archived-code-general

[quick token]

oh, that makes a lot more sense

[hardy pasture]

That's how we end up with untested spaghetti code with files 15000 lines long, classes 6000 lines long, and functions 2000 lines long with 10 different arguments. Because programmers think that DRY, SOLID and GRASP is unnecessary

And it's all untested, of course

[cold parrot]

right and you think a fully DRY & SOLID codebase is something you actually want to have, you try that and report back in a couple of years.

[vestal arch]

it's an upcast to Object to ensure the right bool conversion is called

isn't it necessary though

[quartz folio]

Why?
https://github.com/Unity-Technologies/UnityCsReference/blob/master/Runtime%2FTransform%2FScriptBindings%2FTransform.bindings.cs#L149

[vestal arch]

ah there's no cast at all there

[knotty sun]

You should go had have a look at the very popular Easy Save asset, their code is full of it
it's called boxing and unboxing btw and it's horrible

[thin aurora]

Ah yes, very hard to understand that the first three parameters are always x, y and z

If this is your argument then just use the Vector3 overload

[vestal arch]

(isn't it this signature though https://github.com/Unity-Technologies/UnityCsReference/blob/master/Runtime%2FTransform%2FScriptBindings%2FTransform.bindings.cs#L147)

[thin aurora]

Also, if you have unnecessary arguments then often the answer is "pass an object". This is not applicable here, though. But neither is it necessary anyway because it's very clear what the usage is

Actually, it would be applicable. But guess what, there's the Vector3 overload so it already applies

[vestal arch]

ah it's already typed as Transform which doesn't override (or inherit an override) to Object's conversion, so no upcast is necessary, right?
@quartz folio

[quartz folio]

To be clear that isn't boxing, that's UnityEngine.Object

[hardy pasture]

It stacks together in the solution once you take this approach and say "it's fine". But yeah, Translate(0, 10, 0); is harder to write and harder understand than TranslateY(CharacterHeight);.

Magic numbers, too.

[thin aurora]

BTW the part you reference here talks about passing boolean parameters

Idk how this is relevant

[hardy pasture]

The relativeTo is used as boolean in the if statement.

How is this not relevant

[thin aurora]

So like I said, if this is an issue use the Vector3 overload

Translate (Vector3.Up * 10)

[hardy pasture]

It's used as a boolean in the function

[quartz folio]

It's just checking that the object exists

[thin aurora]

Also, TranslateY(10); still uses magic numbers so I don't see your point

[vestal arch]

that's not a fair comparison at all lol, think about Translate(0, CharacterHeight, 0) instead, or Translate(Vector3.up * CharacterHeight) if you really don't want magic numbers

[chilly surge]

Until the day someone copies TranslateY() 3 times and forgot to change it to TranslateX but no one noticed because it's just one letter difference.

[vestal arch]

it isn't; unity overrides bool conversion to check if an object is valid

[hardy pasture]

That's very verbose, now we're doing math.

[thin aurora]

Are you telling me Vector3 is not used for math?

My life has been a lie

[hardy pasture]

Well, it's Translate(0, CharacterHeight, 0) vs TranslateY( CharacterHeight)
The Translate still has 2 magic numbers in it.

And 2 more arguments

That you have to repeat all the time

[vestal arch]

0 is not really a magic number

[hardy pasture]

When you translate by 1 axis

[quartz folio]

0 as a magic number is hilarious

[vestal arch]

you could use like, uint32.MIN_VALUE, i guess? lol

[thin aurora]

Those are not magic numberssssssssssss

10 is

You still use 10 with TranslateY

[hardy pasture]

It is, but you can encapsulate it and not repeat it. Again, DRY.

[vestal arch]

i feel like you're just spewing buzzwords without really understanding how to properly apply them now

[thin aurora]

I feel like this whole discussion is based on somebody being overly confident in the wrong approach and not accepting valid criticism

[chilly surge]

Next thing someone is going to tell me if (list.Count == 0) abuses magic number and I should refactor to if (list.Count == MIN_LIST_COUNT).

[thin aurora]

Wait until you see my constants file that's just 100 aliases for the number 0

[quartz folio]

The whole convo is a great way to not get anything done. Either way just write the damn code, extension methods or not it doesn't matter one bit if you're getting shit done

[vestal arch]

zero, z, zreo, zeor, Z, Zero, ZERO, invS1, nul, nil, nada

[cold parrot]

conventionally -1, 1, * 0.5, / 2, 0 and the like are not magic numbers

[hardy pasture]

That's what I started with: I don't want to unnecessarily write Vector.right, multiply, create new vectors, etc.
What we really want to write is: transform.position.y += characterHeight; and my solution is the closest to that.

[knotty sun]

where is if (List.IsNullOrEmpty(list)) ?

[cold parrot]

!.Any()

[knotty sun]

Linq, no way

[hardy pasture]

Nah, you guys are just doing personal attacks, because your arguments aren't accepted. So far Anikki did multiple personal attacks on bob and now you did one on me.

[thin aurora]

Jokes on you, just wait until 0 is no longer truly zero and my constants allow me to refactor my code instantly

[vestal arch]

this is a hilarious conclusion

[hardy pasture]

Nope, I'm looking for better practices all the time and accept new ways of doing things all the time, once it's demonstrated that they're better.

[vestal arch]

because your arguments aren't accepted
this is exactly what qyou is saying

[quick token]

came for the code, stayed for the entertainment

[still jungle]

i guess you cant because its Unity Native Object with shallow manager on C# side so it requires syncing and due to that directly manipulating member of reference is not possibble

[vestal arch]

you can manipulate a member of a reference
you can't manipulate a member of a value

[hardy pasture]

No, the way you refactor it is:
if (list.IsEmpty()) and if(list.HasItems()). That's better way to do it.

[thin aurora]

Because 0 is a magic number?

[vestal arch]

...well, i think that proves quite a few of our points

[knotty sun]

calling a method is better than checking a constant?

[vestal arch]

smh Count is already basically calling a method in disguise

[thin aurora]

I wonder how Count/Length == 0 is not already clear enough

[still jungle]

this make a sense, and since you know that, you already know what happend under the hood

[chilly surge]

That's a lot worse. list.Count == 0 is a universially understood way of checking if a list is empty, the fact that .IsEmpty doesn't exist and you have to go out of your way to make a method, makes everyone who reads your code to do a double take "why isn't it simply written as list.Count == 0, why does this project adds its own empty check extension method and calls it? I better look into that method to make sure I'm not missing anything."
It wastes so much time.

[quartz folio]

I've had so many physical double takes when reading code like that lol

Recoiling in confusion

[vestal arch]

i think this boils down to using known ways of doing things when working with other people
when you ask here, or when you're talking about that kind of thing here, you're implicitly working with other people

if you're using it in your own project alone, that's fine, but you're trying to justify it to a wider audience here, and that's just not gonna fly

[thin aurora]

I think this is honestly just a matter of experience and understanding the point of KISS

Adding a ton of aliases and extensions is a good way to bloat your project

[hardy pasture]

Yes, because the code explains itself and tells the reader what is being done here, explicitly.
The IsEmpty explains the reader what is being done on the conceptual level and allows the reader to read and understand the implementation if the reader wants.
The Count == 0 only shows the implementation without explanation and forces the reader to reverse-engineer the code to understand what's being done here on the conceptual level.

Refactoring.

[vestal arch]

I better look into that method to make sure I'm not missing anything.

[knotty sun]

this is Game Dev not App Dev

[vestal arch]

and forces the reader to reverse-engineer the code to understand what's being done here
what even???

[quartz folio]

But the implementation is the explanation

[vestal arch]

it's literally asking, "is the size 0"

[quartz folio]

It's so simple it's self apparent

[vestal arch]

i don't think this justification would fly in app dev either

[knotty sun]

it could, depending on the specific implementation and the scope of the App

[vestal arch]

i think if this isn't synonymous with "empty" in your mind that kinda speaks more to your experience

it's a thing that works in every language, every context

well maybe not cstrings

[hardy pasture]

The implementation is the implementation, it doesn't tell you what's being done, you have to understand and then reverse-engineer it.
Also, we're examining trivial examples. In the real world on real projects you end up with this kind of code with the approach you guys are suggesting:

That's just a small part of the method.

https://youtu.be/SVRiktFlWxI

[quick token]

who is uncle bob and why do we keep citing him

[vestal arch]

yes, Translate is a trivial example, we're all talking about trivial examples here
we're saying you're making something trivial harder to parse

[hardy pasture]

This is programming, the principles and design patterns apply universally.

[quartz folio]

I'm not talking about what this is, and nor is anyone else. We're all talking about trivial examples

[vestal arch]

they absolutely do not

[knotty sun]

wrong

[chilly surge]

Similar situations exist everywhere not just list.Count == 0, for example in a backend app if you see someone writes a SQL query SELECT user FROM users WHERE primary_key = ? LIMIT 1, everyone reading that code will be confused by LIMIT 1 ("why is it there, how is it ever possible for there to be more than 1? Am I misunderstanding how users table is set up?")
Perhaps you don't have experience working with other people, but it's often not about what you subjectively think good code looks like, often it's way more important to have everyone on the same page and work effectively. A big part of achieving that, is to write code in a way that everyone understands.

[vestal arch]

embedded systems, competitive programming, shell scripts, etc will have vastly different ways of going about things

those are the outliers

in different situations for more "similar" kinds of programming, there will still be different ways to think about stuff

even more so with different paradigms

[hardy pasture]

They absolutely do. Here's a few examples:

1. Strategy pattern. Used everywhere: Java library, C# .NET, C++ library, Unity, Unreal, Sprint boot framework. I can continue forever.
2. Single responsibility principle. Same list as before.
3. DRY. Same as before.
   Are you going to tell me those 3 don't apply to Unity?

[vestal arch]

for example, SOLID doesn't apply universally, they only apply to OOP patterns

[hardy pasture]

No, wrong. Uncle Bob uses functional languages, it's his favorite, actually.

https://www.youtube.com/watch?v=GcJgGy-dfvE&pp=ygUgdW5jbGUgYm9iIGZ1bmN0aW9uYWwgcHJvZ3JhbW1pbmc%3D

[cold parrot]

he's an enterprise software architecture consultant who co-wrote the agile manifesto and wrote a bunch of books that argue for specific principles as the solution to all OO software development problems. He has now personally abandoned OO in favour of FP. Generally in game-dev his ideas are seen as somewhat counter-productive and missing the mark. they are generally not helpful in writing performant code or code that is too complex to be developed in a test driven environment.

