archived-code-general

[rocky helm]

But bounds has got a whole bunch of other stuff that works toohttps://docs.unity3d.com/ScriptReference/Bounds.html

[lean sail]

All options would work the same, the absolute easiest is to plug the reference in inspector.
If you have a lot of objects it needs to be plugged into then you could search the parent transform, it won't be the most efficient thing if you have a deep list of children but the performance is probably negligible/expected since this is once at start

[spring creek]

Dependency Injection seems like the best method for you if you don't want to deal with serialized references

[rocky helm]

That should be the same as Bounds.extents.y without the /2

[wraith raven]

do u mean "colliderObject.position.y" here?

[normal arch]

ok ty

[normal arch]

also this @rocky helm

[lyric atlas]

yeah, the hierarchy is shallow and doing it once on start will be fine, I'd much rather that than do insepctor references even though there aren't a ton, it just streamlines my workflow a tiny bit

[normal arch]

i assume you mean only the y value

[wraith raven]

ya

[normal arch]

also i replaced colliderObject.position with other.bounds.center (other is the collider taken from the on triggere enter function)

[lean sail]

Could also do it in the reverse of what I said, instead of searching the parent for a manager just have the manager search its children to see where to plug its reference into

[lean sail]

Which ties into the dependency injection said here

[lyric atlas]

I could but that requires the parent checking all of the hierarchy to make sure it has all child script references, I think the initial way of children being responsible for themselves is better as ideally the only interaction the manager has with the children is Invoking the event they're listening to

[rocky helm]

"here" isn't very specific, but I assume you mean colliderBounds.y/2.
That means the length of the green stick on the picture, where the blue shape is the collider and the orange is the bounding box/ bounds

This is also equal to Bounds.extents.y

[lyric atlas]

this does avoid going over the same objects in the hierarchy multiple times though but it should be negligable

[normal arch]

he menas that you cant add a float and a vector2 together so where it says cs colliderObject.position it should probably be cs colliderObject.position.y

[heady iris]

I am using the Localization package to display lots of localized strings. A frequent pattern I've noticed myself writing goes something like this:

void Start() {
  something.label.StringChanged += ShowString;
}

void OnDestroy() {
  something.label.StringChanged -= ShowString;
}
 
void ShowString(string str) {
  myText.text = str;
}

Bit of a nuisance. Are there any strategies for "automating" that setup? I'd like to be able to just do this:

something.label.StringChanged += str => myText.text = str;

...but I'm pretty sure that means that, as soon as this object (and thus its text component) is destroyed, changing languages will throw errors as it tries to interact with the destroyed component.

[rocky helm]

Oh that, yeah I meant that

[normal arch]

k ty

[wraith raven]

I meant in the first and second line

float bottomBound = colliderObject.position - colliderBounds.y/2;

colliderObject.position is a Vector right?
But we need only the y value.
So,
float bottomBound = colliderObject.position.y - colliderBounds.y/2;

[heady iris]

Ideally, this would cause all of my subscriptions to be removed when the object is destroyed.

[rocky helm]

Yeah

[heady iris]

I guess this problem actually extends beyond just localization -- it's events in general

[spring creek]

#archived-code-general message

[normal arch]

k ive fixed all the errors im trying it out now

it didnt work sadly

it isnt returning correct numbers

[rocky helm]

Show me what you tried.

[normal arch]

Collider2D platform = GetComponent<Collider2D>();
            float bottomBound = platform.bounds.center.y - platform.bounds.extents.y/2; //Gets the bottom most y-level of the collider (given that the position is at the middle of the collider)
            float upperBound = other.bounds.center.y + other.bounds.extents.y/2;
            float value = (other.bounds.center.y-bottomBound)/platform.bounds.extents.y; 
            Debug.Log(value);

they aren't tabbed like that in the actual script that's just a pasting error

[rocky helm]

platform is supposed to be the player right

[normal arch]

platform is the collider

[rocky helm]

so other is the player?

[normal arch]

like the collider that we're testing how far the player is along

ye

[wispy raptor]

I wanna do a custom editor for a custom tile palette, then i would like to show the texture and then select tiles of the palette with the mouse and then modify specific attributes of selected tiles. How can i do a canvas on the window where i can interact with the mouse?

[normal arch]

the player's collider specifically

[rocky helm]

extents is already /2 so you only need to do platform.bounds.extents.y instead of platform.bounds.extents.y/2

[spring creek]

#↕️┃editor-extensions is probably gonna be more helpful

[normal arch]

ok ill try that ty

[wispy raptor]

thanks i wil lask on that channel

[rocky helm]

Also upper bound is still supposed to be the platform's upper bound not the player's upper bound.

[normal arch]

oh yeah thats my fault

i tried it and its close, but im getting some negative values

also its between 0 and 2 i think

[rocky helm]

Thats because /platform.bounds.extents.y is supposed to be /(platform.bounds.extents.y*2), as we want to divide the height by the max height, not by half of the max height, which is extents

[normal arch]

k thats the last thing

[rocky helm]

It works then?

[normal arch]

btw i think im getting negative values because im coming in from the bottom

so i think it does work properly

maybe i should just do if its less than 0 value = 0

[rocky helm]

Show me the improved code, that shouldnt be happening

[normal arch]

Collider2D platform = GetComponent<Collider2D>();
            float bottomBound = platform.bounds.center.y - platform.bounds.extents.y; //Gets the bottom most y-level of the collider (given that the position is at the middle of the collider)
            float upperBound = platform.bounds.center.y + other.bounds.extents.y;
            float value = (other.bounds.center.y-bottomBound)/(platform.bounds.extents.y * 2); 
            Debug.Log(value);

sorry

[rocky helm]

You are missing the first 2 checks there which prevent it from going negative and above 1

[normal arch]

oh yeah i removed those because i assumed it wouldnt do that since its called in an on triggerenter2D function

[rocky helm]

The player collider is probably not a dot, so at first the players position would be a little below the platform when the player first touches the platform with the top of the collider

[normal arch]

yeah im adding them, it shouldnt hurt performance too much

[rocky helm]

Performance difference at this scale should be negligible

[normal arch]

i was thinking because i think with what i eventually want to do, i'd need to add this to an update function

also i changed them to instead of setting the value to 0 and 1, i just return

[rocky helm]

Given the info I have, I'd say that you would only need to call the function whenever the player is touching the collider, but even putting this into update shouldn't leave a slightest mark on the performance

[normal arch]

k ty

thanks alot for the help

you saved me a lot of thinking

[hard viper]

I have like 12 fields in one class that should only be used in one outside class. Would it make sense to put them all into a little class inside?

[latent latch]

A struct? Sure why not.

[hard viper]

it needs to be rapidly altered, a lot

I worry that encapsulating it in a class might make it slower, since this is used for physics calculations, for what will be an extremely expensive part of my program.

[rigid island]

first two is ok , did you get it working?

[primal needle]

Hello. I am loading an Obj file in a Corutine. But, when I run the Coroutine and run yield return www.SendWebRequest();, the code never executes after the yield. The object loads fine (no errors and I can see it in the scene). But, no code runs after the yield. I can confirm this by using the debugger and seeing it just stop after the yield as well as none of my Debug logs write out. Here is the full code, if it helps: ```
public void ButtonClick()
{
""ObjSAS":"http://127.0.0.1:10000/SAS_URL_FOR_OBJ\", " +
""ImgSAS":"http://127.0.0.1:10000/SAS_URL_FOR_JPG\", " +
""ObjMatName":"Item.jpg"}";
StartCoroutine(RunCode(creds));
}
private IEnumerator RunCode(string creds)
{
Creds jsonObject = new Creds();
jsonObject = JsonUtility.FromJson<Creds>(creds);

    Debug.Log("Running code with key: " + jsonObject.ObjSAS);       
   
    OBJ_SAS = jsonObject.ObjSAS;
    IMJ_SAS = jsonObject.ImgSAS;      
    Debug.Log("Getting Obj SAS: " + OBJ_SAS);
    using (UnityWebRequest www = UnityWebRequest.Get(OBJ_SAS))
    { 
        yield return www.SendWebRequest();
        //NO CODE EXECUTION AFTER THIS
        if (www.isNetworkError || www.isHttpError)
        {
            Debug.LogError(www.error);
            Console.Write(www.error);
        }
      ....```

[rigid island]

the code is cutoff

[primal needle]

I can provide the full code, but I know there is a cap to the number of characters. Nothing runs in my code after the yield return www.SendWebRequest(); line

[rigid island]

full code is fine, here use links !code

[tawny elk]

Posting code

📃 Large Code Blocks

Use 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 format as C#, add cs to the first line:
```cs
// Your code here
```
Add a comment with a line number if there is an error message.

[primal needle]

Let me know if this works: https://gdl.space/ezerexikul.cs

[rigid island]

should put debug.log/breakpoint in that else statement

[primal needle]

I did. I just removed them for previty

[rigid island]

are you sure there are no console errors or maybe they're collapsed?

[primal needle]

yeah I'm sure

[rigid island]

I'm thinking that or your server is just not sending back the response

unless whatever you called this coroutine from or on has been disabled / destroyed

[primal needle]

ugh. I just restarted Visual Studio and Unity and now the break points are gettting called. Sorry about that

[rigid island]

😮

[primal needle]

that being said, the code still does not seem to be moving the obj into view of the camera or adjusting the pivot point of the object (the pivot points of the models are all off and I cant edit the files)

If you have any experience with that?

[rigid island]

where exactly is it moving it to view? perhaps the -center is wrongly calculated

is it moving at all? , does scaling part work

[primal needle]

The camera is fixed. and the model is a GameObject that I have centered in the camera. So, I thought that by attaching the loaded objet to it would work?

Scaling seems to work but let me change some values to check

[rigid island]

and the model is properly loaded at least I assume?

[primal needle]

Yeah, I can see the model when it loads in the Scene view

hmm. I dont think the loaded object is being added as a child to the GameObject. Let me see if that would make a difference

[rigid island]

yeah not sure, don't really see that in the code you sent 😛

[primal needle]

Yeah, I realized that too lol

trying it now

well, adding it to a parent game object works well. The problem is still, what I think is a pivot point issue, but maybe I am wrong? When the model loads I can see the Red/Green/Blue arrows are not in the center of the GameObject (or its parent). This causes a problem when I then try to rotate the objet, because it rotates around the wrong point. It also means when I try to move it, I think its moving from the wrong origin. I am not sure how to fix that (the actual Obj files I dont have access to edit)

[narrow yoke]

I cloned this project.But it pops this error and editor crashed

[rigid island]

File not found

[narrow yoke]

I found there did not have Megacity.unity file. Is it auto-generated?

[rigid island]

Wdym auto generated. Don’t think that’s the case here

[narrow yoke]

Anyone has ever downloaded this project?

[rigid island]

Yeah that’s a tough one , Unity doesn’t allow to fix the pivot in editor or /editor script

Working with object with incorrect pivots is a pain

[primal needle]

it sure is!

