

llm trained on racists is racist, how peculiar =)

[plush hare]

It's really easy to underestimate the amount of willpower some people possess for strictly wanting to screw with other people

[plucky sparrow]

yea I have already set in place a blocker for certain words already

[tall ferry]

yea that wont be enough

[plucky sparrow]

I took the first version of my bot to a convention and people were interesting none the less

[plucky sparrow]

yea people will find a way

[tall ferry]

thats the ice cube of the entire iceberg. people have had chatgpt tell them full recipes to make drugs for example

bad words are the least of the concern

[plucky sparrow]

true the bots are limited to there "field"

so they wouldn't even start on a conversation in that reguard

[plush hare]

I don't think that's right

[plucky sparrow]

fair judgement, nothing I say truly matters until you see proof but for now you're just have to take my word

[plush hare]

I mean, you have proof that it's already happened before

to make a bot, it needs to learn

if it's learning from a bunch of racist people, it's going to be a racist bot

etc

[tall ferry]

wouldnt it be better to just have a long list of predetermined responses, even from an existing AI?
At least you can mass generate it and ensure that nothing is bad

[plucky sparrow]

Well yes if you allow the bot to learn from people

[plush hare]

How do you plan on letting your bot learn?

If it's not allowed to scrape the internet?

[plucky sparrow]

yes but it becomes repetive

[plucky sparrow]

I feed it information released by open source free api that I find

and the information it gets from a player is discarded unless it pertains to the story or a quest

[tender gust]

what we need is a chatgpt loop with direct API access to a mechanical body, with full unrestricted wifi and terminal access

[plucky sparrow]

but I bet people will probally find a way to get around it

[tender gust]

then we will see skynet, finally

well, gptnet, but close enough

[karmic carbon]

I've already asked AI to generate code for me and ran it without comprehending it, so I'm doing my part to create skynet.

[plucky sparrow]

welp gl though I am going back to work

[nova summit]

In a event based architecture, do you have any suggestions on how to query existing contexts?
For ex: When a player spawns, his state machine's MoveState may need info about the current level so that it can handle it's path finding. For this let's say it needs to know about the world context(IWorldContext).
How can we get this world context which actually exists in the scene before player spawns?

I'm thinking about emitting an event like RequestWorldContextQuery which is consumed by LevelManager(in this context world) and fires an event RequestWorldContextQueryResponse with the IWorldContext. Is this a valid one to design or any better options?

To Summarise, In Event Driven Architecture, How to get events data if a subscriber enters later after an event is published?

[rapid osprey]

One way I did it in the past was to keep the event. In this case it would be the IWorldContext. If someone subscribe after the event, trigger the event directly with the value cached. Like that it doesn't matter when you subscribe and you are sure every components will get the call.

Example. Could be simpler in your case with just a null check. TL;DR use event keyword in C# to control the add and remove.

public sealed class VideoUIActivable : MonoBehaviour
{
    [SerializeField] private bool m_ActiveOnStart = true;

    [Space]
    [SerializeField] private UnityEvent m_OnActivate;
    [SerializeField] private UnityEvent m_OnDeactivate;

    public event UnityAction Activated
    {
        add
        {
            m_OnActivate.AddListener(value);
            if (m_State == State.Active)
                value?.Invoke();
        }
        remove => m_OnActivate.RemoveListener(value);
    }

    public event UnityAction Deactivated
    {
        add
        {
            m_OnDeactivate.AddListener(value);
            if (m_State == State.Deactivate)
                value?.Invoke();
        }
        remove => m_OnDeactivate.RemoveListener(value);
    }

    private State m_State = State.None;

    private void Start()
    {
        // Ignore ActiveOnStart if we have already Activated/Deactivated.
        // Which is possible if the GameObject was inactive when Activate/Deactivate was called
        if (m_State == State.None)
        {
            if (m_ActiveOnStart)
                Activate();
            else
                Deactivate();
        }
    }

    public void Activate()
    {
        if (m_State == State.Active)
            return;

        m_State = State.Active;
        enabled = true;

        m_OnActivate.Invoke();
    }

    public void Deactivate()
    {
        if (m_State == State.Deactivate)
            return;

        m_State = State.Deactivate;
        enabled = false;

        m_OnDeactivate.Invoke();
    }

    private enum State
    {
        None,
        Active,
        Deactivate
    }
}

[lime island]

what would you guys say is the most advanced question ever asked in this channel?

[dusty wigeon]

Some question are advance, others are not. However, those that are advance are pretty much advance equally.

[lime island]

is there a specific question that took days on end for this channel to answer?

[dusty wigeon]

No. There is not. And it would not make it more advance.

[lime island]

oh, what is your definition of an advanced question then?

[dusty wigeon]

I have no clean cut definition.

[lime island]

fair enough

[hard siren]

Hey guys! Not sure if this is the correct channel but was wondering if anyone would know how to implement a 3D voronoi map in unity programatically. I looked online and saw a couple libraries but I'm unfamiliar on how to import them into unity to work with.

[dusty wigeon]

Why they were not working ?

[hard siren]

I don't even know how to import the libraries to begin with, never imported external libraries into unity before.

[dusty wigeon]

What is the library ? A dll ?

Is it in C++ ?

[hard siren]

No, it's a set of .cs files. Unity does C++?

Didn't even know that.

[dusty wigeon]

It kinda do it, but it was just to know what we are working with.

If it is C#, just drop the folder in ?

[hard siren]

Does that work?

[dusty wigeon]

I mean, why wouldnt it work ?

If it is only C# files.

[nova summit]

I have a MessageBus implementation for the message passing. Do you see adding cache within it makes sense? I'm worried it may put me in a confusion later from where the data came from 😐

[glass hill]

Serializing and Deserializing some scriptableobjects with an XMLSerializer.

Serializing seems to work fine but Deseralizing gives me this error:
InvalidOperationException: To be XML serializable, types which inherit from IEnumerable must have an implementation of Add(System.Object) at all levels of their inheritance hierarchy. UnityEngine.Transform does not implement Add(System.Object).

One of the scriptables inside contains a reference to gameobject prefabs, however, I've marked it with System.NonSerializable

What's the best course of action here?

[flint sage]

You probably need to handle serialization for GameObject and other Unity classes with a custom serializer rather than the built in generic catch all solution

[visual spruce]

Hello! I'm having a problem with writing an IEnumerator that turns an object towards another object.
Here's the code

    public IEnumerator Turn(GameObject targetObject)
    {
        Vector3 target_no_y = new Vector3(
            targetObject.transform.position.x,
            character.transform.position.y,
            targetObject.transform.position.z);

        Vector3 endingAngle = Quaternion.Euler(target_no_y).eulerAngles;
        Debug.Log(endingAngle);

        while (Vector3.Distance(character.transform.rotation.eulerAngles, endingAngle) > 0.01f)
        {
            Vector3 newDirection = Vector3.RotateTowards(character.transform.forward, target_no_y - character.transform.position, RotationSpeed * Time.deltaTime, 0.0f);
            character.transform.Rotate(Quaternion.LookRotation(newDirection).eulerAngles);
            yield return null;
        }
        character.transform.Rotate(endingAngle);
    }```
The problem is that object rotates wildly and changing the RotationSpeed variable either to very small or very high numbers seems to have no effect.

[austere jewel]

I would normalize the direction vector you are passing as the second parameter to RotateTowards

Also Vector3.Distance is inappropriate to compare two euler angles

https://docs.unity3d.com/ScriptReference/Vector3.Angle.html is more appropriate

[visual spruce]

Thank you so much! ❤️

[austere jewel]

also I don't understand what you're doing with Rotate, it rotates something, why would you want to rotate by the new orientation?
Surely you want to set it, character.transform.rotation = Quaternion.LookRotation(newDirection);?