[vestal arch]

how does the L in SOLID apply to functions

[vestal arch]

are you trying to say those are the only 3 patterns that exist? or that all patterns are universal?

observers aren't used much in js.
solid isn't used in c.
neither are really used in shell scripts.
etc

[hardy pasture]

For FP, it can be applied to generic types. Function with arguments of a generic type should ensure that the implementation works correctly, regardless of the concrete types of passed arguments.

https://medium.com/@mkocik/solid-principles-in-functional-programming-b9b83aeddf80#:~:text=Liskov Substitution Principle&text=Paraphrasing it%3A,not to break their behavior.

[cold parrot]

the problems that SOLID tries to address dont exist in FP or any non-OO code really

[hardy pasture]

Those are just 3 examples. All of them apply universally, you just can't apply them in all situations all the time, of course.

[vestal arch]

lmao what?

[hardy pasture]

But saying that X pattern/princple can't be appliied in the whole industry is just wrong

[vestal arch]

how are you gonna apply solid in shells

or cp

of course those don't really fit "industry", but the original claim was about programming as a whole

[knotty sun]

no it's not, there are many specific situations where ALL of your principles and patterns have to go out of the window. Driver software to name just one

[cold parrot]

these principles typically apply most in large team, long term, high turnover, monolithic, java/.net development situations.

what also applies to these projects is "excessive funding and waste".

[hardy pasture]

They're not. Again, he says that these apply in small teams that want to do small things. And the way to do big things is to use a lot of small teams that do small things.

[knotty sun]

not to mention shit performance

[hardy pasture]

You guys just saying false things

[vestal arch]

who is "he"?

uncle bob?

[hardy pasture]

And misrepresenting my position.

[vestal arch]

you really gotta come up with better references than your idol

[hardy pasture]

Yes.

[vestal arch]

yikes

[cold parrot]

"he says" is never a good argument

[hardy pasture]

So far only you and Anikki did personal attacks.
I'm not saying it's right, because he says so. In this case I said he says is because Anikki misrepresented the position

Where "it only applies to big teams"

Which is not what it's for

Even if you disagree with it

You can't just take and misrepresent the original position like that

[vestal arch]

i have an engineer uncle of my own
his name is uncle dane and generally he solves most of his problems by jumping really high and building guns behind the enemy

[vestal arch]

it's a technique derived from ninja code called ninjaneering

[hardy pasture]

Yeah, yeah, idk what your problem is. Just because someone doesn't take your arguments, it doesn't mean you have to start personal attacks.

Are you 20 or something?

Anyway, let's hear what engineering practices you guys are following? Or is it just "anything that comes to my mind today" goes?

[vestal arch]

im not sure where you're seeing personal attacks

i'm asking for better references

1 guy on a youtube video is not a very good reference

[vestal arch]

we're following similar principles, we just know when they don't apply

[quick token]

not sure if this helps the argument at all. but i tend to find that working code is better than clean code (i work alone though so )

[knotty sun]

basically yes, backed up by 50 years of professional experience. The solution should fit the problem not the problem is shoehorned into my desired solution

[thin aurora]

Generally you should first make it work, then worry about making it pretty

[cold parrot]

I think the misunderstanding we're having here is that Kula thinks we don't understand SOLID & Patterns and reject it categorically.

[thin aurora]

Making code pretty has the effect that code might not end up scalable enough and/or working in general

[vestal arch]

i guess that would fit the reactions, yeah

[hardy pasture]

i feel like you're just spewing buzzwords without really understanding how to properly apply them now

you need to stop drinking uncle bob cool aid
this stuff (ideas of clean, dry, solid etc) is extremely outdated/incomplete and in many cases tangential or irrelevant to gamedev.
Anikki also changed his message

[thin aurora]

For example, making aliases for Translate hides the extra parameters that might be included, so you just make it worse for youself

[quartz folio]

I personally think the Translate overloads are totally fine, clear, and not impeding anyone if they're seen by people in your project as useful shortcuts to a common pattern

[vestal arch]

idk what anikki's on since im not him, but

i feel like you're just spewing buzzwords without really understanding how to properly apply them now
this wasn't meant as an attack
im commenting on how you're using the buzzwords, and how lackluster your references/sources have been, and that you should probably consider if you do know how to properly apply them
was the implied question/reflection prompt not clear enough?

[hardy pasture]

Do you mean you Steve? Are you 70 or something?
if not, then what you're following is just a different approach to other people then Uncle Bob.
Which is cool by me.
You're the ones started talking shit and telling me my solution is bad by your standards, then I presented that I have good reasons to do what I do, to which you continued attacking me and what I do.
If you're following an alternative school, that's cool by me.

[quartz folio]

but globally applying these "clean code" ideas is absolute nonsense made by people with very narrow experience, applying patterns too broadly—often at odds to those who first espoused the ideas

[knotty sun]

yes I am 70 or something

[hardy pasture]

So you're saying you persosnally have 50 years of programming?

[vestal arch]

to be clear; im not against the overloads. they're probably fine if you use them a ton
im against your justification, and how you're presenting dry/solid as strict rules to abide by
they are not. they are guidelines

definitely agree with vertx above, i think he said it clearer

[knotty sun]

yes, exactly

[wheat spruce]

what would be a smart way to implement a free transform object, like what you get in photoshop/illustrator, that will be used to drive methods from my XForm interface that my shapes contain?

public class BoxShape : IShape, IXForm
{
  ...
  ...
  public void Translate(Vector2 translation) { }
  public void Rotate(float angle) { }
  public void Scale(float scale) { }
}```
I need each of my shape to have the tool, but I'm not sure if theres any special class types that unity provides (like a scriptable object) that would be suitable for it

[hardy pasture]

but globally applying these "clean code" ideas is absolute nonsense made by people with very narrow experience
This is again, false.
Uncle bob, again, programmed in assembly back in the 70s, then wrote C and C++ and programmed hardware, then Java and C# for web.
That's like more broad than 99% of people.
If Steve is telling the truth about his experience, then he probably did the same or very similar in the type of languages he used.

[vestal arch]

ok so

[wheat spruce]

my Shapes arent actually unity objects, so its not like I could give them a Unity Component

[vestal arch]

if you're saying uncle bob is like, the only resource you have

[knotty sun]

Actually I did a great deal more than that

[hardy pasture]

Yeah of course

[vestal arch]

and you're saying uncle bob would agree completely with your stance, that dry and solid are strict rules

[hardy pasture]

Because it all changed so much compared to the 1970s

[vestal arch]

yeah i don't think i trust this uncle bob, or at least your portrayal

[vestal arch]

do you have a ref on this part

globally applying these "clean code" ideas

[knotty sun]

no, nothing has changed, we are still dealing in zeros and ones, the difference is we, who were programming back in the '70's' know how they work

[vestal arch]

i think you're conflating what you've heard with your own interpretation

[hardy pasture]

No, he wouldn't agree. As I wouldn't. They're not strict rules.

You don't apply them all over the place all the time.
You apply them for reasons. In this case I do it, because I don't want to repeat and write more code that's necessary, because it's more text, harder to understand, more errror-prone.

[knotty sun]

Does Uncle Bob work for the RaboBank?

[quick token]

this chat seems a little busy rn, (+ i am not smart enough to understand that so) maybe forward it to #archived-code-advanced (unless someone answers, in which case just ignore this i guess)

[vestal arch]

im not sure how 3 numbers is harder to understand

[hardy pasture]

I know how they work, too. I can read this just fine:

https://i.sstatic.net/8NQjj.png

[vestal arch]

😒 what is this now, an old age contest

[knotty sun]

ok, so how does that fit into your patterns and principles?

[quick token]

what does this have to do with the translate method again?

[knotty sun]

yeah, but could you write it from scratch?

[hardy pasture]

Yes

[wheat spruce]

the issue isnt clean code, the issue is when developers get into a mindset that their code has to use patterns, without considering if they need to really use one

[knotty sun]

absolutely

[wheat spruce]

like somebody might think theyd need a whole finite state machine, to avoid the idea of having

if(x) {}
else {}
else {}
else {}
else {}```

like its not the greatest approach for that

but some people have a mindset that anything like this block, is terrible, so it HAS to be a state machine

[vestal arch]

it's more text
this is generally not a justification. ask about more complexity or repetition
harder to understand
is it? using a 3d vector, or 3 numbers, literally the common data structure/format for dealing with 3d space in unity
more errror-prone
how so? in my experience, well-tested library code is often more resilient than your own code, especially since you can modify your own code

i'd get the justification if it were about, say, repetitiveness, or personal preference, but you gotta stop pretending these objective/general reasons exist

[wheat spruce]

especially when the justification is "thats very slow to have those else statements" and not really understanding that the speed of that else block, means absolutely nothing for the vast majority of cases, especially in the early days of code or when youre learning

[hardy pasture]

how so?
You can pass the arguments in wrong order and when you're reading, you need to remember which order they're in.
Repetitiveness is also the main factor, that's what I asked my question about: how do I remove having to duplicate my code.

[wheat spruce]

"is it slower to use function X or is it slower to do function Y" really shouldnt ever really be a concern

[vestal arch]

You can pass the arguments in wrong order and when you're reading, you need to remember which order they're in.
it's literally xyz
how would you mess this up
it's xyz everywhere

[wheat spruce]

not unless youre doing something really technical, but if that was the case, its not a question youd ever need to ask a discord server

[vestal arch]

and what coordinate system unity is in doesn't matter, since your own utility methods are also using xyz

[wheat spruce]

honestly, a lot of people working with unity, think they have to make all their code exist purely for the sake of unity

like people who think every class they write has to derive from MonoBehaviour

[vestal arch]

with your utility methods, you could also call the wrong function. x and z are quite close on the keyboard
noticing the difference would be on par with noticing wrong argument order, but wrong argument order would be harder to accidentally happen since that's separated in time, you can't really typo that

[wheat spruce]

its something like they think they cant use anything that Unity has, unless its a MonoBehaviour class

[vestal arch]

Repetitiveness is also the main factor, that's what I asked my question about: how do I remove having to duplicate my code.
this, this is a good justification

the rest was not

[hardy pasture]

with your utility methods, you could also call the wrong function. x and z are quite close on the keyboard
Sure, but it explicitly shows you right there which method you called: the code explains itself explicitly. That's the point.