[lean sail]

That is a scene named Megacity. Scenes are not auto generated by default

[leaden ice]

You have to edit the mesh in a 3d modeling software or Probuilder to truly fix this.

Not sure how it's possible that you wouldn't have access for this.

The bandaid fix is to give it an empty parent

[primal needle]

I can edit the obj file in code (i.e. when I load the file I can make changes to the values in code). I just cant edit the files where they reside (they are read only and I dont have admin access). Yeah, what I did was create offsets and attached it to an empty game object

[leaden ice]

Duplicate the file

Edit the duplicate

[primal needle]

This is for a webgl export to run in a browser. the browser will dynamically provide me with the file location. so, it would be a lot of work to replicate all the files.

[main coral]

Best way to save login details? PlayerPrefs?

[normal arch]

so i have a math heavy function that only affects the y value upwards of like everything it does, however i want it to also affect the y value downwards, x value upwards, and x value downwards and i feel like there's a better way to do this than just do an if statement for all of these cases, how could i do this?

it would probably involve more maths

[steady moat]

What you just wrote make no sense for me. The first step to solve any problem, is to clearly define it.

[leaden ice]

This is a bit too vague to make any recommendations for

[normal arch]

so like i have a function that only uses the y value of some vectors, i want it to be able to use the x values too, but the only way i can think of doing tat is with if statements which doesnt seem right to do, i feel like there's another way

[steady moat]

myvector.x

done

[normal arch]

the only way?

[steady moat]

You are being way to vague...

[normal arch]

ill paste the ode and show you what i mean

[steady moat]

Whatever you do though, will need to access the value x if you want to actual use it.

[normal arch]

float bottomBound = platform.bounds.center.y - platform.bounds.extents.y; 
            //Gets the bottom most y-level of the collider (given that the position is at the middle of the collider)
            float upperBound = platform.bounds.center.y + other.bounds.extents.y;
            float value; 
            if(other.bounds.center.y < bottomBound) 
            {
                value = 0;
            } //The player is below the collider
            else if(other.bounds.center.y > upperBound) 
            {
                value = 1;
            } //The player is above the collider
            value = (other.bounds.center.y - bottomBound)/(platform.bounds.size.y); 
            //Gets how far up the player is on the collider and puts it onto a scale from 0-1 by dividing it by the maximum

see how whre basically all the vectors are it references the vector.y

i want it to sometimes be the vector.x dependin on a variable that i have

[steady moat]

Create a function...

[normal arch]

that does?

[steady moat]

exactly that

[normal arch]

thats exactly what im asking

[latent latch]

give an example of it with your desired if statements

can probably be all you need

[steady moat]

private void mMyFunction(float upper, float lower, float value)
{
    if(value < lower) 
    {
        return 0;
    }
    ...
}

[normal arch]

if(slopeDirectionUpwards == Vector2.up)
{
            float bottomBound = platform.bounds.center.y - platform.bounds.extents.y; 
            //Gets the bottom most y-level of the collider (given that the position is at the middle of the collider)
            float upperBound = platform.bounds.center.y + other.bounds.extents.y;
            float value; 
            if(other.bounds.center.y < bottomBound) 
            {
                value = 0;
            } //The player is below the collider
            else if(other.bounds.center.y > upperBound) 
            {
                value = 1;
            } //The player is above the collider
            value = (other.bounds.center.y - bottomBound)/(platform.bounds.size.y); 
            //Gets how far up the player is on the collider and puts it onto a scale from 0-1 by dividing it by the maximum
}
else if(slopeDirectionUpwards == Vector2.right)
{
            float bottomBound = platform.bounds.center.x - platform.bounds.extents.x; 
            //Gets the bottom most x-level of the collider (given that the position is at the middle of the collider)
            float upperBound = platform.bounds.center.x + other.bounds.extents.x;
            float value; 
            if(other.bounds.center.x < bottomBound) 
            {
                value = 0;
            } //The player is to the left of the collider
            else if(other.bounds.center.x > upperBound) 
            {
                value = 1;
            } //The player is to the right of the collider
            value = (other.bounds.center.x - bottomBound)/(platform.bounds.size.y); 
            //Gets how far up the player is on the collider and puts it onto a scale from 0-1 by dividing it by the maximum
}

[latent latch]

function

[normal arch]

thats basically it but twice more times because slope could be going down the y axis, or down the x axis

[steady moat]

Just do what I did.

[normal arch]

i wasjust thinking is there a different way to do this than to create a massive block of code

[steady moat]

Yes, make a function

As I stated 3 times alredy

[normal arch]

i have made a function

this is inside the function

[steady moat]

Make a new function

You know, like considelate all the repeating operation

[normal arch]

i dont get how this is supposed to help

[steady moat]

What is that

What is the differenec

Are you able to do a function called difference by example ?

If you do that with the upper bound to, maybe you could do a function call interval ^

[normal arch]

the point of that is to find out how far along the player is along a collider

that's what the variable value is

[steady moat]

I dont even need to know what it does

To be able to simplify it

[primal needle]

anyone have experience with ProfilerRecorder? I would like to display FPS for the scene but I dont see an FPS option in the API?

[normal arch]

alright im not sure if this is what you wanted me to do but i made a function that did this

public float ReturnCorrectAxisValue(Vector2 vector)
    {
        return (vector * slopeDirectionUpwards).x + (vector * slopeDirectionUpwards).y;
    }

slopeDirectionUpwards is always a Vector2 value, with y being either 1 or 0, and x being either 1, or 0 so if i times that vector by the vector i want it should return a float value of the vector axis that actually matters. thanks for the help

[spring creek]

@normal arch I'm confused by the two if and else if blocks where you set value
#archived-code-general message

You immediately assign value right after, overwriting all that code

[hard viper]

you can see FPS by clicking the button at the top right of the little game window

[steady moat]

it is not at all what I was saying

[normal arch]

yeah thats my bad

[primal needle]

I want to display it when the game is running

in WebGl

[hard viper]

what are you trying to do? are you calculating a slide along a slope?

[normal arch]

that isnt really the problem thoguh, ill fix taht later; it doesnt change much

[normal arch]

calcualte how far an object is along a slope

[spring creek]

I know, it changes nothing. It is just... bloat I guess. Seemed like you wanted shorter code. You could delete those big parts and it would be the exact same outcome

But yeah, not part of your issue. Just an aside

[hard viper]

can you just get the contact point?

[normal arch]

you mean using a cast?

[hard viper]

oh is this 2D or 3D?

[normal arch]

2d

[hard viper]

what is your input? a collider and two points?

[normal arch]

i guess its more of how far an object is along a collider because the slope is actually falt because its 2d

[hard viper]

or are you calculating from a contact point

[normal arch]

yeah

[hard viper]

so, this might not be well defined because the line segment of the slope might skewer the collider

are you sure it won’t?

[normal arch]

the slope isnt actually a slope, in reality its flat

its just graphics which make it look like a slope

[hard viper]

the angle is irrelevant here

you want to get a point on a line segment that is closest to a given collider

[normal arch]

no im getting a point inside the collider and based on the colliders size it returns a value from 0-1, 1 being fully along the collider, 0.5 being halfway along, etc

[hard viper]

i understand that, but is the slope defined by 2 points, or a contact point, or a collider?

you could have this:

or the box not touching the slope

[steady moat]

Bro, I was going to simplify your function but you probably made an error...

All of that does nothing

It is overwrite directly after

[hard viper]

you can only really define one specific spot along a slope if you can collapse the query to a single point

then the math is very simple

i think you need to work out the math better, which will dramatically reduce how many lines of code you need

because it will be more obvious what you need to do

right now, you are trying to split into cases by each corner, which is going to make your life suck

[steady moat]

Also,

value = (other.bounds.center.y - bottomBound)/(platform.bounds.size.y); 
=> implies
value = (other.bounds.center.y - (platform.bounds.center.y - platform.bounds.extents.y))/(platform.bounds.size.y); 
=> implies
value = platform.bounds.extents.y / platform.bounds.size.y;
=> implies
value = platform.bounds.extents.y / platform.bounds.extents.y * 2;
=> implies
value = 2;

Did you even test your code ? Or am I missing something here

[normal arch]

well this is clearly not the case because the actual logic of the code works fine

[steady moat]

Alright, I see there is an other platform in the mix.

[unborn sphinx]

General question:
Is there a way to create and merge Sprites/Textures to a new one and store the result in a Scriptable Object for subsequent runs? I kinda made a code that is supposed to do that, but I get a "Type missmatch" in the ScriptableObject field during the rund and a "Missing reference" when the run is stopped.

[hard viper]

not really

sprites and textures are really different

[unborn sphinx]

no not merge sprites with textures

[steady moat]

@normal arch so yeah, it is: value = (other.bounds.center.y - platform.bounds.min.y)/(platform.bounds.size.y);

[unborn sphinx]

but sprites with sprites or textures with textures. Either is fine

[hard viper]

you mean you are creating a new texture from a sprite and/or texture, and storing the result?

that is possible, but does make files, and is not super simple

you’re basically writing a new texture pixel by pixel.

[unborn sphinx]

Yes, I'd like to store it in a scriptable object, for later

[unborn sphinx]

I did

wait... let me bit bucket the code

[normal arch]

ty for this

[hard viper]

if you have a reference to the new texture, you can just assign it to an SO

creating the texture is the hard part

[unborn sphinx]

and as far as I know (but can't see it), I get a result

[normal arch]

so ic an just remove the whole upper and lower bound thing

[steady moat]

Yes, you only need 1 line of codes per direction

[hard viper]

i remember instantiating/creating sprites/textures is a bit tricky in unity

[unborn sphinx]

It actually is just a merge from existing ones, so rather "simple"

[hard viper]

are you sure you don’t want to just use masks and shit to show two together?

because that is a common and normal thing

creating a whole new object is not so simple, and not so well supported

everytime I try creating new textures, the answer is almost always “that is the wrong way to do it”

[unborn sphinx]

Thing is, I'd like to make a game about figuring out alien languages, sort of.
Where with each new run (kinda rogue lite) "words" and "letters" are combined to new structures to figure out.
Each "symbol" would represent a word and it is assembled from "parts" (predefined sprites), that get rotated and such during generation.

[hard viper]

ok. then you really don’t want to make textures like thisp

[unborn sphinx]

It worked making "Prefabs" for each Symbol, where each "part" is a child stuck to a plane

[hard viper]

no no no

[unborn sphinx]

maybe there are better ideas

[hard viper]

Prefab for a generic alien symbol

[unborn sphinx]

sure

as is the nature of a prefab 😉

[hard viper]

alien symbol has script to: input variables of what it represents, and 2) create the image

alien symbol can be created programatically, using several sprites

and it should do this

[unborn sphinx]

I already have that XD

my problem is not it's creation, but proper storage

[slate trellis]

I am trying to make a "domain" script where inside the area domain a few are registered per enemy inside it and target it. To damage the enemy I need access to the enemy's stats handler (a genera purpose script that handles all stats in the game). I do this using gameObject.TryGetComponent<StatsHandler>(out StatsHandler enemyStats) . Problem being, shows a weird warning that I haven't seen in my other references callled Name Can Be Simplified (IDE 0001). With no errors in the script, I try to run the game. When I try to cast the domain, it tells me an Object reference not set to an instance of an object Error. I cannot find the cause in my script, although, my code is very messy so I have trouble either way.