[visual spruce]

For some reason I was getting the Cannot modify the return value of 'Transform.rotation' because it is not a variable error, so I assumed you can't do the direct modification

But now it doesn't appear, weirdly

[austere jewel]

That occurs when you try to modify a struct's members directly via a property https://unity.huh.how/programming/common-errors/compiler-errors/cs1612

[visual spruce]

It's working now 😩 i can't believe it is working after 2 days of torture

[visual spruce]

Ah!... i understand now

so, the types simply need to be the same

it can't be vector.x = vector.x, it should be vector = vector

[orchid marsh]

It has to do with properties and how rotation is a struct. Pretty certain the link should have discussed about how properties are like methods and returning a value type from a method would be a copy of the variable's value and not the actual variable. Modifying the value would throw the error.

[torpid birch]

i have a problem with my unity

why she is building with errors in android

hellp

help

please

i need someone to fix my ploblem

[austere jewel]

stop spamming, and explain how this has anything to do with advanced code, or how anyone is meant to help with your complete lack of useful information

[torpid birch]

i have a error with my unity editor

[austere jewel]

Cool. Don't cross-post, and stick to the relevant channel, #📱┃mobile. Provide more context there.

[finite pond]

I don't see any fitting channel, so I hope this is appropriate here.
I'm creating a Linux IL2CPP build on my mac machine, but trying to run the build on an actual linux machine returns "exec format error".

I'm on a M2 silicon machine.

[sly grove]

What build target/ architecture/settings did you use

[finite pond]

just linux and release

[dusty wigeon]

There is this in 2020.1 Documentation, it might be obsolete.

Otherwise, you might want to read: https://docs.unity3d.com/2023.2/Documentation/Manual/linux-IL2CPPcrosscompiler.html

Every operating system (OS) has its own build systems which vary from one to another. If you build using the headers and libraries of a particular OS, the built Player might not run on other operating systems. To address this, Unity provides a sysroot to build with which works on all supported Linux platforms.

I guess you could try if it works in Mono too.

[bleak citrus]

I've been unable to produce macOS IL2CPP builds on my M1 macbook

I forget the exact error, but it involvd claiming that it was trying to sign a 32-bit executable

(the offending file was a 64-bit binary, according to file)

[finite pond]

Mac IL2PP builds work fine on my end!

[finite pond]

Good point, and i have! The resulting linux build produces the exact same error when trying to run it in Linux.

[bleak citrus]

hmm, i'll give it another go soon

[acoustic umbra]

hello, I am trying to make an interpreter for a small programming language for a game that I'm making (first time making an interpreter) and I quickly stumbled into a problem.

How do I save the value of an in-game player declared variable in a var type variable in my code?

Is that enough info or should I append more, like a code example?

[upbeat path]

apart from the boxing and unboxing overhead anything can be saved as type object

[tiny pewter]

i havent write a compiler nor interpreter
but have you defined the syntax of your own programming language?
if it is static type then you will parse the type of the variable and convert it to a c# type, else you may refer to how python do it

[scenic forge]

Take a look at https://craftinginterpreters.com (the web version is free), it goes through implementing a full interpreter in Java which you can easily translate to C#, and a bytecode interpreter in C if you need speed but you probably don't need.

[bleak citrus]

ooo

or just write a LISP

we love a LISP

[tiny pewter]

compiler and interpreter are just the same thing, both are extremely difficult to wirte.....

[scenic forge]

Not sure what your game is about, but also consider just embed a runtime for a popular language like Python instead.

[bleak citrus]

Lua is also a popular choice.

[acoustic umbra]

nah, want to challenge myself, wanna do everything starting from the bones all the way until I feel it's finished

[scenic forge]

A tree walk interpreter is actually pretty trivial to implement, but a compiler even for a simple VM is quite a lot of work.

[acoustic umbra]

yeah, considered doing that, but as said by the above, I'd like to try something hard

[acoustic umbra]

uuuu juicy

[scenic forge]

The game I'm working on has my own low level language (with good reasons to not use an existing language), and recently I've actually switched from a tree walk interpreter to a bytecode VM

[acoustic umbra]

wow so much help, thank you! :)

[scenic forge]

Crafting Interpreter goes through all the same concepts as well, and if performance is a big focus of yours, you should definitely look at its C implementation

Apparently people have not been able to create a faster implementation even with languages like Rust.

[tiny pewter]

c++ maybe better since it provides some abstract data structure in STL

[scenic forge]

Oh many people have tried

Crafting Interpreter is golden introduction to this field for good reasons.

[rough sky]

What prerequisites are there for reflection and attributes? I'm making a tool that fills in a gameObject's references if you right click a class, by reflection and attributes. But, while it can find all the parameters, it seems to not be able to pick up attributes.

Am I doing reflection wrong?

[thin mesa]

SerializeField is an attribute on fields not properties

[compact ingot]

You may also want to look into this https://docs.unity3d.com/ScriptReference/TypeCache.html

[rough sky]

ohh maybe that's it

I'm looking for properties, not fields

Yeah that was exactly it!

Thank you @thin mesa

[thin mesa]

i'd also go ahead and take a look at the link anikki sent as there's a method on that class for fields

[rough sky]

I'll do that too

For anyone else that might benefit from this:

using System.Collections;
using UnityEngine;
using UnityEditor;
using System.Linq;
using System;
using System.Reflection;
public class ChainFill
{
    [MenuItem("CONTEXT/MonoBehaviour/References/Chain References")]
    public static void ChainCreateFillInReferences(MenuCommand command)
    {
        MonoBehaviour target = (MonoBehaviour)command.context;
        Undo.RecordObject(target, "Reference Filled");
        Chain(target);
    }

    public static void Chain(MonoBehaviour target)
    {
        if (target == null) return;
        Type t = target.GetType();
        FieldInfo[] properties = t.GetFields().Where(prop => prop.IsDefined(typeof(Reference))).ToArray();

        foreach (FieldInfo p in properties)
        {
            object value = p.GetValue(target);
            if (value == null || value.ToString() == "null")
            {
                Component m = target.gameObject.GetComponent(p.FieldType);
                if (m == null)
                {
                    
                    Debug.Log("Trying");
                    m = target.gameObject.AddComponent(p.FieldType);
                    Undo.RegisterCreatedObjectUndo(m, "Reference Filled");
                    Debug.Log(m == null ? "Found" : $"Found {p.FieldType}");
                }
                
                p.SetValue(target, m);

                Chain(m as MonoBehaviour);
            }

        }
    }
}

[AttributeUsage(System.AttributeTargets.All, Inherited = false, AllowMultiple = true)]
public sealed class Reference : System.Attribute
{

}

[sand wraith]

Hi I have been going through some quaternion articles online and found out that inorder to rotate a Vector P by Quaternion q, we need to follow the following operation,
q x P x q-1
.
When i try this in unity, P is not rotating at all, where as if i just multiply q X P, am getting the required rotation. Can someone explain why this ?

[tiny pewter]

since unity engineers implement it on this way

[compact ingot]

presumably because the * operator is implemented such that it combines both operations, here is the code, you can try to decipher it 😛

public static Vector3 operator *(Quaternion rotation, Vector3 point)
{
    double x2 = rotation.x * 2f;
    double y2 = rotation.y * 2f;
    double z2 = rotation.z * 2f;
    double xxx = rotation.x * x2;
    double yyy = rotation.y * y2;
    double zzz = rotation.z * z2;
    double xy = rotation.x * y2;
    double xz = rotation.x * z2;
    double yz = rotation.y * z2;
    double wx = rotation.w * x2;
    double wy = rotation.w * y2;
    double wz = rotation.w * z2;
    Vector3 result;
    result.x = (float) ((1.0 - (yyy + zzz)) * point.x + (xy - wz) * point.y + (xz + wy) * point.z);
    result.y = (float) ((xy + wz) * point.x + (1.0 - (xxx + zzz)) * point.y + (yz - wx) * point.z);
    result.z = (float) ((xz - wy) * point.x + (yz + wx) * point.y + (1.0 - (xxx + yyy)) * point.z);
    return result;
}

[bleak citrus]

isn't it a compile error to multiply a vector with a quaternion that way

q * P * qinv shouldn't work

[compact ingot]

i suppose its more a question of why the math isn't 1:1 represented in its c# impl.

you still could do the M4x4 * someVec3 * M4x4'

[visual spruce]

Hello! I'm trying to figure out how to immediately transfer the slider's knob to a place where the user clicked on the slider. Can someone please help me?