You don't have to do any reverse engineering whatsoever

The code you guys showed and wrote requires the reader to do reverse-engineering

[vestal arch]

xyz also explains itself, i have no idea where you're getting this idea of needing to "reverse engineer" a few numbers

it absolutely does not

[hardy pasture]

My code explains itself well and there is nothing you have to reverse

[vestal arch]

that is not what reverse engineering is

and there really is no extra mental load

[wheat spruce]

I'm sure theres at least a dozen people in this server, who would look at this and think "how is this able to use Mesh or Debug.Log, when Zen isnt MB"

using UnityEngine;

public class Zen {
private GroundPoint_deprecated[,] _centroids;

public Zen(GroundPoint_deprecated[,] centroids) {
  _centroids = centroids;
}

public Mesh GenerateMesh() {
  Debug.Log($"New Mesh Generated");
  return new Mesh();
}
}```

[vestal arch]

if you're using TranslateX/Y/Z, you already have to be familiar with the XYZ coordinate system
and using 3 numbers for XYZ is very straightforward

[quick token]

got any examples?

[hardy pasture]

Ok, the translate method kinda does. I agree. But.
You still have to repeat the zeroes all over the place when you have to translate by just 1 axis and you can confuse the axes by mistake when calling that function.

[thin aurora]

I think it's a very valid question considering this could be related to heavy math computations where extra operations really matter. In other cases it would be way less important. Having these extra methods make a massive difference unless you inline them.

[vestal arch]

you absolutely cannot confuse the axes

they are xyz

[quartz folio]

Clean code thread

[wheat spruce]

the solution is to come up with a constant rule for your code and never break it

[hardy pasture]

While passing the numbers, you can pass them in the wrong order.

[quartz folio]

#1324346939658997780

[vestal arch]

by mistake? then you misunderstood xyz to begin with
by typo? that's just not gonna happen

[quick token]

[quartz folio]

#1324346939658997780

[azure frost]

Here's a look at that rigidbody-based character controller thing I made a while back... it's not working right.

[quick token]

whats wrong with it?

if it's the floaty jumps, you can add a gravity multiplier (just an addforce going down) near the end of the jump or something

[azure frost]

Well, other than all-around jank:

-There are times where the jump button doesn't respond.

-I want to implement the character model independently looking from the camera.

-The jump pad particle thingy doesn't launch me correctly.```

It'll take a LOT more showing than just the game view as to what's going on here.

[quick token]

-There are times where the jump button doesn't respond.
that's typically caused from checking inputs in FixedUpdate

[azure frost]

Well, there's a rigidbody system that might cause the body not to detect the ground as it should.

[quick token]

-I want to implement the character model independently looking from the camera.
i dont mess with models much so i dont know exactly how to fix that. i'd imagine you could just have the animated model be a seperate object

[quick token]

could you be a little more specific? whats the code look like, if you dont mind sharing

[azure frost]

I'd show more of how this janky setup works, but I can't screen-stream here.

(I am also terrible with code...)

[earnest gazelle]

Collider2d.bounds is not Bound2d !!!
and Contains(point) method checks in 3d space :/

[azure frost]

Here's what scripts I currently have in here.

https://www.youtube.com/watch?v=qdskE8PJy6Q
The rigidbodyforce script is kinda derived from this video.

[quick token]

o hey, that's a pretty underrated channel. they're code is really good

also, i am actively reading the code, gimme a sec

have you tried just making the onGround variable the raycast?

a raycast just returns a bool anyway

grounded = Physics.SphereCast(transform.position, GroundCheckRadius, -transform.up, out RaycastHit groundHit, GroundCheckDistance, whatIsGround);

(from a recent, similar character controller of mine)

guh, i have no idea how to properly use discord formatting

[azure frost]

According to my bro, the groundCheck function checks how far the ground is from the player.

[quick token]

i think the code should look like this

    RaycastHit groundCheck()
    {
             RaycastHit groundDetection;
             onGround = Physics.Raycast(transform.position, -transform.up, _rideHeight, out groundDetection)

             return groundDetection;
    }

[quick token]

yeah it should, looks correct

you sure its not the inputs?

[vestal arch]

generally they just check whether or not the player is on the ground rather than distance

[azure frost]

There's also the whole "Ride Height" and "Coyote Time" to consider...

But, before we get to that...

Compiler Error

Argument 4 may not be passed with the "out" keyword.```

@quick token

[quick token]

i didn't import the code into a project so i dont have the autocomplete

might just be the wrong ordering

[quartz folio]

To be clear, you're getting errors and autocomplete in your IDE, right?

[azure frost]

I'm using visualstudiocode.

The errors show up in Unity's console.

[quartz folio]

!vscode

[tawny elk]

Visual Studio Code guide

If your IDE is not underlining errors in red or autocompleting code,
please configure it using the link below:

https://on.unity.com/vscode

[azure frost]

With how janky and dysfunctional my setup is, I am seriously considering building off of movement lab.

https://www.youtube.com/watch?v=9_Q9UxK7VF8

[quick token]

sure, that works. i think there's a bunch of character controller scripts on the asset store too. should be a couple for free

[frail halo]

Hello so i got this code here and there seems to be a problem. This is supposed to show a dialogue once a character is near an object and the players presses E. but for some reason it wont work The PlayerIsClose bool never turns true even when the player is near the box collision of the object. Although when i manually set it to true it does seem to work. Was wondering if anyone knows a fix or advice https://paste.ofcode.org/7ESpgUdrKJJYuGBX78iH49

[quick token]

OnTriggerEnter only checks the first frame of the collider colliding with the trigger

try changing it to OnTriggerStay

you should also consider adding some kind of check before, to make sure that just any object cant trigger it

[leaden ice]

YOur logic doesn't make any sense:

            if (Input.GetKeyDown(KeyCode.E))
            {
                playerIsClose = true;
            }```

pressing the button doesn't make the player close

OnTriggerEnter makes the player close.

[quick token]

yeah i was going to say that too, consider changing the variable name

[leaden ice]

You should have that bool and then in Update handle the input

void Update() {
  if (playerIsClose && Input.GetKeyDown(KeyCode.E)) {
    ShowTheDialog();
  }
}```

quick dirty example

OnTriggerEnter/Exit should just set the bool true/false

in fact line 25 of youir code is ALREADY doing this

Just get rid of the input check in OnTriggerEnter

[quick token]

https://paste.ofcode.org/gpDFDi7DdVuSCyrp9mrX8i could someone help me clean this script up a bit. it was going fine until i tried adding the ability to play previous songs

the whole script seems to work fine. but when i get to the last song the "previous" function stops working

i get an error saying the queue is empty
Playlist.PlayPrevious () (at Assets/Music/Playlist.cs:78)
InvalidOperationException: Queue empty.

[thin aurora]

previousMusicQueue is empty, you need to enqueue audio clip instances before you dequeue

Alternatively, use TryDequeue

[trim schooner]

Unless I'm missing it, you never add anything to the previousMusicQueue?

[thin aurora]

I don't see Enqueue added anywhere anyway. Source: ctrl + f

[quick token]

    public void PlayNext()
    {
        previousMusicQueue.PlaceAtFirst(musicSource.clip);

        musicSource.clip = musicQueue.Dequeue();
        musicSource.Play();
    }

yeah its a bit messy, im trying to figure out how to do it more cleanly

its in the commented out part at the bottom

[trim schooner]

When a song finishes (or is skipped) you need to add that song to the previous queue - you are not currently adding anything to it, so you get an empty error

[quick token]

previousMusicQueue.PlaceAtFirst(musicSource.clip); < this line is a function i added manually (im trying to replace it with a list but im having some trouble with it)

[frail halo]

oh lol i get it now thanks guys ill tweak the code a bit

[quick token]

i feel like i should be able to replace all this with a list, but using a queue seems

until i try adding the ability to play the previous song

[trim schooner]

never heard of Queue before

[vestal arch]

shouldn't "previous" be a stack

or you could just use a single list that you transverse, instead of a queue

[trim schooner]

Just use a List, and get the last element of it

[vestal arch]

that way choosing the previous song would just be going backwards through the list

[quick token]

oh yeah that should work. i kept thinking i'd have to use a queue for that for some reason

thanks for the help, i'll try fixing it now

[vestal arch]

if you do want a queue, the "previous song" queue would have to be a stack, so you get FILO order

the most recent song is the first taken out (which i guess is more like LIFO)

[thin aurora]

I mean the whole point of a queue is to return items in a FIFO order. Using a list with a manual index just reinvents the wheel

You just have to properly add items to the list of previous songs as a stack and it works fine

[quick token]

i just googled what a stack is, how have i never seen this

that sounds a lot more like what i was trying to do with the messy queue

[thin aurora]

Also, considering Dequeue/Pop expects an entry you either have to check if it's not empty and/or use TryDequeue/TryPop instead

[vestal arch]

but wanting the extra layer of going back to the previous element kinda defeats the purpose of the queue
you'd have to put the current element at the front of the queue

[quick token]

oh yeah i did get an error from not having that. it was only when the queue was initially empty so i left it out. would be good to remove another error though so, will do 👍

[still jungle]

if the list of songs its constant, simply manipulating the index of array would be the best approach

[thin aurora]

I suppose it could, but seeing as the code had a shuffle system it would bring unnecessary complexity

[vestal arch]

the constraints you have here are

• playing stuff in insertion order (so, queue, list, array)
• adding new stuff? (not array)
• being able to go back (not queue)
  so you will need some kind of list

oh if there's also shuffling then a list makes even more sense

[still jungle]

its shuffled on awake, so from this point of view its constant, the array doesnt change its element order/size

[thin aurora]

I wasn't disagreeing, just pointing out the obvious

[vestal arch]

you might be interested in #archived-code-general message

[quick token]

oh good point, i never actually thought of that

[thin aurora]

If I were you I'd stick to having a Bag like a List or array where all your songs are, and a queue for the items in the current playlist. Lastly, have a Stack for previous songs.