These are the relevant scripts:
PlayerAttacks (developing domain): https://pastebin.com/4xnNAQkS
StatsHandler: https://pastebin.com/rnnxvbvV
The error is at line 46. I have tried for a while and I cannot find documentation or forums relevant to my problem so I am asking here, any help appreciated of course

[unborn sphinx]

I dont want to create so many prefabs for each symbol

[hard viper]

you will not

[unborn sphinx]

And with prefabs that worked

[hard viper]

you make one single prefab. period

that one prefab takes in programatically what it does

[latent latch]

collider.gameObject.transform.parent.gameObject.TryGetComponent<StatsHandler>(out enemyStats);

[unborn sphinx]

hmmm and instead just store instructions how to assemble a sybmol at runtime

[latent latch]

I think you can scratch off gameObject after the parent. I think it's complaining you're doing a repetitive dot operation there

[unborn sphinx]

like, which part and rotation value etc

thought so too

at least I can now stop fidgeting with texture rotation matrix math 😛

[slate trellis]

thanks for the notice, it is my first time using it so I got confused. Although it still doesn't solve the main issue

[latent latch]

The issue can be many things when you've a bunch of accessing going on there

[unborn sphinx]

so called Train Whreck

[latent latch]

so ideally you want to log each variable you're accessing from the collider

and check if it's null

[slate trellis]

can you please ellaborate?

[latent latch]

check if the collider is null -> check if the gameobject is null -> check if the parent is null

[unborn sphinx]

instead of one long row, make several lines and store subsequent values, that you can log 😉

makes debugging easier too

[slate trellis]

I think I figured out the problem

[unborn sphinx]

collider.gameObject.transform.parent.gameObject.TryGetComponent<StatsHandler>(out enemyStats);

to 

GameObject go = collider.gameObject;
Transform tr = go.transform;
Transform parent = tr.parent
etc

[slate trellis]

it is getting the collider from the Floor placeholder I have, which does not have a parent of type GameObject (it's right under Scene)

[unborn sphinx]

still, long train whrecks can be "dangerous" and harder to debug step by step

[slate trellis]

if that is the case, it should be an easy fix. If not, that will need to do a lot of debugging

[latent latch]

ye, just when you get a null ref at a line and you've a string of reference dot operations going on, always log from the beginning or break it up like nico does.

[unborn sphinx]

I tell my students that they may use method chaining and such only if they are 120% sure they know (no assume to know) what the type and value is in each step. "With great power comes great responsibility"

[slate trellis]

okay, it's not the floor, as it now only shows up when I have a player in the domain. It is however on this line if (isOnEnemyLayer && enemyStats.entityType.Equals(StatsHandler.EntityType.LowEnemy)) and in the inspector it doesn't get the stats handler (although they have it)

[unborn sphinx]

What type is "LowEnemy"?

[slate trellis]

it an enum

[unborn sphinx]

also: comparing entityType with entityType.LowEnemy?

looks like a missmatch?

[slate trellis]

thats the localized reference to the enum

[unborn sphinx]

so they are the same by type

[slate trellis]

not really. The enum has some set values, but the localized reference is a referene to the enum, which means it can take any of the values the enum contains

I have 5 values in the enum. The value of the localized reference can be anything of those 5 so I need to compare them

[heady iris]

EntityType.LowEnemy, not entityType.LowEnemy

Very important distinction

StatsHandler.EntityType.LowEnemy looks at the StatsHandler class and gets the EntityType member from it. It then gets the LowEnemy member from the EntityType type.

I wouldn't call this a "localized reference". You've just made a nested type.

public class Foo {
  public enum Bar {
    Baz
  }
}

Foo.Bar is an enum type. Foo.Bar.Baz is a valid value for that enum type

public class Foo {

}

public enum Bar {
  Baz
}

Bar is an enum type, and Bar.Baz is a valid value for that enum

[unborn sphinx]

oh i missread the capital E XD

[slate trellis]

I am not too familiiar with types, but I somewhat know how to make basic uses of enums

[unborn sphinx]

what? how? WHY?

[heady iris]

https://learn.microsoft.com/en-us/dotnet/csharp/programming-guide/classes-and-structs/nested-types

[unborn sphinx]

that's the first thing to brush up on

types

[slate trellis]

oh. Lol

[heady iris]

you might already know how they work without really knowing the words

since they are, indeed, fundamental

[swift falcon]

my player is levitating and the box collider from grass and the characther are matching

[heady iris]

Nesting just affects how you refer to the type

[heady iris]

e.g.

[slate trellis]

yeah, I am just learning CS in my language so I forgot that in english they are called types

[unborn sphinx]

exactly, why one should not just intuitively only know how they work, but use them distinctively

[slate trellis]

also the fact that I am mostly learning python without my self learning journey

[unborn sphinx]

outch, good luck with the transition here then 😄

[slate trellis]

will be easy, since I already know most of the stuff they teach in english terms. I have around 2-3 years of C# experience so the difficult transition was python

[unborn sphinx]

back to your problem then

[slate trellis]

yup

[unborn sphinx]

I'm not sure why I got a lecture on Enums and such, but yeah

feels like this one sentence was what I asked about

maybe, I got lost track - long day

[slate trellis]

can completely relate. after 12 hours of work I only managed to work on the health bar and was too exhasted an unfocused to do anything else

Anywho, the problem is that tryGetComnponent<StatsHandler>(out enemyStats) doesnt return a value even when the object has the, ohhhhhhhhh.

I friggin hate my life

[unborn sphinx]

what do you get instead?

[slate trellis]

I think I figured it out

I am getting the collider from the capsule and trying to get the component from EnemyGFX. instead, the root object is the one that has all the relevant scripts

I am so mad if thats the cased

[unborn sphinx]

happens ^^

[slate trellis]

long day, as you said

[unborn sphinx]

duck typing helps
We serve as each other's ducks here 😄

[slate trellis]

I can't find a capable plushie to satisfy my needs

[unborn sphinx]

anyway, lets not get reprimanded for OT

[slate trellis]

I agree, although the channel IS also for discussing

[eager fulcrum]

Hey, so i have this method

private void SaveToJson(string fileName, object obj)
{
    string save = JsonUtility.ToJson(obj);

    if (!System.IO.Directory.Exists(_path))
    {
        System.IO.Directory.CreateDirectory(_path);
    }

    System.IO.File.WriteAllText($"{_path}/{fileName}.json", save);
}

and I do something like

SaveToJson("Players", _players)

private List<SavePlayer> _players = new();

thats the list im trying to save

SavePlayer is a simple serializable class

and for some reason

save string is just {}

why?

[slate trellis]

question. if I do a foreach loop with a debug, will it debug all the values?

// example
int[] list = {1, 5, 8, 12, 3};
foreach ( i in list)
{
  Debug.Log(i);
}``` how many of the 5 values would it print on the console?

I tested it and it should show all the items of a list or array in one frame. The problem now being I try to debug the containers of an overlap shpere but it only gives me 1 output, while it has 5 colliders inside

[leaden ice]

with a semicolon like that? No

Yes of course it will print all of them

[slate trellis]

then this is a problem

[leaden ice]

you'd have to show the code

[slate trellis]

here it is as well as a detailed breakdown of my problem

from what I can understand, it literraly breaks the foreach loop when I ask it to TryGetComponent

[hard viper]

Could someone explain simply where the need for depenetration comes from when doing physics calculations?

[leaden ice]

Debug.Log is your friend

print the size of the array etc

[slate trellis]

the code has changed a bit minimally tho, but not significantly

[leaden ice]

Debug.Log($"Array length is {collidersInDomain.Length}"); absolutely works 🤔

[slate trellis]

the array length is 5

[leaden ice]

ok then it will do 5 iterations unless you get an exception or something

or have a break; or return;

Since you said you are getting a NRE then it's quite likely breaking out due to the exception

[slate trellis]

problem being, it a) doesn't do 5 iterations and b) it breaks at that point

[leaden ice]

Object reference not set to an instance of an object
You are getting an exception

exceptions stop the execution of your code

[slate trellis]

didn't know loops worked like that. I have researched a lot about them but that never came up. thans

[leaden ice]

has nothing to do with loops

exceptions break out of ALL code

that's why it didn't come up - the loop is irrelevant

[slate trellis]

oh, that makes so much more sense now that I think about it.

[leaden ice]

btw you are completely ignoring the return value of TryGetComponent here, thus defeating the purpose of using it in the first place
collider.gameObject.transform.parent.gameObject.TryGetComponent<StatsHandler>(out enemyStats);

if you're going to ignore the return value you might as well just use GetComponent

this is why you're getting an exception too

you're trying to use enemyStats which is null since that component doesn't exist on that object

that's why TryGetComponent should always be used with an if statement

if (someObject.TryGetComponent(out MyScript theComponent)) {
  // you can access theComponent safely here
}
else {
  // the component didn't exist on the object. Trying to access `theComponent` will throw a NullReferenceException
}```

[slate trellis]

but doesn't TryGetComponent get the component only if it can be found?