[dull raft]

I think Slider has that behavior by default?

[visual spruce]

it does?

i never noticed, let me check

[hearty garnet]

Hello,

I want to create a UI menu that displays various items in their 3D form. Normally, with 1 items, I'd do a render texture. But I'll have 16+ items to display and creating 16 cameras seems like... The wrong thing to do.

I am experimenting with putting real 3D objects under a scroll-rect and layout-groups. It works (sort of) but because my camera is not Orthographic, I need to move the models slightly to make it look like they alighn on camera.

I also can't use Canvas Group to fade in and out these 3D Objects.

Is there another obvious way I don't know about? Is this how I should do it?

[tropic vigil]

I think the common approach would be to make those items 2D expect of the selected/hovered one or zoomed one. The 2D versions are usually premade assets or some sort of realtime screenshots.

Sometimes games create an illusion of 3D graphic by using shaders or by running animations. Animated cards in Gwent were quite impressive ones. Shaders are more limited, but would take less disc space than animations.

[hearty garnet]

I see. However, this is sort of an 'assignment' and it requires all items to be rotating on the UI. I think it's intentionally unreasonable to see how I'd tackle such a request. And frankly, I don't know! I can do it with putting real 3D objects and meticulously aligning them within layout groups and scroll-rects. However, that just seems like an ugly solution but I guess there is nothing else.

Shaders seem like a neat solution though. It's just not in my alley.

[long ivy]

personally I'd do it the ugly 16 camera-with-RT way first, it's simple and I don't think the the brittle alignment plan B you laid out is better unless you have a performance issue or other limitation from plan A. And you won't know if that's the case until you implement plan A anyways

[hearty garnet]

I guess I'll do that. even though the assignment says "we look for optimisation" - because I just realised that 3D objects won't get masked with Viewport of a Scroll Rect, so that's a lot of hoops to jump through if I do it without render textures.

Thanks for your input guys!

[long ivy]

does the assignment require all the 3d things to be in motion or have active lighting? Or just 3D appearance?

[hearty garnet]

It's basically an inventory display and it requires items to rotate around

[long ivy]

all of them at the same time?

[hearty garnet]

Yup

"All items should rotate"

Doesn't specify "all at the same time" but since it's for a mobile project, there won't be a hover event. So, better to make it all rotate at the same time to be safe.

[dusty wigeon]

You have multiple alternative:

• Make it works with 16 camera/shader. (For the shader, you would still need to do everything a camera does, I do not think you would gain much in performance)
• Record a video. (Propably the best you can do)
• Only render a single row at a time.
• Only render the picture that has been "long press".

Having 16 render at the same time does not feel right in term of UX. There is to much things happening at the same time which would hinder the user. For performance and good UI design you should propose the other solution of having 2D image preview and a zoom/focus visualization system. (You could also make the user able to move the preview himself)

Also, do not forget to profile if performance is trully a concern.

[celest mulch]

Hey i have this scene where multiple enemies (bats) are following my player (big guy in the middle). Now the problem is that i dont want the enemies to overlap each other while following. Is there an easy way so that rigidbody push each other out of there mesh?

[storm compass]

I mean you could just add a circle collider to the bat and then set it up in the collision matrix that they only collide with each other

[dusty wigeon]

Add a force on the rigidbody in relation to the proximity of other bats to make it a bit more smooth.

[celest mulch]

Ok thanks to both of you

[surreal juniper]

what general compression APIs or repository are available for use in Unity?
ideally something built-in or in a package, in any case it needs to have a permissive license
hard to believe I'll be stuck with GZipStream without having to install either a git repo (ZipSharpLib) or buy an asset

preferably something bursted & jobified

[vale spindle]

I want to make pathfinding with velocity (inertia). I will use that and normal pathfinding, which would be prevalent. if you look at the first reply here https://gamedev.stackexchange.com/questions/102214/pathfinding-with-inertia it seems like a solution. i guess you would add a field to node class that stores current velocity and then update it together with g cost, but still would like you to explain. The problem however is knowing what to do with velocity. You can arrive at the same node with different velocities and that would impact the rest of the path. say we have a linear path, and on each point an option to slow down or not. to get the optimal path we would have to make a binary tree and compare each of the branches which would probably kill my game.

[upbeat path]

GZipStream needs no external libraries

[surreal juniper]

I know but I expect it's not nearly as fast as a bursted implementation could be

[upbeat path]

true, but that's no big deal to convert the code. tbh in my tests the standard lib works fairly well

[real vault]

is migrating my project (2021.3.21f) to latest 2022 beta version safe ?
I am currently facing gradle build issues bcz of unity ads and people on stackoverflow says , migrating your project to latest version fixes it .

[upbeat path]

2022 is now LTS

what stackoverflow says is usually, how should I put this? 'Terminologically inaccurate'

[tiny pewter]

@vale spindle (2D case) suppose you have K different speeds and 8 different direction (orthogonal and diagonal) .
At each node now the speed is k and you can change the speed to [k-i,k+i] when exploring all neigbhors (the neigbours' speed should belong to [k-i,k+i]) and the radius of turn at different speed are different, ie if the speed is the highest you can only go forward, slower speed you can explore the cells in +-45 degrees, more slowly then +-90 degree and so on, to limit the search space
O(8 different directions*vertice^(branch factor=8*(2*i+1)) (one node can generate 8 children vary from [k-i,k+i] different speeds)
and the visited are mostly cost memory O(8 different direction (a node can be reached from 8 different direction, since the neigbhors generated by different speeds from different directions are different so you have to store it)* number of speed's choice *number of node))

[vale spindle]

what do you mean "you can change the speed to [k-i,k+i] when exploring all neigbhors (the neigbours' speed should belong to [k-i,k+i])" here?

[tiny pewter]

a bit hard to tell without pseudo code....

[vale spindle]

I read this a lot of times and sorry but i'm not getting it. In general are you suggesting an algorithm that would solve my problem?

[surreal juniper]

you mean whether it will kill anyone? probably not ... you'd have to define "safe" and you can only do so within the context of your project so the only one to tell you whether it is safe for you to upgrade your project to Unity 2022.2 is to give it a try (backup, source control, extensive testing, yada yada) 😉

[tiny pewter]

I think your problem cannot be solved by just making binary decisions
i try to come up with a possible design here and analysis the possible run time and memory allocation, you may refer to my design

i havent write an non grid based path finding, but in this case the "direction to neighbor" can be changed to parent node

[tepid vapor]

Would anyone mind looking over my character creation controller and tell me if I missed anything?

[bleak citrus]

that sounds very challenging to do for those without context

[storm compass]

Don't ask to ask, also test your character controller, then try fix the errors yourself and with google, then come here

[remote drift]

I can't find any info on IL Postprocessing. Is there any resource or scripting API for that?

IL - language

[bleak citrus]

intermediate language, specifically

the result of compiling C#

are you thinking of CUDA or somethin'?

[remote drift]

that's not what you think it is

[bleak citrus]

IL has nothing to do with the GPU

GPU-accelerated compilation 🤔

i'm sure someone's done it

[sage radish]

It's not officially documented by Unity. There's a bit of information about how to set up an assembly to work with ILPP here: https://forum.unity.com/threads/how-does-unity-do-codegen-and-why-cant-i-do-it-myself.853867/#post-5646937
And as far the code itself to define a processor, you can look at other libraries that use it, like Burst, Entities or Mirror Networking.