Depending on your use case, separating them in three might be better here instead of a single collection, considering possible added complexity when you want more features such as shuffling. YOu might not wan tto shuffle your main collection that you are listening through

Also, for example, shuffling the main list might give you previous songs you don't want. Instead you want to keep those as "listened" and not repeat those until you finish the list

@quick token

The Bag class in this case would be List<> or Collection<>

Also depends on use case but this is just a quick thought whilst considering what should be possible

[quick token]

for a little extra context, the musicInitial array is all of the songs in the initial order.

will do with the rest though, thanks for the help :D

[vestal arch]

i mean, you could just shuffle the part of the array/list after the playhead

[thin aurora]

Then you would no longer have the unshuffled list

Like with Spotify, you can shuffle or unshuffle

[vestal arch]

i haven't seen how spotify does it but im pretty sure youtube handles that by just

if !shuffle
  play next
else
  play random
```over the entire list
and it's not too bad, tbh

[still jungle]

or just track the position of a song in inital array, each song have its id by it. Then simply another list with those ids being pushed in on song play, so you have tracked history of played songs.
Simply:

1. Array of all songs
2. Array of played songs ids
3. Temporary array for shuffled ids of songs to play

[thin aurora]

Yes but how would you unshuffle a shuffled list

Because pretty sure even in Youtube's case you'd unshuffle and get the original list back

[vestal arch]

youtube doesn't shuffle the list

the shuffle option just ignores the list and picks a random entry

[thin aurora]

At this point just use two separate lists lol

A list of ids is less convenient than a list of references

[vestal arch]

youtube mobile (a while ago, idk if it still does this) has a shuffle button instead of a shuffle option lol

[thin aurora]

But then you lose the ability to know which song comes next unless you keep track of the list of ids that come after

Also, risk of playing the same song twice in quick succession

[vestal arch]

yeah, sometimes that risk isn't too bad

[vestal arch]

not really? just pre-choose

[thin aurora]

pre-choose?

[vestal arch]

uhhhh not sure how to explain what i have in mind

[thin aurora]

Compared to just not thinking too complex and having a second queue instead which does allow it? 🤔

[quick token]

(not sure if im talking about the right part here, but the initial list of songs stays constant)

[vestal arch]

the shuffle option as i described before, but randomized 1 song in advance

[thin aurora]

Yes so you have no idea what comes after that single randomized song

[still jungle]

but you shuffle it

[thin aurora]

Also you'd still have to track the next song

[vestal arch]

yeah that's typically not really an issue

[thin aurora]

Might as well use a second queue for the list of songs

[vestal arch]

i mean... you wanted to, right? are we talking about the same thing?

i thought you were talking about showing the user what song is next?

[thin aurora]

Yes, but if you're going to do that for a single song then I don't see why a second queue is not the better option lol

[vestal arch]

less to move around

i still don't think queues make sense in the context of being able to shuffle

[thin aurora]

And in return you can only shuffle a single song beforehand with no way to extend it

[quick token]

i suddenly remember the second part of my problem

[vestal arch]

that's.. not an issue, though?

[still jungle]

isnt it better to have 3 arrays? one for all songs and 2 others (played songs cache, shuffled songs cache) with just the index where its the song stored?

[thin aurora]

The whole point was limiting yourself just to be able to use a single queue

I think it's pretty clear it's not a good idea

[trim schooner]

it helps to share the error...

[thin aurora]

Idk what PlaceAtFirst is

[vestal arch]

i think we have very different ideas of what the constraints are

[quick token]

error is just that it doesn't exist anymore. sorry i thought the context was in the script link i sent earlier

[vestal arch]

i don't see how the problems you mentioned are problems

i just think a single list is a simpler design

not a better design, just a simpler one

[thin aurora]

By using a List you're unable to get the original order of songs after shuffling

Idk how I can be more clear

[trim schooner]

it may well have been, but no one is actively going to remember those details as soon as moving on!

[quick token]

yeah my bad lol

[vestal arch]

if you shuffle the list directly, yes
but do you need to get the original order? that's just not a design requirement in my mind

[still jungle]

just shuffle the cached randomly pulled indexes

[thin aurora]

Taking Spotify and Youtube as an example which allow it, it's a very user friendly approach to be able to unshuffle a play list

[vestal arch]

if you want to preserve the original order, just randomize next instead of shuffling

[thin aurora]

Imagine wanting a song that has now disappeared because you shuffled

[still jungle]

dont event need to shuffle when its randomly pulled each time

[vestal arch]

youtube doesn't have unshuffle

youtube doesn't really shuffle the list, it goes to a random entry when the current entry is done

[thin aurora]

yes it does, it just doesn't show the next song in the shuffled list

Exactly

But you can still turn it off

That's my point

[vestal arch]

it doesn't have a shuffled list

[thin aurora]

Okkkkkkkk then just look at Spotify

Point is, there's a very good reason to not make it so compact because you missed on a possibly crucial feature

And that's why I pointed out to not do that

You just can't do these things with just a single List of entries

[vestal arch]

nvm it does it just doesn't show it in the ui

[vestal arch]

possibly crucial feature
i don't see how they're crucial yet

[thin aurora]

I thought that was what you meant, it does have shuffling in general just not the UI

[vestal arch]

yeah i thought it didn't but just checked and it appears to

anyways; we're talking about different things here

you're talking about a general solution

im talking about a simple solution, which just covers very basic design requirements
i know it doesn't have those extra features, but i don't see how they're necessary at this point
like, we're just talking about a basic system right? were there advanced design requirements discussed earlier that i missed or something 🫠

[thin aurora]

Their requirements were related to a bug and they didn't need to have more compact code as a solution

At this point this is just a matter of the right approach and I doubt Jordan even cares 😄

[vestal arch]

that's, not really what im talking about

[thin aurora]

I don't really think it matters anyway because we're both talking to a brick wall that thinks they know better

[quick token]

im still lurking, i got a little confused a while ago though ._.

[vestal arch]

my point is; if you don't need the advanced features, a simpler system is easier to maintain

[thin aurora]

An extra queue if very complex indeed

You know if we talk complexity then the queue fits in better over refactoring to use a list anyway

[vestal arch]

that's also not what im talking about, but ok 😮‍💨

[thin aurora]

And their initial problem was solved by checking for an empty queue and making sure songs are added, not refactoring to use a list 😄

[vestal arch]

im not talking about the problem, either...

[quick token]

tried making it a single array and now i can only hear coconut mall

[thin aurora]

Probably forgetting to update the index

A queue automatically plays the next one, but a collection needs its index updated

Otherwise share code @quick token

[ocean mirage]

Hi, im fairly new here, but using unity for a while.
Does someone know how i can relate an item with an inventory?
I feel like, i cannot use the grid layout, because its only for single slot items.
I already found very useful interfaces like IBeginDragHandler, IDragHandler, IEndDragHandler.
The tutorials on utube arent helpful for me.

[thick terrace]

is this for a game with an "inventory tetris" style system?

[quick token]

https://paste.ofcode.org/38RC9bMxSFmFRt2uy7kRxun (== on line 29 is meant to be !=, just fixed that but that didn't fix the unending coconut mall )

[ocean mirage]

i think so, its for an action rpg like diablo.

[quick token]

its not currently shuffling, index seems to change randomly (when i play next or previous)

[vestal arch]

is the issue with shuffle on or off?

also next/previous should be wrapped on the 0/array length, or requeue, so you don't index out of bounds

[quick token]

o yeah, just fixed that lol

[quick token]

doesnt seem to be the issue

i think the issue is around line 29

[thick terrace]

i think you could kind of work around it with the builtin grid layout since children (eg the item graphic) can be larger than the layout size of their parent, so for example that hat could be laid out in the top-left square and have an image inside which overlaps the adjacent squares... but it might be less confusing to implement something custom

[vestal arch]

(btw as you have it it'll start on the 2nd song instead of the first)

[quick token]

h u h

that is what's happening

i have no idea why

did i miss something obvious?

[vestal arch]

you increment before assigning the clip

[quick token]

OH

it was something obvious

[vestal arch]

currentSongIndex starts at 0, you do playnext, it goes to 1, then it starts

[vestal arch]

no im asking what setup you have where you're having the issue

shuffle/loop/loopsingle, on or off

[quick token]

need to clean up the inspector a bit

[vestal arch]

ok so it's just requeuing endlessly

currentSongIndex == musicQueue.Length is false
loop is true

probably gotta clean up the logic there

if loop and loopsingle are false, does it just stop when it gets to the end?

(as in, is that what it should do?)

[quick token]

if both are false it just plays the next song (which will requeue if the next song is the last song)

[quick token]

yeah this definitely needs some clean up

[vestal arch]

i mean the end of the list

like for example in that screenshot setup; if loop is false, and price p5 ends, what should happen

[quick token]

should just requeue and then stop. requeueing doesn't cause the audiosource to play (i think, maybe i changed something and forgot)

i think i've almost fixed it up now

[vestal arch]

if musicSource.time >= musicSource.clip?.length && !loopSingle:
  if currentSongIndex < musicQueue.Length || currentSongIndex == musicQueue.Length && loop:
    PlayNext()
```and then have PlayNext handle requeuing & loopSingle?

[quick token]

yeah sure i can add that. forgot that loopsingle doesn't even need to be in update

[swift falcon]

I have arrived to a problem which is oddly similar to the circle vs ellipse problem, where B is A but B must not inherit some of the A method. What is the solution to the circle-ellipse problem?

[leaden ice]

Decompose A into C and D. Give B a C but not a D.