Description

Gets the component of the specified type, if it exists.
``` Unity Docs

[leaden ice]

yes indeed - and it returns true when it finds it. It returns false when it can't

you can't ignore that

if it doesn't exist, your variable (the out parameter) will be set to null

[slate trellis]

and how would I avoid the NullReference exeption?

[hard viper]

if TrygetComponent {
in here you won’t get a null ref exception}

[leaden ice]

explained already

use the if statement

[slate trellis]

will try rn.

[eager fulcrum]

bumping my question

[leaden ice]

You can't serialize raw lists/arrays with JsonUtility

you need a wrapper objct

Also make sure ALL CLASSES and STRUCTS in the data you are trying to serialize are actually serializable

[eager fulcrum]

i mean, i can see the list in the inspector

so it seems that these are serializable

[leaden ice]

refer back to my first sentence

[slate trellis]

I changed it to what you told me. It just throws an Object Reference not set to an instance of an object exception

[leaden ice]

you would have to show your code and the exact line on which the error is thrown

[eager fulcrum]

wdym?

[leaden ice]

you need a wrapper object.

[Serializable]
public class SaveData {
  public List<SavePlayer> PlayerData = new();
}```

[polar marten]

do you have a screenshot of your game?

[slate trellis]

https://pastebin.com/VRz6qgKB
line 48

[leaden ice]

collider.gameObject.transform.parent.transform.parent.TryGetComponent

this will throw that exception if the object doesn't have two levels of parents

it's also unnecessarily verbose

you could just do collider.transform.parent.parent

you have a lot of extras in there for little benefit

[eager fulcrum]

can i make something like?

[Serializable]
public class SaveData {
  public List<T> PlayerData;

  public SaveData(List<T> playerData)
  {
    PlayerData = playerData;
  }
}

[leaden ice]

why not

[eager fulcrum]

aight

[slate trellis]

I have no idea how transform.parent works as not a lot of references explain how to get 2 layers of parents (first time using it). How would I go about avoiding the check if the gameobject doesn't have 2 parents? Would a layermask to check if it's in the correct layer work?

[heady iris]

transform.parent gives you the transform's parent. that's it.

there's nothing special about going up two parents

you can go up three parents with transform.parent.parent.parent if you want to

[slate trellis]

that much I figured, but it seems I needed 3 lessons to learn how it fully works

[leaden ice]

In this circumstance you should just do this:

EnemyStats es = collider.GetComponentInParent<StatsHandler>();
if (es != null) {
  // whatever
}```

that'll be the cleanest way by far

and it will handle any hierarchy depth

[slate trellis]

won't work for me as I need it to go 2 parents up

[leaden ice]

it will work

as I just said

it will handle any hierarchy depth

Also if you are trying to filter by layer you should not be doing that manually

use a layermask in the OverlapSphere call

[heady iris]

GetComponentInParent looks all the way up, and GetComponentInChildren looks at all of the transform's children

GetComponentInParent would not be very useful if it only looked at the immediate parent!

[slate trellis]

now that is just blatanlty confusing. I looked through docs and missed it and assumed it gets from the immediate parent (as Children is plural and parent is singular)

[leaden ice]

https://docs.unity3d.com/ScriptReference/Component.GetComponentInParent.html

This method checks the GameObject on which it is called first, then recurses upwards through each parent GameObject, until it finds a matching Component of the type T specified.

[slate trellis]

At least I am learning new things through this trouble

[slate trellis]

I really need to learn how to look through docs correctly

[swift falcon]

I feel you...

[rocky basalt]

What's the best way to copy only the yaw of a transform.rotation to a different gameobject/transform? Like, when you maniuplate only Rotation.Y in the inspector.. that's the type of rotation I want to copy, while ignoring everything else.

I'm in Quaternion/Euler hell and the manual's still leaving me confused

[leaden ice]

Something like this:

Vector3 direction = Vector3.ProjectOnPlane(sourceObject.transform.forward, Vector3.up);
destinationObject.transform.rotation = Quaternion.LookRotation(direction, destinationObject.transform.up);```

[rocky basalt]

Ooh those are some new ones for me. Gonna check it out

[hard viper]

if I serialize a field, set it in inspector, and then in runtime it changes via code, does it get permanently overwritten? yes right?

[leaden ice]

depends. On a scene GameObject? No.
On an asset - yes but only in the editor

[hard viper]

ty. I realize I usually just make my inspector-modified serialized fields not change.

[heady iris]

the latter isn't always the case

notably, with scriptable objects

the change will persist until the object unloads

which might not happen until you quit the editor

[leaden ice]

It's precarious

best not to modify assets in your game code as best practice

[heady iris]

indeed

I avoid touching serialized fields on my scriptable objects, just so that I don't get confused by it

[slate trellis]

I am exhausted. It FINALLY works without an error, and just like how I wanted it, and flawlessly at that. Thanks to @leaden ice @latent latch and @unborn sphinx for the help. I will make sure to give you credits upon final release

[heady iris]

I'm using non-serialized fields to store runtime data in them right now

it's just annoying that I can't see them in the inspector. custom editor time...

[polar marten]

do you have a screenshot of your game? *

sorry it glitched on the message i was replying to

[hard viper]

is there a way to check if a number is NaN?

[leaden ice]

if (float.IsNan(theNumber))

[brave pollen]

hi! does someone know why could the character not be jumping in my project? i just did this code and i think i messed up the gravity but cant figure out how to fix it, the input of the jump is working but cant point the exact error and ive been like 1 hour trying to figure out, thanks in advance

[rocky basalt]

this worked out the box. ty!

[rocky basalt]

So, since struct is a value type, can you easily serialize a specific struct configuration as a reference on a class instance?

Say MyClass has a reference to MyStruct myStruct. Struct values like myStruct.myFloat1, etc. are set at runtime. Can it be serialized? I also have Odin btw

May have just answered my own question as it pertains to Odin.. it shows in the inspector accurately, so, seems legit

[untold siren]

quick question - am I able to implement a lerp inside of an extension method?

[lean sail]

implement it meaning just use it? yes, why wouldnt u be able to

[untold siren]

assuming I'll have to define the function as an IEnumerator right?

[cold parrot]

A lerp, a coroutine and an IEnumerator are three separate concepts, depending on what you actually want to do, it may be possible or not

[untold siren]

looking to lerp the FOV of a cinemachine virtual camera, wanting to define it as an extension method

[cold parrot]

That’s not impossible yet

But the coroutine needs to run on a game object, so…

[untold siren]

ah rip, is there any viable method of still doing a smooth lerp without a coroutine / ienumerator definition?

(as an extension method)

[cold parrot]

you can pass a gameobject as argument to your extension method and run the coroutine on that

you can also lerp without any coroutine just in update on a monobehaviour

you can manually iterate an IEnumerator (coroutine is just an automation of it)

you can use DOTween

you can use async, (if you'd use unitask that would make this kind of thing quite easy really)

[untold siren]

what would be the easiest / best implementation

[cold parrot]

first impulse would be DOTween

a coroutine is fine if you want to avoid assets

manual in a monobehaviour is the most plain approach but thats ofc not really reusable

[untold siren]

DOTween has a lerp for a normal camera FOV but not for a virtualcamera unfortunately

unless I'm mistaken

[cold parrot]

you dont have to use the readymade convenience functions, you can lerp any value with dotween

its completely generic

[untold siren]

ahh ic

        public static void LerpFOV(this CinemachineVirtualCamera vcam, float endValue, float timeToTake)
        {
            var start = vcam.m_Lens.FieldOfView;
            var timeElapsed = 0f;
            Debug.LogWarning("Start FOV: " + start);
            while (timeElapsed < timeToTake)
            {
                var fovValue = Mathf.Lerp(start, endValue, timeElapsed / timeToTake);
                vcam.m_Lens.FieldOfView = fovValue;
                timeElapsed += Time.deltaTime;
            }
        }

this is what I have currently, I'm assuming if I've already defined this method the best approach would be to manually iterate an ienumerator?

[hard viper]

you should only use dotween once you have mastery over basic movement without it

[cold parrot]

the generic API http://dotween.demigiant.com/documentation.php#genericTo

[hard viper]

like, if you struggle to make a coroutine that just moves something in a straight line, or has constant acceleration or something, then you aren’t ready for dotween

[untold siren]

I know how to use coroutines lmao

it's just for extension methods

[hard viper]

ok, it wasn’t clear from convo. sry

[untold siren]

all good dw

[hard viper]

dotween is good for quick, fancy movement

[cold parrot]

        public static void LerpFOV(this CinemachineVirtualCamera vcam, float endValue, float timeToTake)
        {   
            vcam.StartCoroutine(DoIt);
            IEnumerator DoIt() {
              var start = vcam.m_Lens.FieldOfView;
              var timeElapsed = 0f;
              Debug.LogWarning("Start FOV: " + start);
              while (timeElapsed < timeToTake)
              {
                  var fovValue = Mathf.Lerp(start, endValue, timeElapsed / timeToTake);
                  vcam.m_Lens.FieldOfView = fovValue;
                  timeElapsed += Time.deltaTime;
                  yield return null;
              }
            }
        }

[untold siren]

works perfectly

you legend thank you sm

my first time actually working with extension methods which is why I was confused

o7

[cold parrot]

its not a particularly clean solution yet

you can easily start multiple such coroutines and they will fight each other

[untold siren]

for this implementation I know that I won't be lerping multiple FOVs at the same time so it should be all good - that will be an issue I know how to fix when it comes to it

[eternal kestrel]

is it possible to change a gameobject´s tag through script?

[spring creek]

yeah

https://docs.unity3d.com/ScriptReference/GameObject-tag.html

[eternal kestrel]

i need it so only the player cube can raise the score when collecting coins

i have a script that assigns a random cube to the player, so in this case i need to change the tag of the others

since theyre all duplicates they all have the same tag

[spring creek]

Just making sure, you saw the docs I linked, right?

It shows exactly how to change the tag in code

Also, wouldn't you want to ONLY change the tag of the player, instead of all the others?

I feel like that would make more sense

[eternal kestrel]

yeah i guess so lol

it worked, thank you @spring creek

[wintry stone]

I have a bucket with a handle on it what would I need to use to make the bucket swing on the handle?

https://forum.unity.com/threads/how-can-i-make-the-players-handheld-object-swing-around-with-the-players-movement.544796/
something like this if you need more specifics

[eternal kestrel]

another question, i want to add a simple win condition where if the player or enemy gets to the finish line, it will display a win or a lose panel respectively. how do i make it so that the panels wont overlap when for example the enemy gets to the finish line after the player does

[steady moat]

You could use a vertical layout group.

https://docs.unity3d.com/Packages/com.unity.ugui@1.0/manual/script-VerticalLayoutGroup.html

[still carbon]

Why does Unity temporarily allow some letters to be input into a script's public integers and floats in the Inspector? For example...set up a script with a public integer or float variable. Attach the script to a GameObject and modify the value in the Inspector. Try typing "123abc". You'll see "123ac" - the 'b' will be blocked (as it should) but the 'a' and 'c' will be allowed. The 'a' and 'c' will be removed when you press return but the question is why it allows those characters to be entered at all, and why there's a difference between 'b' (which is blocked) and 'a' and 'c' (which aren't.)

[proper trellis]

#⚛️┃physics message
im just gonna link my preexisting message but basically i need help with slopes

[steady moat]

Probably because of the implementation of https://docs.unity3d.com/Manual/EditingValueProperties.html (You can also type Infinity)

[elder flax]

ive got an int[,] representing my tiles in my map and I need to get all outlines in the form of lists of points

image to help explain

note i need seperate outlines because my map might have like concave spaces and stuff

whats the best algorithm/approach?

[fervent furnace]

the red dots?

[elder flax]

both

the coloured dots are just like the list sof points I basically want

[fervent furnace]

floodfill the white tiles

[elder flax]

and then like find all of them with neighbours?

[fervent furnace]

i am wrong, you should flood fill the empty tile, if the neighbors of empty tile is white tile then it is boundary

you may get duplicate points so you need to check if the point is added to list

[elder flax]

the white tiles are empty tiles

[fervent furnace]

then flood fill the black tile....

i think the black one means empty at first....

[elder flax]

oh i didnt think of that for some reason

[hoary monolith]

Hello, I have a method the OnSceneLoaded function of scenemanagement like this:

public void OnSceneLoaded(Scene scene, LoadSceneMode mode)
{
    this.dataPersistenceObjects = FindAllDataPersistenceObjects();
    Debug.Log(dataPersistenceObjects.Count);
    Debug.Log(scene.name);
}

and this is the method

private List<IDataPersistence> FindAllDataPersistenceObjects()
{
    IEnumerable<IDataPersistence> dataPersistenceObjects = FindObjectsOfType<MonoBehaviour>(true).OfType<IDataPersistence>();
    return new List<IDataPersistence>(dataPersistenceObjects);
}

in the debug.log it shows that I do have 1 dataPersistenceObject but when my saveGame method is accessing it, it says that dataPersistenceObject is null.

public void SaveGame()
{
    if (playerData == null)
    {
        Debug.Log("No Data was found. Need to start a new Game.");
        return;
    }

    if (this.dataPersistenceObjects != null)
    {
        foreach (IDataPersistence dataPersistenceObj in this.dataPersistenceObjects)
        {
            if (dataPersistenceObj != null)
            {