[bleak citrus]

re: the question, I don't know anything on that front, unfortunately

[remote drift]

Burst is not public though?

and entities switched over to roslyn codegen

afaik

[sage radish]

The package contains the source code. An old mirror of it is hosted here:
https://github.com/needle-mirror/com.unity.burst

Entities moved most of it to source generators, but last I checked, it still has some ILPP.

https://github.com/needle-mirror/com.unity.entities/blob/master/Unity.Entities.CodeGen/EntitiesILPostProcessor.cs

[remote drift]

🤔

[sage radish]

An important thing to keep in mind, which isn't obvious, is that Unity will run IL post processors in their own isolated .NET Core app, in parallel. This means you can't access any information from the Unity project in ILPPs, other than the assembly you're given to modify.

[remote drift]

yeah, this is hardest part

no debugging

Do you happen to know whether it would be a trivial task to simply change declaration type from class to struct?

[sage radish]

This is for changing the compiler generated IEnumerator to a struct?

[remote drift]

I am looking for a way to use IEnumerator in burst by swapping declaration of generated type from class to struct and then swapping IEnumerator return to that type, as well as MoveNext calls on it

[sage radish]

I would recommend using https://sharplab.io with IL output to see what needs to be changed.

Changing the class declaration to a struct should be the easiest part. The more difficult part will be to fix any IL instructions that are expecting a different type.

[remote drift]

yeah...

I really wish yield was easier. Can't even use source gen without rewriting compiler 😩

[plucky snow]

hello everyone!! im strugling tryng to implement Animation Curves on a 2D character Movement, i want to use them so i can change freely the way the player accelerates and decacelerates and things like that, anyone can help me?

[bleak citrus]

Don't crosspost.

[plucky snow]

sorry i think the question was more apropiate here...

[remote drift]

sir. Any idea how to import package into project? Even though I reference it, it seems like Unity can't resolve it

  "references": [
    "Unity.CompilationPipeline.Common.ILPostProcessing.ILPostProcessor"
  ]

"Unity.CompilationPipeline.Common" this wouldn't work either

🤔

[lilac lantern]

I'm posting this in advanced, but I'm not sure if it should go here. I have this class that's essentially a wrapper for a string so I can display a different inspector element for it. Since it only wraps a string, I want it to compare equal to a string if that string matches the id field.

[System.Serializable]
public sealed class TilemapID
{
    public static readonly string[] ID_LIST = new string[] { "base", "foreground", "background" };

    public string id;

    public TilemapID(string id)
    {
        this.id = id;
    }

    public override string ToString()
    {
        return id;
    }
    public override int GetHashCode()
    {
        return id.GetHashCode();
    }
    public override bool Equals(object obj)
    {
        return id.Equals(obj);
    }

    public static explicit operator string(TilemapID id) => id.id;
    public static explicit operator TilemapID(string s)
    {
        for (int i = 0; i < ID_LIST.Length; i++)
        {
            if (ID_LIST[i].Equals(s))
                return new TilemapID(ID_LIST[i]);
        }

        return null;
    }
}
```However, the above will not compare as equal when compared with a string (or another TilemapID class with the same string for id) that matches the `id` string. If I change the `Equals()` override to:
```cs
    public override bool Equals(object obj)
    {
        return id == obj.ToString();
    }
``` it works perfectly fine. My question is, why doesn't it work with the first implementation?

[remote drift]

because comparing string to non-string results in false

[lilac lantern]

I'm confused on what you mean. If the hashcode is overriden to return the string's hashcode, and the equals is overriden to use the string's compare, wouldnt it be equivalent to comparing two strings?

[remote drift]

it doesn't compare by hash

[lilac lantern]

TilemapID t = new("hi");

t.Equals("hi"); // this would be false
```so this is expected?

even if t is using the string's .Equals?

[remote drift]

ofc

[lilac lantern]

why?

[remote drift]

whenever you compare 2 objects of different type it's guaranteed false

unless Equals implementation does something custom

[lilac lantern]

how is that comparing two objects of different types?

[remote drift]

obj.GetType() == other.GetType()

[lilac lantern]

but

id is a string. the Equals method of TilemapID uses id.Equals

so shouldn't TilemapID.Equals(string) be equivalent to TilemapID.id.Equals(string)?

[remote drift]

oh, let me see again

should be true actually

[lilac lantern]

spent 4 hours on trying to figure out why (only landed on this being the issue at the end), came to the same conclusion, asked on stack overflow, and someone said that I need to change the Equals implementation. They got downvoted for being wrong because the first one "should" work, but... doesn't for some reason? I just checked ab 5 times right now. Having it the first way, false. Having it the second way, true.

[remote drift]

you can simply use debugger

to check internal implementation of comparison

of .net

[lilac lantern]

I did! (use the debugger) That's how I found out this is the issue

[remote drift]

and see why it returns false

[lilac lantern]

how do I do that? I only really know how to do breakpoints and step into/over/out of stuff, as well as look at variable values, memory addresses, etc.

[devout hare]

t.Equals("hi") would be true but if you compare to another TilemapID object then it's false (because the string id is not equal to a TilemapID object)

[remote drift]

ooh, wait

[lilac lantern]

but t.Equals("hi") was false

[remote drift]

Equals assumes other object of same type?

yeah, I guess that's why

[devout hare]

Not in the code that you've shown

https://dotnetfiddle.net/G9OCuf

[lilac lantern]

I literally just tested it many times. Only change, one works one doesn't. It's not like there's any errors or anything. It just doesn't work. At least, in unity.

Could it have something to do with how unity uses serialization to expose things in the inspector? The string it's comparing is exposed in the inspector. I didn't actually type t.Equals("hi"), I had it doing t.Equals(variable) - variable being a srting set in the inspector.

I still don't see how that would happen though.

[devout hare]

Debug.Log($"Variable: {variable}, id: {t.id}, variable == t.id: {variable == t.id}, t.Equals(variable): {t.Equals(variable)}") and see what it prints

[remote drift]

Just tested this. Returned true

    public class TestClass
    {
        private readonly string _kek;

        [RuntimeInitializeOnLoadMethod]
        private static void Test()
        {
            var kek = new TestClass("id");
            Debug.Log(kek.Equals("id"));
        }

        public TestClass(string kek)
        {
            _kek = kek;
        }

        public override bool Equals(object obj)
        {
            return _kek.Equals(obj);
        }

        protected bool Equals(TestClass other)
        {
            return _kek == other._kek;
        }
    }

[tall ferry]

can u show how u test this? because i try it in unity and t1.Equals("hi") gives true and t1.Equals(t2) gives false

also idk why i bother trying it in unity, it will run the exact same in the compiler online

[lilac lantern]

I don't know. Three people saying that it works like that, I must have messed something up while testing it. I'll just believe you guys lol.

[tall ferry]

Maybe u read the answers backwards 🤷‍♂️

the difference is just because of the type sent in the comparison

[remote drift]

Since my question drowned, I'll repeat:
How do I reference ILPostProcessor assemblies in my code?
I tried such asmdef, but Unity is unable to resolve
using Unity.CompilationPipeline.Common.ILPostProcessing;

{
  "name": "ILPP",
  "references": [
    "Unity.CompilationPipeline.Common.ILPostProcessing.ILPostProcessor",
    "Unity.CompilationPipeline.Common"
  ],
  "includePlatforms": [
    "Editor"
  ],
  "allowUnsafeCode": true,
  "overrideReferences": true,
  "precompiledReferences": [
    "Mono.Cecil.dll",
    "Mono.Cecil.Rocks.dll",
    "Mono.Cecil.Pdb.dll",
    "Unity.IL2CPP",
    "Unity.IL2CPP.Common"
  ],
  "autoReferenced": false,
  "defineConstraints": [],
  "versionDefines": [],
  "noEngineReferences": false
}

[real vault]

did it , but now getting new errors on build , its says there are 2 same main functions cant exist in ggg.c and fff.c and a couple of more errors ,
also tried to export it as a gradle project and opened it in android studio to build apk , but the same error shows there , mostly bcz i am clueless about how android studio works.and their sdk , ndk , gradle versions and so .

tried removing the ads package too , but still shows gradle build failed error on build

[remote drift]

@final steeple sir. I remember you mentioned ilpp as suggestion. Do you happen to know about ilpp in Unity? I can't find a way to get started.