Instead of:

class A {}
class B : A {}```
Make:
```cs
class A {
  C c;
  D d;
}

class B {
  C c;
}```

[swift falcon]

B have the same method as A, which is X(). However X() is works differently than the one in A due to constraint. How do i decompose the method?

[leaden ice]

Use interfaces

[swift falcon]

That was my exact problem

[ocean mirage]

i just got an idea out of that. thank you.

[leaden ice]

class A : IHasC, IHasD {
  C c;
  D d;
}

class B : IHasC {
  C c;
}```

[swift falcon]

But then B is not A if i do that

Is it possible to decompose the method while keeping the "B is A"?

[leaden ice]

Why does B need to be A

Can't B and A just be IHasC?

[swift falcon]

Because by definition it is

[leaden ice]

B is not an A if it doesn't have all the parts of A

you're presenting an impossible conflict of statements.

To use A and B in a common place you can use the interface IHasC

This conversation might be easier if we use concrete terms and concepts

what's the actual use case here?

[swift falcon]

Hmm thanks, that does make sense. Maybe i need to read more about the circle is ellipse problem

[still jungle]

https://paste.ofcode.org/r24MQayTLjTMdeDjmdPhCj

[quick token]

wow, you rewrote the whole thing

i was almost done debugging too lol

i see where i made a lot of errors now ._.

[vestal arch]

please don't spoonfeed

[quick token]

https://paste.ofcode.org/36enCLJaidgaXb2x29smWAa < here's my version

[vestal arch]

musicSource.time = 0;
this doesn't make sense to be inside the if

[quick token]

oh yeah, i was just trying to copy spotify there

oh wait no good point

i'll have to add another check for that one or something

or maybe a bool

[vestal arch]

maybe check if the current time is less than some threshold

[quick token]

yeah, i'll probably add that as a private field to be set in the inspector

also put the whole script (and an example scene) into a package, if you want to give it a try

i've been having (the older version) play in the background of a project im working on. its pretty fun to have some (totally legally acquired) music playing while i play

*may have added a couple too many songs though *

[quick token]

added it 👍

[still jungle]

he already tryed for long time, and keep trying, so thats just a form of inspiration at this point

[quick token]

i do appreciate the script to help too. i dont actually have a formal education in computer science or anything. having a script to look at is definitely more useful than reading the docs for another week lol

• i really need to learn proper naming conventions. my variable names are a mess

[timid comet]

hello im trying to do local multiplayer but im kind confused on why the inputs dont work im using new player input and ive placed the right binds but it wont move the character
if im usinging the normal script it does work but then i cant do the local multi binds

the movment code
https://pastebin.com/EteW434x

[thick terrace]

naming is an art not a science, conventions can only help so much... welcome to the hardest part of programming 😛

[still jungle]

the worst part come when variables are named correctly but are too long and mess up the readibility

[vestal arch]

https://javascript.info/ninja-code

[rigid island]

Only the one with truly good intuition will be able to understand such names. Try to shorten everything. Only a worthy person should be able to uphold the development of your code.
lmao wat?

goofy satire though, love it

[quick token]

i skimmed through and didn't even realise it was satire at first lol

probably should've noticed after seeing "underscores for fun"

[rigid island]

literally the first block

Irony detected
Many try to follow ninja paths. Few succeed.

[quick token]

idk, sounds professional to me

[rigid island]

One-letter variables I would fire you immediately lol

[quick token]

im not gonna stop using one-letter variables until i get through the whole alphabet (i only t lol)

[still jungle]

i got called out by underscores 😆, i use them for private members of a class, so i know instantly its private/protected member of a class, so its easy to know where to look for it when used inside a method

[rigid island]

or just use a proper IDE

[cold parrot]

makes sense to separate serialized from non-serialized private members (if you have conventions about them being handled differently)

[rigid island]

code with _ looks disgusting and its very unecessary imo

[still jungle]

m_ its better?

[rigid island]

just as shit

camel case it, you already know its a private var..

Pascal Case, its a property or public..

[still jungle]

_ are cool for private members

[cold parrot]

if you have conventions associated with a member it makes sense to make it obvious when a that convention should hold if the IDE doesn't help. so the underscore is by no means unneccessary

[rigid island]

go for it if it helps you, I just find it noisy

[still jungle]

btw i have proper IDE, its nvim 😆

[rigid island]

on a serious note :
https://unity.com/how-to/naming-and-code-style-tips-c-scripting-unity
https://web.archive.org/web/20240518091705/https://unity.com/resources/create-code-c-sharp-style-guide-e-book?ungated=true

[chilly surge]

Without _privateMember, its style would conflict with localVariable. Mutating a private member has very different consequences compared to mutating a local variable, distinguishing them is imo helpful.

[rigid island]

I use locals less often, i use _ for local if i must

[still jungle]

for me properties are noisy, 2 methods with backing field in one, with naming convention same as methods

[rigid island]

or just this. for members with same local name

[rigid island]

properties are just fancy methods, they can get messy too

I would not put more than a few lines in a prop though

[cold parrot]

the main use of style guides is to make everyone shut up, and you have truly arrived in your career when you finally get to mandate silence over it and stomp on everyone who objects. Then make people work with what they got or quit.

[rigid island]

ya as long as the team agrees to the same style, it doesn't matter. A good dev can use whatever the team wants to use for that project

[vestal arch]

no one-letter variables? got it

class M : UnityEngine.MonoBehaviour {
  void Update() {
    this.gameObject.GetComponent<UnityEngine.Rigidbody2D>().velocity = UnityEngine.Physics2D.Raycast(this.gameObject.GetComponent<UnityEngine.BoxCollider2D>().bounds.center, new UnityEngine.Vector2(0, -1), this.gameObject.GetComponent<UnityEngine.BoxCollider2D>().bounds.extents[1] * 1.2, 1) ? new UnityEngine.Vector2(UnityEngine.Input.GetAxisRaw("Horizontal") / 1.8f, this.gameObject.GetComponent<UnityEngine.Rigidbody2D>().velocity[1]) : new UnityEngine.Vector2(UnityEngine.Input.GetAxisRaw("Horizontal"), UnityEngine.Input.GetKeyDown(UnityEngine.KeyCode.Space) ? 10f : this.gameObject.GetComponent<UnityEngine.Rigidbody2D>().velocity[1])
  }
}

i spent way too long doing that

[rigid island]

jesus lol

[cold parrot]

ive had to use the most inane styles at work, and i would never agree to their sanity, but the silence regarding style questions is really liberating

[cold parrot]

there is no need for anyone to agree, thats the point

[still jungle]

ninja order

[hardy pasture]

Is this the type of code you write?

https://cdn.discordapp.com/emojis/1223322441540177970.png?size=48

[still jungle]

he is back

[rigid island]

ya its like fighting over how to style your shoe lacing lol

[leaden ice]

Having a code formatter among the static analysis checks your CI/CD system does is ❤️

[chilly surge]

Speaking of formatter, I can recommend CSharpier if you find dotnet format to not be strong enough for you.

[hardy pasture]

@vestal arch Are you back home yet? I really wanna see that juicy top tier code that you were talking about

[hardy pasture]

Yeah, they all ran away when were challenged just a little. But I had to go work. And I still do.
Like Anikki and that guy went back to chatting here after being challenged.

[steady bobcat]

getting flashbacks to es lint dumbness such as 80 char line limit 😐

[still jungle]

i gave you the response why you cant manipulate members of struct of UnityEngine.Object, so we done with that topic i guess

[vestal arch]

colab also has that dotted line at 80 chars by default
guess it's an old standard lol

[hardy pasture]

I understand why I can't do it, it's a property with a backing field. My question on what's the solution to this without duplication, and while the solution with just .Translate(0, 10, 0) works, the reason for that convo was because I presented the closest to the original transform.position.y += CharacterHeight, which is:

public static class TransformExtensionMethods {
    public static void TranslateX(this Transform Transform, float X) {
        Transform.position += new Vector3(X, 0, 0);
    }

    public static void TranslateY(this Transform Transform, float Y) {
        Transform.position += new Vector3(0, Y, 0);
    }

    public static void TranslateZ(this Transform Transform, float Z) {
        Transform.position += new Vector3(0, 0, Z);
    }
}

But then got many personal attacks and talked down upon instead of arguments after I challenged their worse(in my opinion) solutions.

[steady bobcat]

ages ago i was working on a JS/TS project and for some reason it had auto es lint applying configured for when you committed (git hooks). I hated it i removed it.
it would do lots of dumb stuff that just annoyed me such as that stupid default line length.

[rigid island]

pretty much no reason to even do this lol

[hardy pasture]

But yeah, that's usually happens when people are actually challenged: they just leave and ignore. Everyone can do personal attacks and talk down to other people, but most just run away when challenged to see what they got.

[chilly surge]

Because people don't care. I as well, left the thread almost right away, but joined in again only because the thread ended up being borderline people attacking each other, and I felt bad that you had to fend for yourself, so I decided to at least offer something so you wouldn't walk away from this whole drama with nothing.
I see that what you really care about is just one upping other people.

[hardy pasture]

It's shorter, more expressive, less error-prone.

[vestal arch]

"run away" buddy, i don't care. and neither should you
let it go

[rigid island]

I'm not seeing it, but if you do then more power to you..

[hardy pasture]

No, that's not what I care about. See how you feel when you get trash talked for no good reason.

[quick token]

oh this guy is still going, i blocked them a while ago

[hardy pasture]

I literally got talked down and trash talked instead of arguments.

[chilly surge]

Well, you are certainly still caring about it, so I'll leave the mods to it.

[hardy pasture]

And now I'm getting accused of one upping

And being the one to start this

[hardy pasture]

because the thread ended up being borderline people attacking each other
They started it. Don't give me "I care about one upping people" routine when I presented arguments and got personal attacks in return.
And then they just ran away after I challenged them to show the code they write.

#archived-code-general message

But yeah gtg now, I don't want to get dragged into this again

[still jungle]

the backing field is on native side, the value have to change so the shallow manager of c# side notify serialization system to write it to native side. Hanging member of value doesnt change the value itself so serialization system doesnt catch it. I know you know this but other reads and this is some good knowledge to know.

[vagrant blade]

Then don't, when you return. If you want to continue debating these things, use the thread provided earlier.

[reef elk]

Hey. Im new to Unity and was just wondering if any of you guys know how to transfer gameObjects between scenes with all the values being there still. Im making a turn based RPG and have a battle controller object in a battle scene that has references to all the player objects in the scene. Im trying to make a rest point in the dungeon scene that regenerates all the players MP and HP (I used playerprefs for this). When I load into a random encounter and use the DoNotDestroy method, it nulls a lot of the values in the battleController object that was transfered from the previous battle scene (not all the values from the battle scene battle controller are intact such as the player references). The screenshot I provided is the battleController that is passed from the doNotDestory method: Battle Controller Class: ``` public static BattleController Instance;

 private void Awake()
{
    if (Instance == null)
    {
        Instance = this;
        DontDestroyOnLoad(gameObject); // Keep it persistent across scenes
    }
    else
    {
        Destroy(gameObject);
    }
}```

`

[leaden ice]

it nulls a lot of the values in the battleController object that was transfered from the previous battle scene
Nothing is changed on a DDOL object or its components when a scene is loaded.

The reason those references appear to be null now is because they are holding references to objects that were destroyed when another scene unloaded.

how to transfer gameObjects between scenes with all the values being there still
In other words, this happens by default. The values do not change.

[wintry crescent]

Is there a way to dictate the order of rendering of UI objects, that does NOT include reordering the hierarchy?

[leaden ice]

Put them on different canvases with different depths

and/or rendered on different cameras in a camera stacking setup

[wintry crescent]

that sounds like the way to go for me, thanks!

[leaden ice]

Or consider switching to UI Toolkit which has its own depth system and doesn't use GOs at all.

[wintry crescent]

Unfortunately I'm making a mobile game that is pretty much 100% UI. UI Toolkit, from what I can tell, doesn't have nearly the flexibility I'd need :/

is it still the case that you can't even use custom shaders in ui toolkit?

[leaden ice]

I'm not exactly an expert in it, but I'm pretty sure you can. #📲┃ui-ux would be the best place for these questions

[wintry crescent]

probably #🧰┃ui-toolkit instead, but yeah

[leaden ice]

oh sorry yes

[wintry crescent]

sadly I don't think it's an option for me, at least not on this project

I guess I really should learn it though... eventually

[reef elk]

Im a little confused. Would I have to transfer the player objects as well? Cause rn it breaks the game cause the values are null. What could I do to tranfer those missing values? When it loads the first encounter, its fine. The values are all there. But when I get into another one, it nulls a lot of the values such as player

[leaden ice]

Based on your original screenshot roughly 90% of the job of this object is to just hold onto references to other objects in the scene. Therefore it makes little sense to make this thing DDOL in the first place. What's the motivation behind making it DDOL?

The other thing is it looks like this script is doing way too much. Break up the parts that store data that needs to persist and make those DDOL. The part that's largely just holding references to other objects in the scene shouldn't be DDOL, make it a separate component that lives and dies with the scene.

Again it doesn't "null" anything

you are simply holding references to objects that have since been destroyed.

All the things that say "Missing" have been destroyed.

[reef elk]

Oh I see

[sacred python]

is there a way to blend two material at runtime
im trying to make a day night cycle with two skybox materials

using UnityEngine;
using System.Collections;

public class ChangeSkybox : MonoBehaviour
{
    public Material targetSkybox; 
    public float transitionDuration = 2.0f; 
    private bool isTransitioning = false;

   
    void OnTriggerEnter(Collider other)
    {
        
        if (other.CompareTag("Player"))
        {
           
            if (!isTransitioning && targetSkybox != null)
            {
                StartCoroutine(SmoothTransitionToSkybox(targetSkybox));
            }
        }
    }

    private IEnumerator SmoothTransitionToSkybox(Material newSkybox)
    {
        isTransitioning = true;

        Material initialSkybox = RenderSettings.skybox;
        if (initialSkybox == null)
            yield return null;
        
        Material blendMaterial = new Material(initialSkybox);
        RenderSettings.skybox = blendMaterial;

        float elapsedTime = 0f;
        while (elapsedTime < transitionDuration)
        {
            elapsedTime += Time.deltaTime;
            float blendFactor = Mathf.Clamp01(elapsedTime / transitionDuration);

            blendMaterial.Lerp(initialSkybox, newSkybox, blendFactor);
            RenderSettings.skybox = blendMaterial;
            DynamicGI.UpdateEnvironment();

            yield return null;
        }

        RenderSettings.skybox = newSkybox;
        DynamicGI.UpdateEnvironment();
    }
}

[leaden ice]

https://docs.unity3d.com/6000.0/Documentation/ScriptReference/Material.Lerp.html perhaps?

[hexed pecan]

They seem to be using that

[sacred python]

yeah the problem is that it start from the initial skybox from the scene, than to completely black over time, than straight to the target skybox

[leaden ice]

indeed - I didn't look at the code 😬

[dusk apex]

What's the wanted behavior?

[sacred python]

to go from the inital skybox to the target skybox in a smooth way

[dusk apex]

So simply without the completely blacken material?

[sacred python]

yes
i have no black material

[dusk apex]

Log the initial, blend and new material prior to blending.
Assuming the initial and new materials are correct (as you're able to acquire the targeted skybox) the blend material might have some data unexpected. Printing the properties wouldn't hurt.

[hardy pasture]

MonoBehavior still receives OnTriggerEnter2D message even after getting disabled:

public class MyComponent : MonoBehaviour {    
    void OnTriggerEnter2D(Collider2D Other) {
        Debug.Log("OnTriggerEnter");
    }
}
```This runs when I disable this component.
I can work around it by checking if it's enabled:
```C#
public class MyComponent : MonoBehaviour {    
    void OnTriggerEnter2D(Collider2D Other) {
        if(enabled)
          Debug.Log("OnTriggerEnter");
    }
}
```But I thought this is what this property is for: to not get the Unity messages anymore. Is it the default/preferred way to do it for your own components: to check the enabled variable?