                dataPersistenceObj.SaveData(playerData);
            }
            else
            {
                Debug.LogWarning("Null dataPersistenceObj found in dataPersistenceObjects.");
            }
        }
    }
    else
    {
        Debug.LogWarning("dataPersistenceObjects is null.");
    }

    if (this.cloudDataHandler != null)
    {
        this.cloudDataHandler.SaveData(playerData);
    }
    else
    {
        Debug.LogWarning("cloudDataHandler is null.");
    }
    Debug.Log("Successful Saving");
}

[lean sail]

When is SaveGame called in relation to your data persistence objects being found?
This might be simpler instead of your data persistence objects just add itself to the save system on start, then remove themselves on destroy

[hoary monolith]

SaveGame is called on everytime a player upgrades/unlocks/equips a skill and OnApplicationQuit()

[lean sail]

Then you'll also need to ensure that the list gets updated, since objects can be destroyed or created after OnSceneLoaded

Try doing it in reverse, instead of searching for the objects to save, have them register themselves in the save system

[hoary monolith]

Okay, thanks for the help. I'll try doing that

[lean sail]

It just avoids the need to call .Find more often, but now you'll need a way to get a reference to the save system. The easiest way is singleton, since you probably dont need multiple instances of a save system

[hoary monolith]

Oh this is already a singleton. that makes it easier then

[dawn canopy]

any idea why this condition is still getting checked even though I'm on Windows?

 private void FixedUpdate()
    {
#if UNITY_ANDROID
        if (aimDone) return;
#endif
      ...

[lethal rivet]

Anybody know why the red one is going in circles? https://i.gyazo.com/467991ee6ba7770012c6398fefcab76e.gif

//Walk code
Vector3 direction = target.position - transform.position;
float yRotation = Mathf.Atan2(direction.x, direction.z) * Mathf.Rad2Deg;
transform.eulerAngles = new Vector3(0, Mathf.Lerp(transform.eulerAngles.y, yRotation, turnSpeed), 0);

Vector3 moveVector = (transform.forward * walk).normalized * speed;
rb.velocity = new Vector3(moveVector.x, rb.velocity.y, moveVector.z);

The target is the gray one.

[prime sinew]

is your current unity platform set to Android

[dawn canopy]

Oh, yes. Is there a work around for what I want to do? maybe I should use DEVELOPMENT_BUILD as well

[prime sinew]

you havent said what you want to do

https://docs.unity3d.com/Manual/PlatformDependentCompilation.html
check out the link

[dawn canopy]

Thanks

[last raven]

looked into TMP tag parser, oh god

why is it so terrible

did they have some intern code that

<br is a valid tag btw

and so is </i="why is this so stupid" why?="kill me">

but not this </alpha> despite being documented

I can't find a single consistent thing about it, every single tag is parsed differently

<size=" aasd 10 asdf" > this is valid

but this is not <size=" aasd 10 asdf" >

the code is impressively bad

[last raven]

even case sensitivity is not consistent

<zWsP> is a valid tag, so is <I angle=50> but not <Size=110%>

[last raven]

and some also ignore <noparse>

holy shit it's so bad

[loud wharf]

    internal class Program
    {
        static void Main(string[] args)
        {
            var foo = new Foo();
            (foo as IFoo).some_init = 1; //Not allowed

        }
    }

    internal interface IFoo
    {
        public int some_init { get; init; }
    }

    internal class Foo : IFoo
    {
        int IFoo.some_init { get; init; }
    }

Found out an interface property can use an init. But i don't see anyway to be able to use it when it's implimented explicitly.

[swift falcon]

Anime girl, that is interesting.

[loud wharf]

Wa?

[swift falcon]

Init is can be only used in constructor, anime girl.

var anime_girl = new Foo
{
    some_init = 1984
};

foo.some_init = 1926; //Not allowed, as its value can only be set once in the constructor

If you want to do that, declare a set property in class.

internal class Foo : IFoo
{
    int IFoo.some_init { get; init; set; }
}

[mellow sigil]

That's exactly the same as what they have + an extra syntax error

[loud wharf]

Not what i'm trying.

[loud wharf]

First example isn't allowed because it thinks it doesn't contain a definition. Which makes sense because this is an explicit implementation.

[latent latch]

What's the idea around init anyway? Instead of using constructors you just make a mini constructor for each property that is also readonly?

[loud wharf]

Essentially yes.

I find it handy imo.

Avoids the MyType() : base().

[latent latch]

Probably clean up some of my constructor parameters ;)

[loud wharf]

Yep, that's pretty much why it's there.

There doesn't seem to be a way.

And i can't find anything about this either.

[quartz folio]

Why explicitly declare the member using the interface? Why not just public int some_init { get; init; }

Also, what does this have to do with Unity

[loud wharf]

I got curious.

[loud wharf]

Nothing actually when we only have a vaccum example like this, now that i think about it, does the current version C# uses even have the init?

[raven orbit]

current version is 9

[quartz folio]

If nothing, then this is the wrong discord

[loud wharf]

Considering we talked about Flash and AS3 at some point or at least didn't stop us, you're randomly strict.

[quartz folio]

I cannot remember the context, but if it wasn't stemming from a Unity conversation then I would have told you to move on

[jaunty plume]

I have a problem: when i reach the point where the light goes between 359 and 1 (i mean when you go under 0, it goes instantly to 360 and vice versa), the camera goes crazy...
I have tried many methods, like adding 360 (it automatically goes to 0)
My apologies for bad code

{
    public bool ActivateScript = true, TouchCamera = true, IsGrounded, LateCameraMovement;
    public Vector3 Jumpforce, MoveVector, Velocity, MouseLookEuler, ZeroVector=Vector3.zero;
    public float MultiplierY = 1, MultiplierX = 1, CameraDelay;
    private void Start() {
        
    }

    private void Update() {
        MoveVector = new Vector3(Input.GetAxis("Horizontal"), 0, Input.GetAxis("Vertical"));
        //Velocity = GetComponent<Rigidbody>().velocity;
        MouseLookEuler += new Vector3(Input.GetAxis("Mouse Y")*MultiplierY, Input.GetAxis("Mouse X")*MultiplierX, 0);
    }
    private void FixedUpdate() {
        if(ActivateScript){
            if(IsGrounded){
                if(Input.GetButtonDown("Jump")){
                    Velocity = Jumpforce;
                }
            }
            Velocity += MoveVector;
            transform.position += Velocity;
            IsGrounded = false;
            GetComponent<CharacterController>().Move(new Vector3(MoveVector.x, 0, MoveVector.z));
            transform.Find("MouseLook").eulerAngles = MouseLookEuler;
            if(TouchCamera)
        {
            if(LateCameraMovement)
            {
                Camera.main.transform.eulerAngles = new Vector3(Vector3.SlerpUnclamped(Camera.main.transform.eulerAngles, MouseLookEuler, CameraDelay*Time.deltaTime).x, Vector3.SlerpUnclamped(Camera.main.transform.eulerAngles, MouseLookEuler, CameraDelay*Time.deltaTime).y, 0);
            }
        }
        }
    }
}

Edit: removed +360 parts

couldnt find any alternative method

[leaden ice]

First person camera control tutorials are all over the place and they all solve this problem.

The basic gist is stop relying on transform.eulerAngles

[jaunty plume]

i remember your name, long time no see

[heady iris]

You should set the camera's euler angles, rather than trying to change them

So just store the pitch and yaw you want, and adjust those values based on user input

[jaunty plume]

what do you mean by this?

[heady iris]

float yaw;
float pitch;

void Update() {
  yaw += inputStuff;
  pitch += moreInputStuff;

  cameraTransform.eulerAngles = new Vector3(pitch, 0, yaw);
}

e.g.

I forget if that alone will work right. I tend to explicitly compute two rotations for the yaw and pitch

var turnRot = Quaternion.AngleAxis(yaw, Vector3.up);
var lookRot = turnRot * Quaternion.AngleAxis(pitch, Vector3.right);

like this!

[jaunty plume]

i think i understood something, let me do some changes

[heady iris]

This computes the Z-axis rotation, then combines that with the X-axis rotation

[jaunty plume]

also the script is based on lantern rotation, not by mouse

if you meant input as mouse or something

[heady iris]

i don't know what a lantern rotation is

[heady iris]

this just shows the strategy you should be using

[jaunty plume]

it first rotate the lantern by mouse and the camera is smoothly rotating by lantern's position

[heady iris]

if you want to make the camera smoothly follow something's rotation, use Quaterion.RotateTowards and Quaternion.Slerp

[jaunty plume]

ok, let me see what can i can change

[heady iris]

float distance = Quaternion.AngleBetween(transform.rotation, target.rotation);
float t = Mathf.InverseLerp(30, 90, distance);
float rate = Mathf.Lerp(45, 90, t);
transform.rotation = Quaterion.RotateTowards(transform.rotation, target.rotation, Time.deltaTime * rate);

This will rotate towards a target's rotation at a rate that varies between 45 and 90 degrees per second

depending on how far off you are

[jaunty plume]

i changed based on this except for the rotatetowards

Camera.main.transform.rotation = Quaternion.SlerpUnclamped(Camera.main.transform.rotation, Quaternion.Euler(MouseLookEuler), CameraDelay*Time.deltaTime);

IT WORKS!

thanks so much

[heady iris]

there you go (:

Quaternion methods will work best for rotations

I see that you were trying to slerp the underlying euler angles

That causes discontinuities whenever the euler angles wrap around

There are some methods that can handle that

e.g. Mathf.SmoothDampAngle will properly wrap a single angle

[jaunty plume]

this is how it was supposed to work, it works perfectly

[low scaffold]

Think I'm just having a slow moment here, could someone explain why button != 1 would always be true in this case?

[mellow sigil]

x != 0 || x != 1 is always true

you probably meant &&

[low scaffold]

bless

[jaunty plume]

&& would make it ever true

[mellow sigil]

No, || would

[jaunty plume]

|| => if button is either 1 or 0, its true

&& => if button is not 1 or 0 is true

[mellow sigil]

|| => if button has any value, it's true

&& => if button has any value other than 0 or 1, it's true

[jaunty plume]

actually even no value inserted its true i think

[mellow sigil]

int always has a value

[jaunty plume]

i forgot that false false go true

[low scaffold]

it's !=
If button is not 0, and is not 1
Then invert that value so it's true when the value is 1 or 0

[jaunty plume]

yeah, i kinda remembered too late

somehow i made it bouncy

[delicate zinc]

i'mm currently working on an 2.5d fps Videogame, does anyone know of a good video guide i can use to setup stuff like the HUD animations?

the characters themselves are sprites so i'm a bit unsure how to do the animations for the FPS attacks

[steady moat]

There is multiple things that qualify as 2.5D. I guess you are talking about isometric point of view ? In that case, you want 8 direction sprite animation. There is multiple tutorial online on how to do it.

https://www.youtube.com/watch?v=NQN3rYGqqP8

[delicate zinc]

not quite, i should've explained mmyself better

[heady iris]

I don't get what an "FPS attack" would be

[delicate zinc]

yeah fair enough, i was talking mostly the animations that the player ddoes when firing guns n stuff

like in Doom for example, the ones that seem to be part of the HUD

[heady iris]

Oh, so you're just talking about sprite animations.

Nothing weird there.

[delicate zinc]

i suspect theyre indeed just sprite animations

[steady moat]

In DOOM, I'm pretty sure it is a simple overlay.

[delicate zinc]

i'm just relatively confused on how i'd go implementing them

[heady iris]

you'd probably use an Image component to show the image in a UI

so you would animate the sprite field of the image

[delicate zinc]

ahh

[heady iris]

you could also do it entirely through code, of course

(which is what doom did)

[delicate zinc]

aye, i am aware

i was thinking mostly using an AnimatorController for this

[steady moat]

Also

https://www.youtube.com/watch?v=UEeqT2G1TSg

[heady iris]

get the decino video!

[delicate zinc]

heh, i know that one

decino's, that is

i love his content

[heady iris]

https://www.youtube.com/watch?v=zKnaoA6BdJk

[delicate zinc]

had no idea codemonkey actually did this, i'll take a quick look at how he handled that bit

ahah

so yeah he does just use an Animator component with an Image component

that's gonna be easy

thanks you two

[hard viper]

let’s say I have a collider in another collider (but there could maybe be other colliders nearby). any suggestions on how to do depenetration?

Example: I need to tell red square to get out of blue triangle. But at that point in calculation, idk that red square would then be bumping into green circle.

Any general recommendations or resources on how to resolve overlaps?

[delicate flax]

is this a custom physics implementation, you do you want physx to solve this?

[hard viper]

custom physics implementation

in 2D

[delicate flax]

ah, aight, there are many videos, papers and blogs about this. give me a min and I'll see if I can dig up some 'easy' starting points

[hard viper]

ty. i just don’t know how to do this is a non-stupid way

assume i can use basic Physics2D methods to do the basics of casting, detecting overlaps, and getting “distance” between two colliders

[delicate flax]

well to start here is the source code for the built-in 2d physics in unity:
https://github.com/erincatto/box2d

but it can be quite daunting to read, I found this video explains the general algo in a 'simple' manner:
https://www.youtube.com/watch?v=9IULfQH7E90

Sebastian Lague's video on fluid sims also goes over some general concepts in a good way. (also just a really good watch imo)
https://www.youtube.com/watch?v=rSKMYc1CQHE

[hard viper]

oh shit i thought box2D was totally closed

thanks

[hard viper]

i watched, and i think I’m missing the critical thing I need

specifically, how do I resolve overlaps in a non-stupud way

pezza talked about the issue, but I don’t think he actually explained how to solve it

[normal arch]

so yesterday i was figuring out how far along an object is on a collider, but i only want to detect this is if the player came in through the left/right of a collider (or top/bottom depending what i set it to), how could i detect that?

i'd understand if you don't get what i mean

[heady iris]

you could record the direction from the collider's center to the player when the trigger starts

[normal arch]

with what?

a cast?

[heady iris]

player position - collider position

[normal arch]

so i've got this code

if(other.bounds.center - platform.bounds.center != slopeDirectionUpwards ||
                 other.bounds.center - platform.bounds.center != -slopeDirectionUpwards)
``` is this what you mean?