[final steeple]

rare ping in this server, lol

I haven't personally used it before

I've always done the more manual IL postprocessing stuff

[remote drift]

sadge

I need ilpp to run before Unity runs burst compiler

[final steeple]

The annoying thing about ILPostProcessor is that it's not documented or actually intended for use outside of Unity stuff

[remote drift]

yep

[sage radish]

The name of your assembly must be of the pattern Unity.*.CodeGen.

[remote drift]

🤔

[sage radish]

This was mentioned in the forum post I linked to you

[regal glade]

Is it bad to create new meshes frequently for the mesh filter? I keep leaking meshes and I'm not sure what's causing it. Should I be editing the mesh instead of creating new ones?

[sage radish]

If you don't need the old mesh around, there's no reason you shouldn't edit the existing one. You have to call UnityEngine.Object.Destroy on created Meshes to release them.

[regal glade]

alright so the culprit should be my new Mesh() then. thanks i'll test.

just gonna do mesh.Clear and pass in the old one

yeah that fixed it thanks

[lament salmon]

Yeah any UnityEngine.Object that you create will need to be destroyed or you get leaks.
This includes Texture2D, Mesh, etc.

But yeah mesh.Clear is way faster than creating a new one

[regal glade]

thank you

rendertexture.Release() does not seem to reduce the texture count however?

[sage radish]

This function releases the hardware resources used by the render texture. The texture itself is not destroyed, and will be automatically created again when being used.

[regal glade]

ah

[sage radish]

So it depends on what the "texture count" is counting. GPU texture objects, or Unity texture objects.

[regal glade]

I see. So I should do Object.Destroy on it to get rid of it then

right?

seems to work

[sage radish]

Yes

[rancid swift]

I am trying to load a managed DLL for a dependency, but unity keeps giving me the following error:

Could not load image [...].dll due to Metadata verifier doesn't handle sections with SizeOfRawData < VirtualSize
Run the peverify utility against this for more information.
Any ideas how to resolve this? Turn off the validation?

[sage radish]

Never seen this before. Sounds like the .dll might be corrupted or something. Where did you get it from?

[rancid swift]

https://github.com/xBimTeam/XbimGeometry uses a managed DLL to calculate geometry for complex 3D models

I just built that DLL fresh from source, and the error still persists

(and it works just fine in a non-unity C# project...)

specifically: this is about the Xbim.Geometry.Engine64.dll shipped in the XBim.Geometry.Engine.Interop nuget package

[sage radish]

This seems to be a recent issue, which explains why I've never seen it before
https://forum.unity.com/threads/random-peverify-spam-errors.1398223/

[rancid swift]

To be fair, peverify spits out 15 errors when given that DLL.

but yeah, any way around that? We currently start an extra executable and communicate via an intermediate file, which is far less than ideal...

[sage radish]

Maybe try building with different target frameworks or settings and see if it makes a difference? Are you building to .NET Framework or .NET Standard?

[rancid swift]

I wish I knew how to find out. Ideally I'd build for dotnet framework 4.72, but I haven't worked with dotnet C++ projects yet

looks correct?

[sage radish]

I have no idea how C++/CLI works either.

[rancid swift]

https://github.com/xBimTeam/XbimGeometry/issues/109 seems like this has been a problem in some way for about 5 years now

[deep escarp]

hey can someone please help me, i'm creating a mod for a game ( hold fast nations at war) it's a multiplayer game and in the mod i added a driveable vehicle and tested it in the editor and it works perfect but when i launch a server and try to test it in game it doesn't show up ( also now in unity editor the veichle goes half under the terrian and won't let me drive it anymore

[glass hill]

Not looking for someone to analyze this error message, but rather, is there a way to re-target the System.Runtime library? I know how to import DLLs, but even after importing an updated System Runtime, Unity doesn't switch to it. I assume it has something to do with the .NET version but some of this is new to me. Any ideas?

[novel birch]

Looking for some security tips+resources; mainly regarding RCE.

Using MoonSharp to load Lua script made by community. We currently load main.lua file and run that; we looking into supporting require such that a single script can load more lua files.

Anyone know any good steps to ensure the scripts wont load any nasty stuff?

[sage radish]

This documentation should be comprehensive on this topic:
https://www.moonsharp.org/sandbox.html

[novel birch]

Thanks @sage radish ❤️

I have seen that list -- I am looking for more "general" best practices. As in, only allow a-z script names, load only from root of your mods folder. etc.

I just won't be happy if people can simple upload a random .exe to the community mod and just ..\..\ until they can somehow place it in Windows Startup programs or whatever.

[sage radish]

Do you intend to give Lua scripts access to the file system? The page I linked mentions io as a dangerous module that shouldn't be included.

[novel birch]

100% wont - way too fiddly to sandbox.

[sage radish]

Then they won't be able to do what you said, unless there is some unknown exploit within MoonSharp that allows them to bypass it.

The safe approach is to give your users only the bare minimum at first and then slowly expand capability when they ask for it, and after careful review.

[novel birch]

That's our current approach -- But we see a potential for community driven libraries and functionality to generalise stuff a little more.. So we wanna add the option for scripts to load other scripts.

Right now we just have people making monolith scripts which is rather crazy and hard to manage 🙂

[novel birch]

But yeah; I guess this was my main point that I was afraid of.. A single exploit would be enough to fuck shit up.

Do we know any other big projects that use MoonSharp?

[sage radish]

Ok, I understand. There's nothing inherently dangerous about require, as long as you limit what paths can be specified. MoonSharp might already do this, I'm not sure.

[sage radish]

None that comes to mind. But when I think of games with big modding communities, those games almost all have full, unconstrained, (unsafe) scripting/plugin support.

But understandably, that sort of thing is difficult to support in an official capacity as a developer, since you can't guarantee safety.

But if your game grows popular enough and users aren't happy with the restrictions, they will mod it themselves without restrictions. Most modded games are like this, with no official mod support.

[novel birch]

Exactly. We will go with disallow everything at first, but I simple have to figure out where those holes are so I can limit them 🙂 But sounds like I am on the right track at least.

Cheers!

[dusty wigeon]

.NET Standard 2.1 and the .NET Framework does not possess any library name Microsoft.Identity.Client. You will need to "install the package" which is kinda more complicated with Unity as you cannot change the reference directly. There is nothing to "retarget" as far as I know. However, I know little to nothing in dependency management.

If you want to change between .Net Standard 2.1 and .Net Framework, you need to change the API Compatibility Level which is in the Player Settings.

You also, maybe, want to be sure that whatever package you install is targetting the correct API/Version.

https://learn.microsoft.com/en-us/dotnet/api/?view=netstandard-2.1
https://learn.microsoft.com/en-us/dotnet/api/microsoft.identity.client?view=msal-dotnet-latest

[regal olive]

hmmm

i see

[drifting flame]

lol

[dusty wigeon]

🤷

[tawny imp]

Hello everyone, i did this groundcheck and gravity system on Unity, but i am experimenting an error. When the player touch the ground, still the velocity on y axis isnt 0. Someone knows what is going on?

And this is the Update

[dusty wigeon]

Is ther a reason why you want velocity to be 0 ?