[vestal arch]

https://docs.unity3d.com/6000.0/Documentation/ScriptReference/Behaviour-enabled.html
it's not all unity messages, it's specifically Update

(maybe other closely-related messages too? idk, the docs only say update.)

seems like only Update's docs mention "only called if enabled" as well, so i guess it is just that

oh and Start

[dusk apex]

Log cs Debug.Log("OnTriggerEnter"); Debug.Log($"{other.name} hit {name}"); Debug.Log($"Click to see {other.name} highlighted", other); Debug.Log($"Click to see {name} highlighted", other);

[wintry crescent]

Now, I admit that I'm kinda solving a problem before it appears, but is there a way to do a "async-instantiate" if I have a really large single prefab I want to load in?

I know I can just add a whole another scene, but it's really not the most ideal option in my situation

[wintry crescent]

the goal is to hide it underneath an animation

[leaden ice]

https://docs.unity3d.com/6000.0/Documentation/ScriptReference/MonoBehaviour.OnTriggerEnter2D.html

Note: Trigger events are only sent if one of the Colliders also has a Rigidbody2D attached. Trigger events are sent to disabled MonoBehaviours, to allow enabling Behaviours in response to collisions.

[wintry crescent]

hide the wait I mean

[wintry crescent]

There literally exists GameObject.InstantiateAsync what am I even asking...

[dusk apex]

Disable the collider if you're wanting to disable physics interaction with the object.

[hardy pasture]

Thanks, guys.
Also, on the forums people said, I can remove the component from the object.
Yeah, I just disabled the collider atm.

Also, someone mentioned this on the forums:

You could add an ‘isEnabled’ boolean which cancels the execution in TryInteract

[inner wasp]

    public class Chase : StateMachineNode
    {
        /// <summary>
        /// The target to chase.
        /// </summary>
        [DoNotSerialize]
        public ValueInput target;

        /// <summary>
        /// The <see cref="AIController"/> used as a blackboard for the AI.
        /// </summary>
        [DoNotSerialize]
        public ValueInput aiController;

        protected override void Definition()
        {
            base.Definition();

            target = ValueInput<Transform>("target");
            aiController = ValueInput<AIController>("aiController");
            targetCaught = ControlOutput("targetCaught");
            targetLost = ControlOutput("targetLost");
        }
    }

Alright, bear with me here because this ones a bit dumb but I've been having trouble finding an answer. I've decided I want to use VisualScripting for my AI state machine. I'm working on reimplementing the state machine I made for my AI onto classes that inherit from Unit so that I can use visual scripting to link them up. I can see I need to use a ValueInput class to get parameters into my class, I'm not sure how I'm supposed to actually get the value contained within the value input class though.

[vestal arch]

uh, i guess #763499475641172029

[inner wasp]

This kind of sits on the threshhold between the two since I'm writing an actual script rather than linking nodes together. Do you think it fits better there?

[wheat spruce]

!ide

[tawny elk]

💡 IDE Configuration

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

Select one:

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

[vestal arch]

i think it would, the folks there would likely also know about the underlying system more than us

[still jungle]

wonder if there is any good indie game made with visual scripting

[inner wasp]

Honestly, after digging through it there's not a lot of support for it.

I would highly reccomend against making an entire game with visual scripting.

[still jungle]

same

[robust dome]

funny thing is, you asked this in #archived-code-general

[inner wasp]

Yeah, I fussed over where to plop my question. This seemed like the best place, especially since the VisualScripting section was a link to another discord.

[heady iris]

Visual scripting seems most useful if you can hand your non-programmers a pile of custom nodes to perform common game interactions

But it also has to be something you can't just configure in the inspector

I guess it gives you a framework for basic "if this, then that" schemes that can include some complex control flow

[latent latch]

visual scripting is for tool programmers to develop for your team

I wouldn't waste your time on it if you're developing it for yourself

[heady iris]

yeah

[inner wasp]

In my case I want to be able to create a handful of state machine nodes, I have them implemented as a series of C# scripts with a monobehaviour driving the state machine and it works but it's a bit of a pain to set up a new state machine using the same nodes.

[quick token]

while we're on the topic
at what point does spamming unityEvents just become visual scripting ._.

i feel like i've been using unityEvents too much recently but i haven't found a good reason to not use them

i just recently discovered them and they've made setting up stuff way quicker, and made my code more flexible

maybe i just need to read up more on events in general

[cold parrot]

Immediately. Editor assigned events are not actual events in the sense of the observer pattern, they are method calls performed in a visual scripting layer on top of your c# code.

Their problem is primarily that they break easily when you refactor method signatures, they are not reviewable together with code and they cannot be validated by code.

other than that, they are very limiting in what parameters they can pass and you cannot easily see in the inspector what their meaning is, or when lost, what they should be.

they have their place in prototyping but should be removed in production (if you want to stay productive). They remain tolerable when used only inside the same gameobject’s components for simple, non-critical calls (artist usage).

[quick token]

i see, so far i've just been mainly using it within single objects (mainly prefabs). i'll definitely keep all of that in mind though. changing a single variable name would be pretty deadly for the project because of them ._.

[quick token]

that's also a big problem i've seen so far, i wish there were some kind of way to fix that

[cold parrot]

i look at it this way: the time to originally write a piece of code, is typically much less than the time you spend on fixing its bugs later on. This extra time grows when you use lazy patterns in the original code in the hope of saving you a few minutes. So try to eliminate the possibility for bugs right away, make code as simple as possible but not simpler: In scripting: turn only stuff into data that you actually need to be configurable, in library code: protect your code flow as much as you can from config errors, log warnings and errors liberally, be very pedantic to help yourself and others use the stuff correctly. Validate number ranges and value combinations. This will save your mountains of time. Unfortunately, all this is very difficult (impossible) when doing visual scripting.

A single bug discovered a month later can add days to what originally only took 10 minutes. At the very least it will add 30 minutes, more if you're in a team.