[heady iris]

well, you probably don't want to compare vectors for equality

I am thinking of something like this

public Vector3 direction;
public float angleTolerance;

void OnTriggerEnter(Collider other) {
  Vector3 colliderDirection = other.transform.position - transform.position;
  Vector3 expectedDirection = transform.TransformDirection(direction);
  float angle = Vector3.Angle(colliderDirection, expectedDirection);

  if (angle <= angleTolerance) { ... }
}

direction is a local-space direction that you expect the player to come from

We transform it to world space, then check the angle betwen it and the actual vector to the player

[normal arch]

ok

ty

btw for the expectedDirection can i just set that to what i want

[heady iris]

Right. You would pick the (local space) direction you want

So making it [1,0,0] would make the component check if the player is coming from its right

This would be a smart place to add a gizmo

to visually indicate the direction

[normal arch]

k, also btw im in 2d if that makes anything easier

[normal arch]

do you know if there's any math i could do based on the trigger colliders width/ height to turn into the angleTolerance variable|?

[heady iris]

are you trying to detect when the player hits a specific side of a box-shaped collider?

[normal arch]

yeah

[heady iris]

If you get a collision, not a trigger, you could check the normal vector for the collider you hit (on the player side, at least)

for a trigger, you could convert the player's position into local space

then compare the X and Y values of the local position to the collider's width and height, I guess

[normal arch]

sorry i took so long but would it be something like

float angleTolerance = Vector2.Angle(platform.bounds.min, platform.bounds.center)

[heady iris]

no, this doesn't make sense

These aren't directions.

you're finding the angle between two vectors that start at the origin and point to those two positions

I think you could compute it with Mathf.Atan2

[normal arch]

it takes the angle between the bottom corner of the collider and the center of the collider

[heady iris]

You would find the angle from the horizontal axis to the corner

so suppose the collider has a width of 4 and a height of 2

The upper corner would be at [2,1] in local space

Actually, you can still use Vector.Angle

You would want two vectors in local space.

suppose I want to allow the player to approach from the left

I want the angle between Vector3.left and a vector pointing from the center of the object to the bottom left corner

If the player's angle is less than this, they'e coming from the left

I'm not sure what logic I'd use to automatically do this for any direction

tbh I'd just add an empty child and position it in the scene view

That'd let you visibly see the cutoff angle

[hard viper]

If I just use .Cast by Vector2.zero, is that just... the same as checking for Overlap? Is there something I should know?

[normal arch]

so i would do the angle between Vector3.left (well probably vector2 but whatever) and -platform.bounds.extents (this would be a vector pointing from the center to the bottom left corner)

[heady iris]

They're in different spaces

So convert -platform.bounds.extents to local space or Vector3.left to world space

This feels convoluted...

[leaden ice]

Cast with zero doesn't make much sense

[heady iris]

I wonder if there's a better way to handle this.

[normal arch]

ok

[heady iris]

I guess it's weird because we're trying to derive the angle from the size of the box collider

[hard viper]

I'm just wondering if I can avoid making a specific Overlap method, by just using my Cast method with zero input

but yes, zero cast makes no sense

anyway, I am looking to make a simple depenetration algorithm. OverlapColliders might give me what I need, but I'm not sure if what I'm doing is dumb or not.

[delicate flax]

if and understand the problem, and you excuse my poor paint skills.
given a box collider, it has 4 side, aka 4 normals, n_1 -> n_4.

when you collide with it you can get the normal you collided with (hit.normal iirc)
in this example, we've defined the 'up' direction of the collider to be n_4, we can now check the dot product of the collision with the 'up' vector, to get the side we collided with.

does that make any sense?

[normal arch]

kinda

how would i convert that to code

[gray mural]

public static IEnumerator FocusCamera(Vector3 targetPosition, float targetSize, float duration)
{
    Vector3 startPos = _mainCamera.transform.position;
    float startSize = _mainCamera.orthographicSize, elapsedTime = 0f;

    while (elapsedTime < duration)
    {
        float t = elapsedTime / duration;

        _mainCamera.transform.position = new Vector3(Mathf.Lerp(startPos.x, targetPosition.x, t), Mathf.Lerp(startPos.y, targetPosition.y, t), _mainCamera.transform.position.z);
        _mainCamera.orthographicSize = Mathf.Lerp(startSize, targetSize, t);
            
        elapsedTime += Time.deltaTime;

        yield return null;
    }
}

This method focuses camera. It moves camera to the targetPosition and size in specific duration.
So the problem is that it doesn't fully change it. When desired position is far enough, its size and position may differ on e.g. 5f and (6f, 10f) respectively.
Why?

[hexed pecan]

I think you stop the while loop one step too early

Like the t value never reaches 1.0

[gray mural]

so <= ?

[hexed pecan]

Try that

[gray mural]

tried, still the same

[hexed pecan]

Or, just set the values after the while loop

Or, calculate the t after adding to elapsedTime, not before it

[normal arch]

also where is the hit.normal variable gotten from

[delicate flax]

Well this is just pure pseudo from memory, so take it with a grain of salt. and also, this will only work for box colliders

private void onCollisionEnter(Collision collision) {
    var normalThatWeHit = collision.normal;  //get the side/normal of the thing we hit
    var upDirection = collision.transform.forward; // get the 'up' direction of the thing we hit

    var dot = Vector2.Dot(upDirection, normalThatWeHit);

    //dot = 0 means we hit the top
    //dot = 0.5 means we hit left
    //dot = -1 means we hit buttom
    //dot = -0.5 means we hit right
}

[hexed pecan]

collision.transform.forward should be .up, no?

[delicate flax]

in 2d, that might be the case yeah 😅

[normal arch]

well it could be anything; in this case it would be a variable that i have in the script called slopeDirectionUpwards

[normal arch]

btw this is in an OnTriggerEnter2D function not an OnCollisonEnter so how would i get the normal?

[normal arch]

also wouldn't it be
1 if top,
0 if right or left
and -1 if bottom?

[gray mural]

That's the issue. So I changed < to <= and calculated elapsedTime before t. Now it works perfectly, I guess. Thank you 😄

[normal arch]

that's what the docs told me at least

[delicate flax]

btw. using Debug.DrawLine is a life saver in these situations since you can visualize the vectors that way.
and as I said pure pseudo from memory so sure, the values / API may be wrong, but I'm fairly sure the general idea is correct

[normal arch]

yeah that's fine, it's just getting the normal that i need to do and then it's done

[normal arch]

this has helped a lot btw thanks

[delicate flax]

since it's a Trigger and not a Collision, getting the normal can be quite tricky.
my first naive idea is to do a raycast, but maybe someone else here has a more elegant solution:

private void onTriggerEnter(Collision collision) {
    var normalThatWeHit = vector2.zero;

    // prob need to add a mask or something to make sure the raycast is only between, 'this' and 'other'
    // maybe even use Raycast all?
    if(physics2d.Raycast(this.transform, collider.transfrom, out var hit)){
      normalThatWeHit = hit.normal;
    }
   
    var upDirection = collision.transform.forward; // get the 'up' direction of the thing we hit
    // same code
}

[delicate flax]

now that I think about it, I'm not sure raycasts can hit triggers by default...