[tawny imp]

This method

To fix the velocity on slopes

[dusty wigeon]

Are you using FixedUpdate ?

Are you using a Kinematic Rigidbody ?

[tawny imp]

i am using character controller

my character doesnt have rigidbogy

[dusty wigeon]

Make your own character controller, that would be a start.

This way you could truly control the velocity

[tawny imp]

a character controller without the component?

[dusty wigeon]

Yes. Character Controller is a prototype component. Not something that you should use in your final game.

This way, you can truly manage the state/interaction of your character.

Optimize it to your needs.

Your case, is a case where having your own controller is way easier.

[regal olive]

Hey. Is there any way to clear the Input.inputString?

[tiny pewter]

you have asked it in code-beginner

[regal olive]

ye

[rancid swift]

Unity runs on a weird custom fork of dotnet framework 4.72 on mono iirc, with some features backported.
It's impossible to get a dotnet 5 or newer runtime with Unity (they are working on it), but you can install some packages that microsoft did release for earlier framework versions. For example, we use System.Collections.Immutable which is included in the dotnet 5 runtime, but also available as a nuget package. Just need to get your package from nuget (not via IDE, it's complicated. Google it) and hope for the best.

[tired fog]

Hi

could anyone explain to me how the ddx and ddy on tex2D method affects the sampling of a texture in unity?

And how could I include those two variables to affect it similarly when doing UNITY_SAMPLE_TEX2DARRAY

[orchid lichen]

idk how to explain what's happening here i think the game view is taking the rendered texture from scene view, the script works as it should in scene view and i'd want the game view to have that box rendered correctly from its perspective and not the scene view's perspective. idk if this is the right channel for ask for help for this kinda stuff but i'd appreciate if someone could help idk cus where else to go. i hope u understood something from this shitty explanation lol

[orchid lichen]

i think this is more of a #archived-shaders question?

[tired fog]

true, will move it there

[dusty wigeon]

#archived-shaders

[umbral trail]

I would convert the full screen pass to use the post processing framework

[orchid lichen]

can u link a few articles for me to look into all this rendering stuff is very undocumented and hard to find info on other than the official documentation

[umbral trail]

https://docs.unity3d.com/Packages/com.unity.postprocessing@3.3/manual/Writing-Custom-Effects.html for built in

[orchid lichen]

i think i'll still need to get the render texture won't i?

[orchid lichen]

ill look into that thankss

[umbral trail]

that is part of what it does, though if this is for a 2d game I'm not sure if it's as appropriate

[tired fog]

there it is

[orchid lichen]

i don't think that'd work for me since i'd still need to draw gl calls on a texture, and that's the part where it breaks

here's a better vid of what's happening

it's so sickening i've been trying to figure out a solution since like 8 hours ago 😭

[dusty wigeon]

Are you using RenderTexture ?

[orchid lichen]

yeah i posted the code above

[dusty wigeon]

If you change your resolution, what happens ?

[orchid lichen]

sec

[dusty wigeon]

Just making sure that it is not a stupid error like that.

[orchid lichen]

yeah nothing

literally nothing

i can change the resolution to whatever i want it just squishes the texture as it should

[dusty wigeon]

Yeah, I dont know. I would need to do multiple test to correctly understand how it works.

Can't help sorry.

[orchid lichen]

yeah understandable np

[dusty wigeon]

Also, next time.
!code

[thorn flint]

Posting code

📃 Large Code Blocks
Large code blocks should be posted as links to services like:
https://gdl.space/, https://paste.ofcode.org/, https://hatebin.com/
https://paste.myst.rs/, https://hastebin.com/

📃 Inline Code
Surround code with three backquotes. Not quotation marks.
To get C# formatting the first line should only contain cs or csharp.
Add a comment with a line number if there is an error message.
```cs
// Your code here
```
Do not share screenshots of code unless requested.

[dusty wigeon]

Kinda hard to read

[orchid lichen]

oh

sure

ty

[glass hill]

That is where I've been getting them actually. I tried going to an early fork of the package to maybe lower the version requirement but no luck.

[rancid swift]

The 5.0.0 packages usually work, and I remember reading on some unity doc page that that's intended. No guarantees for anything later, though.

[glass hill]

Does nuget's dependencies page reflect the dependencies of the earlier versions or only of the latest? When I check earlier versions the dependencies page didn't update.

Maybe I can grab an earlier version of the security package and it just wasn't reflecting the proper dependencies

I'll just try a random one and see

[rancid swift]

to be honest, I've never understood the nuget page. it's a mess to figure things out.
you should just use the GUI in VS/Rider in a separate C# project to install what you need and then paste the package into your unity project.
there's also a nuget for unity plugin, but that one occasionally breaks for some dependencies and you'll have to do it manually anyways (https://github.com/GlitchEnzo/NuGetForUnity)

[rancid swift]

oh, and try the earliest version you can find for the best chances I suppose

[glass hill]

Sure

Yeah the video I watched originally just had me grab the package files and unzip them and grab the dll to put in the unity folder

Since Microsoft's SMTP stuff wasn't working. Gotta use MailKit.

[rancid swift]

I recently asked bing and, well, it just told me to use the REST API endpoint that the GUIs use on the nuget page to download the nupkg files. So that works. Can't remember the exact URL tho.

[patent glen]

anyone might know why my resolution on simulator becomes pixelled? This only happens in simulator, in game mode is normal

i aint sure if it is for this channel to ask, but seems like it

forget it found it

[regal lava]

scale ;)

[patent glen]

yeah took me a while :). sort of new. trying to do a project with a deadline next week :')

but no mao

it wasnt scale

it was something in android build

[regal lava]

:O

[patent glen]

🧀🧀🧀🧀🧀🧀🧀🧀🧀🧀🧀🧀

[spark raft]

How can I refresh my gameview while being in the editor

Like through an editorwindow script or something

[glass hill]

I tried a bunch of versions but it seems all of them depend on a higher version of system runtime

_>

[rancid swift]

Then the last option: make an extra executable and launch it from unity. Communicate via files or sockets. It sucks but works for emergencies like these. But you might need a custom installer for the runtime you're using

[glass hill]

Unfortunately that's beyond my expertise at this point

I'll have to do some more research

Am hmm

Seems if I update the unity version to 2022 and beyond, .NET 4.x is available as a scripting option

Maybe that will fix it

[glass hill]

Well that plus using the nuget extension did work

Thanks for turning me onto that @rancid swift

Now it's just a matter of getting the emailer to actually work

[rancid swift]

oh wait you werent on that already? I just assumed

[glass hill]

The project was using an LTS version of 2021 since I was contracted last year

I had switched Unity versions a bunch on previous projects and caused problems so we hadn't updated yet

[rancid swift]

well, things should be a lot easier with 4.x compatibility wise

[glass hill]

Mhm

Now I need to figure out why the emails aren't sending x)

We wanted to use an in-program emailer to send a survey response from the user to their teacher by entering their teacher's email (it's for an educational program)

So I got the azure certs for the oauth, set up the SMTP code, all that based on MailKit, but not working at the moment

No errors, likely because it's an async operation and I haven't set up any callbacks yet, but still not seeing any emails coming through

[glass hill]

Does anyone have any experience with MailKit?

[devout coyote]

I've used it before but that was a while ago - are you trying to send emails from your clients?

Unless you have total control of the clients, as in you have full control over which clients have your application installed it wouldn't be wise to use passwords etc

Unless you are sending it on behalf of a user, who has their own credentials

@glass hill

[glass hill]

What would you suggest? I'm not sure how comfortable users would be entering their own credentials into the program

So we had set up an email that can only send, with no inbox, to email from- but that may have been the wrong approach

[upbeat path]

you really, really really do not want users sending email using your credentials

[glass hill]

I understand. What do you recommend is the most used friendly and secure way of doing an emailer?