(apparently people have made entire games using visual scripting, i wonder how they did that though, must take a lot of willpower)

[quick token]

oh yeah i see what you mean. UnityEvents have been really useful for a prototype project im working on, but in a full game i can see them getting messy quickly.

as an example, i've "gun" script that uses 2 events for firing and reloading. for now its only doing effects and audio, but that list of things in the event is going to pile up quickly once i start adding animations, maybe more effects and all the other stuff that's going on

unity events are good for quickly doing that, but i should hardcode some of this in later, considering its only on the gun object anyway

[cold parrot]

i think overall there is little to gain from pushing the scripting responsibility to non-technical people and if you are a technical person you can only loose time (by visual scripting).

i think not even its approachability (because of discoverable API) argument holds much water anymore

[quick token]

yeah that's definitely true. i've always worked alone. but in the few cases i've worked with teams, just having programmer do the programming is a lot quicker than trying to teach the story writer how to lol

[cold parrot]

these are the ones you may even retain in production because the tech-art APIs in unity are so convoluted, so just as a way to unify how you deal with fx.

[quick token]

oh yeah, if i were working in a team i'd definitely leave those in

[cold parrot]

but i'm telling you, the things that always cause bugs late in the development are things misconfigured by designers & tech-art.

not their fault, the just dont have the tools to validate the stuff.

(invest in a tools team!!)

[steady bobcat]

In a project i work in we have lots of objects that need special logic and we use unity events a lot to do lots of them due to the amount of them.
I did make some editor scripting to help such as a button that finds references to a mono behaviour instance by unity events in a scene.

I usually prefer to not use them when i can avoid it, i will always sub to an event in code if i am able as its more reliable. But for that game its often best to avoid a lot of unique scripts for many small cases.

[cold parrot]

maybe if you really cannot see the abstraction in a system that may be a good use case, I would however look very hard into finding one before i'd accept this as a problem.

[quick token]

oh yeah, i've actually been trying to be a lot more mindful of that recently. i've been making a lot more scripts that are general purpose and can be re-used.

[cold parrot]

games are actually quite simple in their abstractions so i would be surprised if this ever really comes up (absent time constraints)

[cold parrot]

there is a danger in that too btw, you can easily end up with little components for super useful little things that never quite fit the problem

[steady bobcat]

The issue is we have like 30 environments and like 100+ items in each and often you need to configure say a specific item to unlock a thing in an env when its picked up but another item needs to go in a container and it is soo many combinations we often use unity events to configure the logic for these (e.g. pick up unity event calls the unlock function)

its a bit vague i know, id want to avoid it myself but we decided making a new script for each and every instance was not a good solution due to the amount of such cases.

[cold parrot]

understood, seems to me you have not discovered the type object pattern yet? https://gameprogrammingpatterns.com/type-object.html

in unity this can be implemented trivially as an empty scriptable object.

and multiple such instances can be aggregated into collections to declare various game design rules like A needs B and C, all with the same symbol that is entirely abstract from your code that runs it.

this way you can abstract most of your design symbols completely from code. you declare only the usual dependencies, like unlock trees, inventories, abilities etc. this frees you to code systems abstractly without requiring a type for everything in your design domain.

[steady bobcat]

With how items are created in this game im not sure this is applicable and most of the initial design was done before i went to work on it anyway.
Perhaps if we knew of such potential features earlier it could have been done better

[cold parrot]

fair

[steady bobcat]

there is much more that needs fixing in that project but i don't get to work on it as much anymore apart from some bug fixing every now and again.

[cold parrot]

well, best not rock the boat then

if you break it, you own it

[quick token]

reminds me i need to look into abstraction a lot more. i've been casually programming for a few years now but i've never actually looked into stuff like abstraction ._.

there's a lot of more complex things i feel like im missing. i've been fine with the basic stuff but i've slowly started to realise that more complex stuff makes programming a lot simpler

[cold parrot]

it goes in waves, you find ways to improve, overengineer them out of excitement, this leads to you recognizing their true nature, then you truly improve by seeing 'the truth' and apply it, then discover new areas of improvement and repeat the cycle.

[quick token]

the programmer cycle
i've got a script from a while ago that uses way too many delegates, from back when i first discovered them

[steady bobcat]

🔥 public event Action myCoolEvent; 🔥

[quick token]

also yeah looking at some of the stuff in my project, this should definitely just be it's own script. it's just a respawning target

[quick token]

i dont use events enough, they seem so useful but i never find a good usecase that isn't just the same as a UnityEvent

[heady iris]

if you don't need to configure it in the inspector, then you might as well use the event keyword

[quick token]

that's the problem, i always configure it in the inspector

[steady bobcat]

an event the "normal" version of unity event basically and ofc therefore faster. HOWEVER you have to set it to null on destroy if used in a mono to avoid GC leaks

[quick token]

oh, that's good to know

i've had a few warnings about leaks in the console but i've never had them in a big enough project to have to fix them

[steady bobcat]

The event keyword ensures you cannot = the delegate and it can only be called inside the declaring class instance
https://learn.microsoft.com/en-us/dotnet/csharp/language-reference/keywords/event

[heady iris]

i'm not sure how you'd get a leak from that: if you destroy the Unity object it's attached to and then lose all references to it, the object will get cleaned up normally

[quick token]

o, that actually sounds pretty useful. i like to try and keep the autocomplete clean so that could help a lot with that (there's definitely better uses but thats my main use lol)

[steady bobcat]

I may have it a bit wrong but its quite easy in unity to have the managed objects leak and not get GCd for a while (even if the native counterpart was deleted)

[heady iris]

well, sure, if you keep any references to the C# objects lying around

[cold parrot]

also "for a while" is by design

[heady iris]

notably, UnityEventBase has no code that would handle object destruction (and there's no way it could even know about object destruction, AFAICT)

[cold parrot]

unity objects typically only leak easily when you use them in (void) async contexts that dont get cleaned up properly

[steady bobcat]

I thought i read this a while ago but perhaps i miss understood (ofc if you destroy a mono and do not unsub a member from an event then that will cause a leak until the event is also able to gc)

[cold parrot]

thats not a leak

[heady iris]

yes, that's just objects not randomly ceasing to exist :p

[cold parrot]

a leak is defined as "the GC or program code does not have a reference to the memory"

[dusky oar]

hello friends I have a little problem that is bothering me I can't find the problem it's not going in the right direction. would anyone help me?

[heady iris]

I guess you could label "I forgor to get rid of the object" as something like a leak

but it's not a memory leak in the proper sense

[steady bobcat]

If class A has an event and class B subs a function to class A. Later we Destroy class b. Class b's instance is not gc'd due to the event sub not being removed from Class A.

[cold parrot]

i would agree that this happens much more easily

[heady iris]

well, yeah

I'd hope that would be the case

The more practical concern is that forgetting to unsubscribe can cause bogus behavior

like unity throwing a tantrum when class B's function tries to access transform

[steady bobcat]

Its ofc not the same as a leak in native code but its still something you want to avoid

[heady iris]

using an event member v.s a UnityEvent is also completely irrelevant here

[cold parrot]

yes but you avoid that by always unsubscribing on destroy

its trivial

[heady iris]

you'd have the same problem either way

[steady bobcat]

but its not guaranteed so its a possibility if you forget

[heady iris]

well, sure

[cold parrot]

this stuff is also easily detected in code reviews

[heady iris]

you could also call Application.Quit()

[cold parrot]

if you want guarantees, use rust

[steady bobcat]

ofc why didnt i think of that!

[heady iris]

The thing is that this isn't a problem with an event member vs. a UnityEvent member

[cold parrot]

i think c# strikes a very nice belance between freedom, performance and convenience

[heady iris]

Neither of those can save you from subscribing with the wrong things

[steady bobcat]

and im basically the only one working on projects im assigned too now so no more code reviews 😦

[heady iris]

You know, that does remind me

I'd like to be able to scan for events that are "suspicious" -- e.g. the number of subscribers is growing without bound

or there are more than zero subscribers after exiting play mode

I have no clue how I'd do that kind of incredibly invasive analysis, though

[cold parrot]

why do you write code that needs such checks... tsk tsk tsk

[heady iris]

I mostly don't (:

also, that wasn't me -- that was six month ago me

[cold parrot]

i remember that one

total idiot

[heady iris]

foreach (var thing in GiveMeEveryGodDamnDelegateObject()) {
  
}

[steady bobcat]

from now on i wont make any mistakes thanks to this chat 👍

[heady iris]

Something like that

[heady iris]

🌠

I like to write validators where I can

I automatically scan for missing references in most of my game's prefabs

[cold parrot]

can you use reflection in the OnExitPlaymode (or what its called) method?

[heady iris]

That's roughly what it'd be; I'm just not sure where I'd start..

I'll have to poke at it at some point

[cold parrot]

get all types, get all instances, do your foreach over them instances

but how do you get all instances

maybe from the GC?

[heady iris]

that's what I was wondering

you'd need to find every single object

[cold parrot]

you could source generate your way into a solution

make your own little runtime for additional checks

[heady iris]

alternatively, I just invoke a bunch of events after the game is supposed to be over

and I see if the console blows up

we call this an "integration test", I think

[cold parrot]

disintegration test

[swift falcon]

hi , I am making a 2d game and am planning to complete it and publish it on both pc and mobile

however before i start making levels, i wanna consider aspect ratios since pc/mobile have different screen sizes etc... this is important bec the camera doesnt folllow the character, it shows the entire level

can any1 help me regarding this matter? Idk what aspect ratio to choose, how to play around this entire concept

[steady bobcat]

Continuing on the topic of "leaked managed objects": https://docs.unity3d.com/Packages/com.unity.memoryprofiler@1.1/manual/managed-shell-objects.html

[late lion]

I've been thinking about making a Roslyn analyzer for detecting when you forget to unsubscribe and maybe an auto-suggestion for generating it.

[cold parrot]

so the rule of thumb could be "make sure to null references to large objects when done with them"

[steady bobcat]

Yea i guess, i didnt really think it was needed before reading this and doing a fair amount of memory profiling for a game as we see lots of graphic related crashes (mobile).

[heady iris]

It's important to understand how destroying an object means basically nothing in C# land

[cold parrot]

i've never had an issue with this kind of thing, cause i generally use a lot of manual cleanup/dispose-pattern in my code but its good to be more aware of it

[heady iris]

we're just telling Unity that the object no longer exists in the game world

that's a neat article though! i should go peek at that..