[normal arch]

but it would be hitting the player right?

which isnt a trigger

[delicate flax]

that's why I called it naïve. you could also do something like

var hits = RaycastAll(.....
foreach hit
if hit.transform == collider.transfrom

and parse out the correct hit, I'm not 100% on-board with this solution, but I think it will work at least. so if anyone want's to chime in with a solution, feel free

well the 'start' point/vector of a raycast is arbitrary. so you can set it to what ever you want

[edgy kernel]

my origin right now is transform.position

[hexed pecan]

https://docs.unity3d.com/ScriptReference/Physics-queriesHitTriggers.html

There's a bool for it

Remember to set it back to false after

[heady iris]

I always just pass the explicit extra argument

[hexed pecan]

Theres also this https://docs.unity3d.com/ScriptReference/QueryTriggerInteraction.html

Which is probably cleaner

[delicate flax]

then set it to what you need it to be? I'm not sure I understand the problem correctly 😅

[edgy kernel]

i got it

thanks tho

[normal arch]

Vector2 normal;
RaycastHit2D[]  hits = Physics2D.RaycastAll(platform.bounds.center, slopeDirectionUpwards);
            foreach(RaycastHit2D hit in hits)
            {
                if(hit.transform == other.transform)
                {
                    normal = hit.normal;
                }
                else
                {
                    normal = Vector2.zero;
                }
            }

why am i getting an unassigned local variable error even though all code paths set the variable

[somber nacelle]

not all code paths assign to it though. what happens if there are no elements in the hits array

[normal arch]

oh ty

[somber nacelle]

also your logic is flawed. what happens if the first element in the array is the same as other.transform but the second element is not?

also why are you comparing to another transform like that in the first place?

[normal arch]

this #archived-code-general message

[heady iris]

What would be a simple way to detect if an object is in the scene when I'm entering play mode?

I have a scene for my settings menu. I want to add a camera and event system if I start the scene directly

But I don't want to create those objects if the scene gets loaded while the game is running.

[hexed pecan]

Singletons?

[heady iris]

I guess I could just check if a camera exists

Would be a good enough proxy

[somber nacelle]

could just use a Find call from an object that is tagged EditorOnly so it isn't loaded in a build

[hexed pecan]

With the usual destroy-self-if-instance-already-present logic

[heady iris]

that'd get rid of any (extremely minor) overhead, yeah

sounds reasonable

i haven't set a tag on a game object in ages

[delicate flax]

simple way? GameObject.FindObjectOfType, or GameObject.Find("Name of Object")

[heady iris]

Oh, I mean for something to decide if it existed when the game started

using UnityEngine;
using UnityEngine.EventSystems;

public class SystemMenuEditorTester : MonoBehaviour
{
    public Camera editorCamera;
    public EventSystem editorEventSystem;

    // Start is called before the first frame update
    void Awake()
    {
        if (FindAnyObjectByType<Camera>())
        {
            Destroy(gameObject);
            return;
        }

        editorCamera.gameObject.SetActive(true);
        editorEventSystem.gameObject.SetActive(true);
    }
}

I think this'll do.

I'm working on the menu, and it's annoying to have to load it in alongside a scene that provides a camera

Yep, bingo. I was overthinking it a little. Thanks everyone (:

(hey, that was an XY problem!)

[somber nacelle]

oh yeah, you could also just check Camera.main if you have your cameras tagged as MainCamera. that would be a bit faster than using Find

[heady iris]

(doesn't that find an object by tag?)

[somber nacelle]

yeah, but it also should cache it too so it can be faster than just using a straight Find ByType or WithTag

[delicate flax]

historically yes, but iirc there is some special optimizations for Camera.main in newer versions of unity

[heady iris]

Ah, neat.

That would make sense, given how often it's used.

[hexed pecan]

Reminder that Finding by tag is also faster than finding by type since unity caches tagged objects in a list

[somber nacelle]

https://docs.unity3d.com/ScriptReference/Camera-main.html
ah looks like it just caches all MainCamera tagged objects anyway. so it's not even just caching it the first time you call it, it should already be cached

[heady iris]

😮

goodbye, serialized references! makes 5000 tags

also, any opinions on this problem i mentioned yesterday? #archived-code-general message

I've replaced some of my code with LocalizeStringEvent components

maybe I should just use more of those

[heady iris]

(e.g. instead of using code to fill in text for headers, I just put a LocalizeStringEvent component on them and assign the localized string reference on that)

[hexed pecan]

So youre wondering if you can use anonymous functions instead? I don't think you can unsubscribe an anonymous function since it isn't the same instance of that function

[heady iris]

That's what I was originally doing

And yeah, that's the rub

(well, I'm still doing it in several places. I need to clean this up...)

The main goal is to get rid of that boilerplate code

• subscribe to the event
• manually set the string
• unsubscribe from the event

It's easy, but it's very repetitive

and I know I'm going to start getting weird exceptions once I actually add more than one locale and switch languages

because some random-ass text element from an unloaded scene is still subscribed

[normal arch]

btw i ended up just asking chatgpt and it gave me this

Vector2 collisionDirection = new Vector2(platform.bounds.center.x, platform.bounds.center.y)
                - new Vector2(other.bounds.center.x, other.bounds.center.y); 
            if (Mathf.Abs(collisionDirection.x) > Mathf.Abs(collisionDirection.y))
            {
                if (collisionDirection.x > 0)
                {
                    // Object entered from the left
                    Debug.Log("Entered from the left");
                }
                else
                {
                    // Object entered from the right
                    Debug.Log("Entered from the right");
                }
            }
            else
            {
                if (collisionDirection.y > 0)
                {
                    // Object entered from the bottom
                    Debug.Log("Entered from the top");
                }
                else
                {
                    // Object entered from the bottom
                    Debug.Log("Entered from the bottom");
                }
            }

which surprisingly worked perfectly, i didn't know it was that reliable, but thanks for the help anyways

[delicate flax]

might be because I haven't worked with 2d for a long time, I would think something like that would not work if the collider is rotated. anyways, glad to hear it worked out 😄

[heady iris]

note that this will give you unexpected results for a non-square collider

[normal arch]

probably but basically all of the code in the script relies on that, so that shouldn't be on a problem

[heady iris]

you could move down onto the left side and get told you entered from the left

it's also in world-space, so the results may be surprising if you rotate the object

[normal arch]

it also relies on it being a box collider too

[wary ferry]

This is definitly the wrong place to ask this question but I figure their is a good chance someone here would know. Anyone know what I need to install to be able to compile c and c++ code? their seems to be a wild amount of slightly disagreeing info online. I have windows 10 installed if it makes a difference.

Just being pointed in the right direction would be massivly helpful

[leaden ice]

Definitely the wrong place but https://visualstudio.microsoft.com/vs/features/cplusplus/

[wary ferry]

Thanks i appreciate it!

[chrome trail]

I can't seem to get the rigidbody's angular velocity to update with this function. The only reason I can think of for why that might be is because the function is being called in Update rather than FixedUpdate but I can't see why that would be the case when I can use this function to manipulate the object's regular velocity just fine from Update. So why isn't angular velocity updating correctly?

    //It would be an issue
    void ChaseTarget()
    {
        //Calculate a path to the target
        NavMeshPath path = new NavMeshPath();
        NavMesh.CalculatePath(transform.position, target.position, filter, path);

        //Try to face the next point
        float targetAngle = Mathf.Atan2(target.position.x - transform.position.x, target.position.z - transform.position.z) * Mathf.Rad2Deg;
        float angleDiff = Mathf.DeltaAngle(transform.eulerAngles.y, targetAngle);
        float torque = Mathf.Clamp(angleDiff / Time.fixedDeltaTime, -moveTorque, moveTorque);
        rb.angularVelocity = Vector3.up * torque * Mathf.Deg2Rad;
        print(Vector3.up * torque * Mathf.Deg2Rad);
    }```

[hexed pecan]

Hmm make sure that the rigidbody's rotation is not locked in the inspector

[chrome trail]

Now that I think about it that might be it

[hexed pecan]

Also yeah FixedUpdate is the correct way, otherwise it will depend on framerate

[chrome trail]

Yeah, seems that was the problem. I was locking the axis I wanted to rotate

[hexed pecan]

Nice

I would also simplify that code by using Vector3.SignedAngle or something but if you love trigonometry, do it your way 😄

[chrome trail]

I have no clue what Vector3.SignedAngle is

If you're talking about how I calculate the target angle that's just how I've always done it

[hexed pecan]

That's fine

[leaden ice]

btw depending on transform.eulerAngles.y is possibly not going to work well due to potential gimbal lock. Euler angles can't typically be taken in isolation.

[chrome trail]

What do you mean?

[leaden ice]

I mean for example these two sets of euler angles:
(0, 180, 0)
(180, 0, 180)
are identical, so just reading the y part might not tell you what you need to know

it's better not to rely on euler angles in this manner

[hexed pecan]

I mean something like cs Vector3 targetDir = target.position - transform.position; float angleToTarget = Vector3.SignedAngle(transform.forward, targetDir, Vector3.up); // Y angle difference (-180..180 range) from current direction to target direction. 'Vector3.up' argument denotes that we want to use the up/Y axis // Then use angleToTarget as the Y torque
It might not be exactly the same result, not sure

[chrome trail]

What should I be doing instead then?

[leaden ice]

quaternions are generally better for this kind of thing, or direction vector manipulation, or some combination of the two

[chrome trail]

I've noticed that quaternions are a bit weird, though, and I don't really understand them all that well

[leaden ice]

They're not that bad, and easier to work with than euler angles once you get familiar with the API

people overexaggerate them

[hexed pecan]

I think the word 'Quaternion' just sounds evil by itself

[chrome trail]

In general the Y quaternion never seems to be the same as Y euler and the Y euler is a lot more legible anyway

[hexed pecan]

You are right they are not as scary as they seem, just use the utility functions (Quaternion.FromToRotation/AngleAxis/LookRotation etc)

[chrome trail]

So I would use what Quaternion function to get the information I want?

[leaden ice]

for example if you want to know the rotation you need to get from your current rotation to look towrads some direciton you can use something like:

Vector3 desiredLookDirection = target.position - transform.position;
desiredLookDirection.y = 0; // "flatten" it
Quaternion desiredRotation = Quaternion.LookRotation(desiredLookDirection); // create a rotation from the look direction
Quaternion rotationDiff = desiredRotation * Quaternion.Inverse(transform.rotation); // this is like desiredRotation - currentRotation```

now of course rb.angularVelocity is a set of euler angles so you can convert back at the end

[hexed pecan]

How would you translate that into a torque/angular velocity?

Just convert to euler from quat?

[leaden ice]

Vector3 eulerDifference = rotationDiff.eulerAngles;
rb.angularVelocity = eulerDifference.normalized * speedInDegreesPerSecond;``` something like this

I think haha 😉

[hexed pecan]

Something like what praetor suggested, but my suggestion still stands, you can get the Y angle difference with a couple of lines of code
#archived-code-general message

[leaden ice]

but yeah actually

SignedAngle is a great approach here too

maybe simpler

[heady iris]

Quaternions are a set of 4 values. The X/Y/Z values do not correspond to euler angles.