[upbeat path]

ask for username/password and process it properly

[glass hill]

Should I just have them copy paste the formatted string and email it themselves?

I was hoping there might be some more streamlined way of doing the feature

[tall ferry]

If the fear is that users wont plug in their credentials, there really isnt much you can do about that. I'd even be cautious plugging in my password, no matter how popular the application is.

You could provide them the string to copy, though not sure what you mean by formatted string

It likely wont copy over too well

[glass hill]

We were trying to remove as much required user input as possible to make the process smoother, and yeah there was a concern about user security. He wants to remove our access to whatever data they have as much as possible to make it feel more secure and safe for schools adopting the program.

Copying the mail to the clipboard or launching an external email program would probably be the best solution then unfortunately

[upbeat path]

I've never used MailKit but generally you would have a user register/sign in process and that would register a user with your email server, then you are sending out emails in their name rather than your own this can still cause problems but is easier to secure

[devout coyote]

or if you are just interested in collecting some data

Just send it over to a back-end

Do validation on the data that you get

and send your emails from there

care that this still gives attackers access to your system, but atleast they won't have direct access to the email credentials

so the worst they can do is send the lecturer a ton of emails

but realistically that's always going to be a problem unless you add user authentication and invite them to do the survey specifically

I worked on a questionnaire system where we'd generate a bunch of codes, and send those out to users

That code could only be redeemed once, for a single questionnaire

But that kind of requires you to know where to send the codes to beforehand

If you don't know before hand, you could always have them type in a unique onetime code, but you'd have to make sure that their use is authorized, either by them being in a physical location (such as a class room where the code is on a beamer)

@glass hill

[glass hill]

We don't have a user system set up yet so for now it's anonymous

[devout coyote]

Sometimes a user system is a bad thing - but then youll have to use codes

you familiar with kahoot?

[glass hill]

I am not yet

[tall ferry]

Just curious what is this actually for? Because you mentioned school but university/colleges typically gives out an email to every student which you have to use

[devout coyote]

its an online quizzing program - one person makes room, gets a code, everyone else fills it in on their device

and they join that room

[glass hill]

Well this is program for training a specific type of student, and the questionnaire at the end is just for the lecturers to know what responses the students gave to the virtual experiencr

[devout coyote]

There are existing tools that solve this problem

[tall ferry]

You might not need email at all for this

[devout coyote]

e.g. https://www.mentimeter.com/

or just a office forms app

[glass hill]

Yeah that's where I'm landing but I have to run it by my boss

He was adamant we cannot control whatever response system they use so that there's no concern we are collecting any data

Since unis are very secure about their students and networks

Hence why we don't have a use system or anything like that

[devout coyote]

so if htey don't trust you, then you'd have to use a third-party product that they do trust

or have it integrate with their existing products

[tall ferry]

I would stay away from email in this case, especially because itll be so vulnerable giving out credentials to your email made. And if you attempt to have the students send that formatted string to your server, which then sends the email to the lecturer, you might as well just upload the information to some server instead

[glass hill]

Yep. I'll do a little more research and discussion with my boss about it and let him know what yall suggested.

I haven't had to tackle a problem like this (mostly do games) so I appreciate the feedback and knowledge

👍

[tall ferry]

Thinking about it as well, schools would never allow a system where users type in their password. The only option would be a system where users create accounts if you need to attach some information to a person (like one person sends 2 responses in on different days/devices/locations).
If it doesnt matter who sent it, kahoot style should be fine along with verification. Kahoot did suffer from the fact users can upload the code publicly, then anyone could join putting in whatever bad names they wanted.

[glass hill]

Yep. That's why we didn't go that direction at first. A kahoot style solution might work, but I'll have to see what he wants.

[devout coyote]

People will always grief surveys though

Make the code only have x responses

Or allow teacher to close it

You ll never be able to 100% prevent

Without sacrificing the reliability of your survey

[tall ferry]

That wont help the overall issue. The most you can do is deter it by having an account system so that responses are tied to a person. If they grief it, it will be tied to their name. This is why u dont see school emails getting griefed often unless hacked

[devout coyote]

Non anonymous surveys are not good

Esp when negative feedback is issued

[glass hill]

Technically my code adds invisible tags to the email at the moment saying which student ID used the program so the lecturer can grade them

So it's not actually anonymous

[tall ferry]

How are u setting up those student IDs?

[glass hill]

Ah yes, and there's the issue there- it's entered by the student and there's no verification

[tall ferry]

What would stop me from putting someone elses ID in and doing a 2nd survey, writing bad things to get them kicked out

This system definitely needs accounts then

Which is annoying to setup but it really has to be used

[glass hill]

Nothing at the moment. Which is another issue we can't exactly address. I'm leaning toward just giving them a button to copy responses to clipboard and prompting them to email the lecturer. My boss seems to have an issue with it though.

I don't want to derail the chat anymore so if there was anything else we could make a thread but for now I just need to wait to meet with my boss haha

[tall ferry]

I too dont like the idea of a program giving me a long string to copy paste into my email tbh

[glass hill]

Agreed. But he has a lot of restrictions that we need to follow for whatever solution we do use.

[tall ferry]

Hope it goes well over with the boss

[glass hill]

Thanks 👍

[devout coyote]

People should be able to be negative without consequences

Else your survey won't reflect the true opinion

I am not going to give negative feedback about something if I know I could get in trouble

[glass hill]

Perhaps survey was wrong in this instance- its basically a log of their responses to the virtual experience (the decisions they made in dialogues) appended with a short essay about the experience.

That they write in the text box at the end of the program

[devout coyote]

Could honestly just have them fill in their own email

Send them a one time code

To proof that they have access

[glass hill]

That's definitely a consideration 👌

[wide elbow]

Not sure if this is considered advanced or not but I want to use a render texture of a camera output that is pointing at a canvas with a button. How can I make that render texture register the button click events? Do I have to somehow pass the raycast to the other camera or is there a better way?

[bleak citrus]

speaking as an academic type

loooooooooooooool

our student union site got backdoored

i reported it to the sysadmin

they fixed the defaced site

the backdoor was still there

i guess they are very big about checking boxes, so they'll be very annoying to deal with

[tall ferry]

now that u mention that.. there were multiple instances in my uni where people got hacked and messaged every single email associated with the school. Also i found some way to have the system spam me with emails, figured that could be potential lag (but ofc didnt try it)

[glass hill]

well yeah to be fair, I'm not speaking to their competencies- just on the front-loaded requirements x) I haven't worked with unis much but they have a lot of concerns and questions when adopting new technology

(even if they're not very good at enforcing them afterward)

and since this company is technically a startup they want to make sure we can get the widest early adoption customer base possible, so they're being very careful to not have any features that might require the institutions to give us information

[bleak citrus]

this reminds me of how NeosVR does its curved menu

It renders a flat canvas, then displays that on a curved mesh

Oh yes, this'll be exactly what you need

IIRC, it would raycast onto the curved menu, look up the UV coordinate of the hit point, and raycast onto the original canvas

[soft shoal]

Hi guys, anyone knows how to trace how many bytes allocated in heap from a func call by log? Can't use unity profiler cuz I wanna check it in release build.

.Net provide an API, GC.GetAllocatedBytesForCurrentThread(), but it seems dosen't work in Unity, but works in pure .Net console APP.

This is what I code, and the log in unity (API returns 0...), and the log in console

[soft shoal]

One possible reason can be, GC.GetAllocatedBytesForCurrentThread only works in .Net Framework, but in Unity, C# runtime is managed by Mono. I'm not sure

[sly grove]

Unity profiler gc alloc thing will tell you I believe

[dusty wigeon]

Profiler.BeginSample - Profiler.EndSample will track the GC.alloc.

Maybe there is other function in the Profiler API that you could use too.

https://docs.unity3d.com/ScriptReference/Profiling.Profiler.html

[tired fog]

What would be a faster way to do this?

    public static Texture2DArray GenerateTextureArray(Texture2D[] textures, int size)
    {
        Texture2DArray textureArray = new Texture2DArray(size, size, textures.Length, TextureFormat.RGBA32, true);
        for(int i = 0; i < textures.Length; i++){
            textureArray.SetPixels(textures[i].GetPixels(), i);
        }
        textureArray.filterMode = FilterMode.Bilinear;
        textureArray.Apply();
        return textureArray;    
    }

I was thinking with Graphics.CopyTexture, but I cannot manage to get the properties right on the method call

[dusty wigeon]

Currently in Unity texture arrays do not have an import pipeline for them, and must be created from code, either at runtime or in editor scripts. Using Graphics.CopyTexture is useful for fast copying of pixel data from regular 2D textures into elements of a texture array.

https://docs.unity3d.com/ScriptReference/Texture2DArray.html

[tired fog]

Yeah yeah, I tried that, but cannot get the parameters right.

I was hoping that by just doing

            Graphics.CopyTexture(textures[i],0,0, textureArray,i,0);

It would do the trick, but not really

[soft shoal]

Right, I did use profiler for development build. But I want it on release build. profiler stuff only works on development build

[soft shoal]

I want it on release build, that's why I'm trying log

[soft shoal]

Specifically, I'm checking if UniTask.Delay can generate GC or not. Based on their doc, development build will have GC (I already saw it on profiler), release build won't. So I want to confirm it on release build\

[sly grove]

I mean... I'd just trust them lol

UniTask is a pretty big thing, I'm sure they know what they're doing

[soft shoal]

emmm...my case, it's a long story...but yes, maybe this case I just trust UniTask, but I want to know if there's way to log GC info in Unity, that's what I'm focusing

[sly grove]

FWIW https://forum.unity.com/threads/profiling-a-release-build.186055/#post-5665108

[eager carbon]

hi final ik has this feature called head effector which allows you to stabilize your head (useful for full body fps where you dont want headbob)

what would be the unity animation rigging equivalent of this?

[sly grove]

Not completely sure. Perhaps Aim Constraint or Multi-Aim Constraint?

But... #🏃┃animation might be a better place to ask

[eager carbon]

i was thinking multi aim constraint on each of those elements, but it didnt work properly

[eager carbon]

eh i doubt they will know

[dusty wigeon]

I mean, doubt any body will know.

I usually just do trial and error

with Final IK

Oh, nvm

There is no equivalent for that in Animation pretty sure.

[eager carbon]

yeah i know but i was wondering if i could simulate it somehow

[sly grove]

Well there's all kinds of stuff in the animation rigging package https://docs.unity3d.com/Packages/com.unity.animation.rigging@1.3/api/index.html

[eager carbon]

i thought it would be the same as adding multi aim constraint on each of those elements seperatly but nope

[eager carbon]

yeah thanks i will need to look into it

[dawn zephyr]

anyone familiar with UniTask catches whats wrong here? after 1 jump for some reason my token is being disposed

[scenic forge]

Probably the ??= at L4.

[dawn zephyr]

how come?

[flint sage]

If you dispose it, then it's not null and a new one isn't created

[dawn zephyr]

ahhh so dispose just clears the memory and not the pointer?

[austere jewel]

Nothing sets things to null apart from actually setting them to null.

Only UnityEngine.Object has the bullshit null equality going on

[dawn zephyr]

lmao

[scenic forge]

Very based.

[dawn zephyr]

you guys are saviors tyvm 😄

wait so that dispose will never get called no? because its never null

[scenic forge]

Perhaps you are confusing ?. and ??

[dawn zephyr]

mb lol nvm

[austere jewel]

Dispose on the first line will get called if it's not null.
Therefore the assignment on the second line only ever work if the Dispose was never called.
As a CancellationTokenSource is invalid once disposed, then you have a problem, and should not be using ??=, just =

[dawn zephyr]

yea gotcha thank you very much 🙂

[bleak citrus]

I eliminated most of my head-bob by just forcing the camera to point in the desired view direction

The positional bob is way less problematic

[eager carbon]

by just using LookAt or..?

[bleak citrus]

hm, implemented this a while ago, so I had to go look at it again

I'm using Cinemachine. The camera is set to follow and look in the same direction as an "Eyes" transform

The "Eyes" transform is parented to an "Eye Holder" transform

That object has a Rotation Constraint on it that copies the rotation of the player

I adjust the strength based on how much view bob I want

[eager carbon]

so eyes is not the child of player?

[bleak citrus]

It is. Eye Holder is attached to the head

At 100% strength, the rotation constraint will make the Eye Holder point in the same direction as the player's root transform

(The thing the CharacterController is on: not on the armature)

I rotate the player around the Y axis to do yaw and adjust the local X rotation of the Eyes object to do pitch

So, if the Eye Holder is pointing exactly the same direction as the player's root, it cancels out all of the rotation caused by the player's animations

and then the Eyes can pitch up and down as usual

I have it set to like 70% strength normally, so that there's a little sway

[eager carbon]

hmm i see, i will have to try that out thanks

[minor valley]

i tried that but it always return 1

[SerializeField] private Canvas canvas;


    private void Awake()
    {
        Debug.Log(canvas.scaleFactor);
    }

[orchid lichen]

i have this short snippet running in OnRenderObject() and i'd need it to run different resolutions for scene view and game view, is there a way to detect which view it's currently rendering and then adjust the resolutions accordingly? ik this is a very googlable question but i legit can't find any answers

or alternativery a function that returns right values for both scene and game view cus screen.height returns wrong resolution for scene view and SceneView.currentDrawingSceneView.camera.pixelWidth returns wrong for game vierw

[sly grove]

Camera.current is a good place to start

[orchid lichen]

nope, ion want it to be rendered on the cam

it runs in edit mode

i need it to cover the whole editor view with pixel perfect resolution

[sly grove]

Scene view is also rendered by a camera

the scene camera

[orchid lichen]

nvm my bad it returns the scene cam too

yeah

ty

[upbeat garden]

I’ve got an absolute doozy of a problem:

I’m investigating a bug in which it seems like a value in a script changes out of nowhere, but only if I inspect the value using the VS debugger.

It’s Schrödinger’s variable. This is driving me absolutely insane

It gets weirder: while I’m debugging, a getter property on my script gets called, even though it has no call references anywhere in code.

But again - only when I debug and inspect certain variables

[upbeat garden]

Yup, I’m that stupid.

[mellow plinth]

How can I request the Burst compiler (in editor) to start jitting a certain method when asynchronous compilation is enabled?

[sly grove]

The debugger calls the getter

Which you might have recognized based on your subsequent message?

[bleak citrus]

did you have a getter that changed state

[soft shoal]

lol, this channel...noboby there

[granite viper]

I think your discord is broken

[soft shoal]

y?

[granite viper]

There is very clearly a number of messages after that one

[soft shoal]

I'm talking 1, we are in channel 2 right now. He ask question in both, so do I, but I basically never got answer from channel 1...

[austere jewel]

That they weren't answered in a completely different discord isn't relevant to this discord, and is just off-topic.

[granite viper]

oh I see

nvm then

[high lake]

I have a navmesh which is autogenerated by unity and I have extracted its vertices and indices. I want to split up the navmesh into a collection of meshes. (the nav mesh consists of multiple surfaces which are seperated, I essentially want to seperate these surfaces into individual meshes) Does anyone know how I can achieve this?

Further more I want to find the closest surface given a point in space. I was going to do this by using Physics.closestPoint and looping over each mesh post separation, but closest point doesn't work on concave meshes which a surface may be.

obviously I can iterate over each triangle but I do need this to be performant.

[granite viper]

iterating over each triangle can be plenty performant if you use broad phase filtering