[chrome trail]

I'm aware

[leaden ice]

you shouldn't ever be looking at the y in the quaternion and indeed it is not the same as the y euler

[hybrid turtle]

hi dumb question about git and unity. I just pulled a repo with missing assets and want to checkout a fresh repo what would be the git commands to do this could I git branch -D name then git fetch origin branch would this delete the local branch then fetch me a copy of the remote?

[leaden ice]

you can do git checkout --hard origin/branch

if you want to pull a repository from scratch you should do:

cd ..
rm -rf myfolder // delete the folder entirely
git clone blahblah // copy the clone command from github here```

[hybrid turtle]

nah like I have another user whose working on it on gitlab and can't see it from my list of branches but only gitlab so last time I wanted to get his work I just git fetch origin branchName but ok will git checkout --hard origin/branch be run on the branch that I want a fresh repo of?

[leaden ice]

Idk what "fresh repo" of a branch means

a branch is part of the repository as a whole

[heady iris]

you can't see the branch you want?

[hybrid turtle]

nah you run git branch -a

[leaden ice]

if you want to checkout another user's created branch you can do:

git fetch
git checkout branchName```

[heady iris]

git won't know about the new branch on the remote until you fetch

[hybrid turtle]

doesn't show but when i run git fetch origin branchName

yes I fetched it

[leaden ice]

just do git fetch

then git branch -a will show it

assuming they actually pushed it up

[hybrid turtle]

yes I checked the logs they pushed it for sure

but on my end the textures all missing

so i want to fetch it again

and delete the local branch

i just ran git fetch again on the branch and pull but didnt do anything

[leaden ice]

git reset --hard origin/branchname will work (after a fetch)

if it's not working your friend didn't push things up properly

[hybrid turtle]

ok and I run this on the current branch

like the one I want to update

[leaden ice]

yes

[hybrid turtle]

ok awesome ill give it a shot

ok didnt do anything

can I just run git branch -D delete the local

then git fetch origin

again

and it wont tamper with his remote branch?

[leaden ice]

it won't tamper with anything but you're not going to get a different result

it sounds like your friend didn't save/commit/push their changes properly

[hybrid turtle]

cuz I've checked the logs in gitlab and seen for sure they were pushed

im confused bruh

[leaden ice]

I mean they have made a commit

it doesn't mean the commit includes everything

they might have forgotten to add files

or to save a scene

or a prefab

before pushing

[hybrid turtle]

yeah idk cuz I checked the commit details and saw all prefabs and textures were staged

so he kept telling me to get a fresh repo

and i was like idk ill fetch another then 😦

[leaden ice]

your friend doesn't seem to understand git that well

there is never a need to "freshly" pull a branch

if the commit is in your local log, you have it

[heady iris]

this means he doesn't know what's going on

[hybrid turtle]

welp then I'll go pester him again

thanks for the help though guys

[heady iris]

either you see the commit or you don't

https://xkcd.com/1597/ moment

[hybrid turtle]

cuz i saw this right here

these are the missing textures he claiming were pulled

[leaden ice]

Is this your computer?

Then you have the textures

so what's the issue?

[hybrid turtle]

nah this is his

not my comp

lol ill remove it cuz idk if he wants his screenshot on

but bruh

i think he didnt push it right

[leaden ice]

he needs to do git status and see what it shows

[hybrid turtle]

yeah i'll tell him idk i thought it was convention to always git status before you commit and push

[heady iris]

it will tell you if:

• the files are completely untracked
• the files are tracked, but not staged
• the files are staged, but not committed
• the files are committed, but there is a commit that hasn't been pushed to the remote

i run git status randomly when i'm trying to look busy

[hybrid turtle]

lol same

[heady iris]

ls; git status; ls; ls; ls; clear; ls

[hybrid turtle]

lmaooooo

[leaden ice]

it certainly is

[heady iris]

If git status reports there are no pending changes and no unpushed commits, then you will need to look at the remote branch he pushed to

[hybrid turtle]

yeah my git status not showing anything so im just going to make him push it again

[leaden ice]

he needs to run git status

and show what the results are

[hybrid turtle]

lol and yes tell him to always run git status and lol make sure that well the textures are all there

[heady iris]

the point is that they should not "be there" when he runs git status

[hybrid turtle]

well yes that he's committed them all

so staged should be empty

[heady iris]

correct.

[leaden ice]

He should see something like this:

On branch master
Your branch is up to date with 'origin/master'.

nothing to commit, working tree clean```

that will tell us he properly committed and pushed everything

[hybrid turtle]

yes exactly

[heady iris]

(so you'd better get that entire message from him)

[hybrid turtle]

yeah lol ill make sure to this time lmao instead of just trsuting it there after he said it pushed

[leaden ice]

if he sees

On branch master
Your branch is up to date with 'origin/master'.

Untracked files:
  (use "git add <file>..." to include in what will be committed)
    example

nothing added to commit but untracked files present (use "git add" to track)```
He needs to do `git add example` or `git add -A` and then commit & push

[hybrid turtle]

ok you can -A

i just do git add .

[leaden ice]

to add all unstaged files yes

yeah that works too

assuming you're in the root

[hybrid turtle]

does git branch -D delete only the local branch then

git push origin --delete remoteBranchName ive been using this to get rid of both but have never checked tbh that what they taught me in class

[lean sail]

github desktop would make stuff like this a bit easier

[heady iris]

Hey, question about dealing with multi-scene design.

I have my settings menu in its own scene. The player can ask to open the settings menu from the game scene. This makes the game controller load the scene.

The settings menu can have many nested screens open. If the user hits "back" (so, the B button) on the settings screen, it exits the current screen. When you exit the root screen, the settings menu needs to close. This means the settings menu needs to be able to say "hey, I'm closing!" to the game controller.

This means that the game controller needs to find the settings menu in the newly loaded scene and subscribe to the event. Right now, I'm doing this:

var op = SceneManager.LoadSceneAsync(systemMenuScene, LoadSceneMode.Additive);

op.completed += op => {
    FindAnyObjectByType<SystemMenu>().Closed += ToggleSystemMenu;
};

I'm...really not a fan of this! It just finds a completely random SystemMenu!

I'm thinking about how to make that connection. Nothing's jumping to mind here...

In this case I could definitely just use a static event

But I'm wondering about what'll happen if I have multiple scenes that need to do the same kind of thing

[heady iris]

I guess I can iterate over the root game objects of the scene and search for an object.

[heady iris]

I could definitely drop a "scene descriptor" component into each scene that holds all the interesting references, then find it and grab what I need

[hybrid turtle]

ok sorry one more question

what does git pull origin do

sorry git pull origin

i saw online it pulls from a branch

if it is not linked to an upstream branch and specifies which upstream to pull from

[leaden ice]

git pull pulls the current branch from the remote tracking branch

git pull origin branchName pulls a specific branch from a specific remote

[hybrid turtle]

oh ok so its just specificity

both do the same?

[leaden ice]

pull is the same as fetch + merge or rebase

[hybrid turtle]

cuz i also tried git pull origin remoteBranch as well

but made no difference then git pull

[leaden ice]

why would you expect it to make a difference

[hybrid turtle]

lol idk false hope

just trying everything while I wait for kicks as well

[rotund burrow]

i want to make a better structure for an inventory system.
Classes ---- Weapon : Item, Recourse: Item, Junk : Item.
Paired classes ---- WeaponController : ItemController, InactiveController : ItemController
when i click on weapon, it shoots. but for junk and recourses nothing happens.
I need something like this dictionary to use (the analogy just makes it simpler):
pairedClasses = { Weapon : WeaponController, Junk : InactiveController, Recourse : InactiveController}
The item class will have method like this
protected ItemController GetController()
return new pairedClasses[this.GetType()]
right now i'm just overriding the GetController method in each item derived class but i want to only have 1 method at the base (item class)

[leaden ice]

Why not just have an abstract void Use(); function on Item and let each subclass implement it.

it seems like you're trying to fight against the type system instead of just using polymorphism as intended.

[rotund burrow]

because items will process inputs and work differently for player and npcs

[leaden ice]

I don't see how that changes anything

Maybe you want some parameter then to distinguish who is using it? abstract void Use(Character user);

[rotund burrow]

consider an item like a bow that charges and then shoots on button up. then normal gun that shoots on button down. then for ai it will work completely differently

one method is not enough for this

[leaden ice]

Right that's why you let the subclass implement things

you're right

one method is not enough

abstract void Use();
abstract bool RequiresCharge { get; }
abstract bool FullAuto { get; }
abstract bool BeginUse();
abstract bool EndUse();```
etc

[rotund burrow]

i just dont want to be constrained in what items can do. i dont think i will be able to fit it like that

[quaint crypt]

I want to have a character preview in the inventory. Much like in wow / rust / zelda etc. Is there any better way to achieve this than: 1. create a copy of the character model, 2. attach afront facing camera to it, make it render to a render texture, add that render texture to a raw image?

[leaden ice]

there's no difference between this and the dictionary thing. It's just a better way of organizing the code

where each class has its own behavior inside itself

instead of some high level controller having intimate knowledge of all the subclasses

[rotund burrow]

where do you put the logic that activates the item? for example, a gun in an fps game. For player you will have to check facing direction, range etc.. For ai it is controlled completely different

[lean sail]

better in what way?

[rotund burrow]

right now i just have it in player and enemy classes but i want it to be cleaner

[hexed pecan]

Again, the way you described is a pretty standard way of doing it

[leaden ice]

In the player controller and in the AI controller

[quaint crypt]

design, approach, performance. any

[leaden ice]

Each of them will call Use on the gun when they want to shoot

[hexed pecan]

As an alternative, you can skip copying the character and just render the existing character, with a culling mask on the camera so that it renders only the character
But that's a design choice, whether you want to have different animations on the preview or not

[leaden ice]

Choosing when and how to activate the item is not the purview of the item itself.

But what happens when the item activates is

[rotund burrow]

what if there are 50 guns? you cant fit it all in one class

all with different logic, like i described with a bow

[quaint crypt]

unfortunately i don't like the fact that some animation sip through and can be seen on the preview

for example, opening the inventory while falling etc

[hexed pecan]

Yeah, I see

[leaden ice]

That's exactly why you do what I'm saying...

[quaint crypt]

anyway, thanks a lot for the help @hexed pecan @lean sail !

[leaden ice]

Your approach is to put it all in one place

[rotund burrow]

i dont think i understand you correctly

[light jay]

HI all,

i got a question about my code:
https://pastebin.com/WJ8UX2Ur

What i expect to happen is that this gameObject with this script get put under DontDestroyOnLoad in the Hierarchy but i doesnt and when i switch scene's ( Main Menu to World ) it gets destroyed
and i cannot figure out where i did something wrong

[hexed pecan]

Make sure the gameobject is a root object