[cold parrot]

it would make sense to have a dispose pattern on all monobehaviours and possibly a Destroyed event

[steady bobcat]

We use addressables a lot so usually textures get unloaded automatically (if we release correctly) but sometimes when profiling ill see stuff still loaded and i have no idea why 🤷‍♂️

[cold parrot]

because something still has the reference (in a shell) and it doesn't get unloaded

[steady bobcat]

well duh but the issue is finding why

[heady iris]

things can also live for a while before the GC actually eats them

[steady bobcat]

Yea thats true, i just wanted to improve graphic memory usage however in this particular instance

[quick token]

this might be kind of obvious but, i've got a player character who should be able to equip 2 weapons in each hand

would it be better to instantiate the weapon into the hand when its equipped, or just have 2 disabled instances of the weapon in each hand and then enable them when they get used

i've only got 4 weapons so far and the inspector is already looking a bit messy ._.

nvm, i think i got it. i think its easier if i just make a new script for this

[compact spire]

Is there a problem with calling Vector3.lerp during update()? I'm moving/scaling/rotating meshes in a grid and the don't end up very aligned at the end.

[somber nacelle]

show relevant code

[leaden ice]

You can call any function you want in Update.

You are most likely not understanding the math behind Lerp and using it incorrectly though.

[rigid island]

tossing this out there just incase
https://unity.huh.how/lerp/wrong-lerp

[leaden ice]

To use Lerp "properly" you need to call it like:

current = Lerp(start, end, t)``` where `start` and `end`  never change throughout the entire course of the motion and `t` changes slowly from 0 to 1 over the desired duration.

Most likely you are doing something like:
```cs
current = Lerp(current, end, Time.deltaTime * something)```
which makes a nice smoothing effect but does not get you ***L***inear Int***ERP***olation.

[compact spire]

I'm using it the first way, t is .1f

buuut

[leaden ice]

t being .1f means you are NOT using it the first way

t cannot be a constant value in the first way

it must go from 0 to 1 over time.

[heady iris]

well, that's extra wrong :p

Both of these are wrong:

current = Mathf.Lerp(current, end, 0.1f);
current = Mathf.Lerp(current, end, Time.deltaTime * 0.1f);

The first one is more wrong

[leaden ice]

Basically that t stands for "what percentage through the motion should we be now?

[compact spire]

lol

huh, I find it funny that it mostly worked then

so t needs to be some value calculated between 0 and 1 using deltatime?

[leaden ice]

Again, lots of people use it the "wrong" way and it makes a nice smoothing effect but what it doesn't do is actual linear interpolation or get you all the way to the final value

[leaden ice]

it needs to be a timer basically, going from 0 to 1

e.g.

float t = 0;
void Update() {
  t += Time.deltaTime;
  float currentVal = Mathf.Lerp(startVal, endVal, t);
}```

[swift falcon]

help please!!!, i need to build and run my game but it says "failure" , " u cant build while comiling or editing" even tho am not editting or compiling

[leaden ice]

(Lerp automatically clamps it at 1)

[compact spire]

I need to set a max of t to 1 right if I'm doing it that way?

[leaden ice]

no ^

[compact spire]

oh

[leaden ice]

What you might find easier than restructuring everything is to switch to MoveTowards

[compact spire]

Does that work with scale/rotate?

[leaden ice]

current = Mathf.MoveTowards(current, target, Time.deltaTime * speed);```

[leaden ice]

the semantics of what you're doing don't matter

it's just moving numbers towards other numbers

[compact spire]

gotcha

[leaden ice]

it can be done with anything

For rotation though you probably want Quaternion.RotateTowards

[compact spire]

Great, thanks, that will save me a lot of time. I spent a while trying to snap the transforms to grid and for some reason that wasn't working either.

@swift falcon unless you need to save your scene, can you just restart Unity?

[swift falcon]

didnt work

i even restarted laptop

[compact spire]

So it's either busy compiling or importing something then? I've never ran into that one

[compact spire]

@leaden ice Would Vector3.SmoothDamp also work?

[leaden ice]

yes but less controllably. If you want a smoothing effect that's what it's for though.

[compact spire]

I'm just very quickly rotating/moving/scaling in a runtime editor, maybe it won't be very noticable

[azure frost]

Okay, I have returned to an initial problem I was having.

My rigidbody-based character controller works, more or less. But there's a problem my bro and I have been wanting to solve.

Basically:

Player can only apply rigidbody forces to reach speeds up to "7" to the character. But the environment can apply more than that.```
But the issue is that when an outside force applies the rigidbody force, the "speed limit" kicks in, slowing the object down unnaturally.

Here's the script, derived from Dave/GameDev's third person controller.

[thin aurora]

Please use a paste site for large code blocks

[tawny elk]

Posting code

📃 Large Code Blocks

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

📃 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.

[azure frost]

https://hastebin.skyra.pw/xesolanama.csharp
Okay, how's this?

[mossy snow]

calculate the player's new velocity in Jump and MovePlayer (instead of SpeedControl) before you apply forces and clamp them to prevent player input from exceeding your limit

[arctic sparrow]

Lemme write some pseudocode

Vector3 inputVelocity = curVelocity+wantedForce;  //Calculate the velocity you'd get after Force

Vector3 envVelocity = Vector3.ClampMagnitude(inputVelocity, maxPlayerSpeed); //Get the clamped speed

wantedForce -= envVelocity-inputVelocity; //Remove the difference from the wanted force

There's probably a much better way to go about this

Seek's use case is pretty specifically tied to Rigidbody physics, which I rarely use in my projects, so idk if this is right

[arctic sparrow]

Is my approach what you're looking for?

[mossy snow]

roughly yes, although his code uses force for player motion. I try to avoid modifying velocity directly personally

[arctic sparrow]

Actually, the velocity from the rigidbody is "curVelocity", which is never modified in this context

[azure frost]

Also, he and I are literally in the same room, and working on the same thing here.

[arctic sparrow]

Though there's another problem. What if you try going in the opposite direction?

[azure frost]

Wouldn't the force vectors add up and cancel out?

[mossy snow]

would what forces cancel out? Don't clamp force if the resulting player speed is smaller than your threshold. If player movement is opposing their current velocity, their new velocity will be lower than the old velocity and you needn't do anything special

[azure frost]

Experimenting. We'll see if it turns out right.

[trim schooner]

not a code issue, delete and ask in the correct channel - #🎥┃cinemachine

[swift falcon]

How do i open metadata in visual studio? I wanted to see the metadata for a class from dll. The way i do it right now is to type the class in code, hover over class name, click on the class name in the popup, and revert the code change.

It will open the metadata. Is there other way to do this without modifying the code?

[swift falcon]

honestly i thought it required code

but am figuring out stuff in it now

it feels cheating asf for havingg such thing not to be coded

[wheat spruce]

@vestal arch I managed to come up with a really nice solution to the shapes! It turned out that having a transform tool would overcomplicate things, for the moment, so instead I opted to go with extensions for the shapes. (plus I can always make the tool later on)

public static class XFormExtensions {

  public static void SetCentroidPosition(this IXForm xform, Vector2 position) {
    if(xform is IShape shape) {
      shape.Geometry.X = position.x - shape.GetSize().x / 2f;
      shape.Geometry.Y = position.y - shape.GetSize().y / 2f;
}}}


public static class ShapeExtensions {

  public static Vector2 GetCentroidPosition(this IShape shape) {
    float x = shape.Geometry.X + shape.Geometry.Width / 2f;
    float y = shape.Geometry.Y + shape.Geometry.Height / 2f;

    return new Vector2(x, y);
}}```

then its so much easier to work with the shapes myBoxShape.SetSize(boxSize);

took a little while to work out that I could write if(xform is IShape shape) to be able to access the shape from IXForm

anyway, the question is. What is that "xform is IShape" thing called?

[thin aurora]

Pattern matching?

[wheat spruce]

I dunno you tell me 😆

but yes, it seems like thats what it is

[thin aurora]

I dunno what your question is about

xform is IShape is a pattern match

[wheat spruce]

sweet

most of my usage of the word is has usually been as an equality check

[thin aurora]

It's a convenient way to find relations between types without doing work on the actual implementation

[wheat spruce]

or whatever its called if I wrote if(foo is null)

[thin aurora]

is null is something people swear on but I don't see the point

[wheat spruce]

when I came across "is" I saw it used in that way, and I made the assumption that it was an alternative way to write ==

made sense that x is y and x == y both sound like theyre doing the same thing

[thin aurora]

The way it works in Unity kind of sucks because Unity only has a very small part of what pattern matching as a whole is in .NET

Also the whole "the gameobject is null when destroyed" really ruins it

Or well, pattern matching is probably a lot bigger now .NET Standard 2.1 is a thing

But is doesn't use standard equality check, that's something to remember

It might look the same but depending on implementation of equality checks between types it can be an even bigger difference

[wheat spruce]

I just wish Unity used a newer C# version

theres a few quality of life features I got used to when I've worked outside the context of unity

like namespaces have to be enclosed with {} in the version unity uses

namespace G.Shapes; you cant write something like this

not like it's caused problems or anything, it just means anything I write is forced to always have an additional indent

nor can structs be written as

  public struct mytest
  {
    public float foo = 4f;

    public mytest() { }
  }```

[steady bobcat]

Hopefully in a few years they will replace shitty mono with the CoreCLR which will give us the new features and greatly improve the compiled IL (i've seen people show how bad mono can be with optimisations in unity).

[simple void]

You could just not indent tho

[wheat spruce]

VS is set up to do smart indenting, so I'd be fighting against what its configured to do

[heady iris]

file-scoped namespaces would be nice to have

[leaden ice]

Finally got to use those in my Roslyn code generator 🙌

[simple void]

Not very “smart” indenting, yeah but I get that would be very nice

[swift falcon]

why when i use string rather than actionAfterDelay()

the StopCouritine instantly works but with () and not string, it is buggy asf?

[leaden ice]

Because you aren't using StopCoroutine properly

[swift falcon]

may u pls explain

or send me a doc link

explaining

[leaden ice]

Improper usage:

StartCoroutine(MyCoroutine());
StopCoroutine(MyCoroutine());

Proper usage:

Coroutine c = StartCoroutine(MyCoroutine());
StopCoroutine(c);```

[heady iris]

https://docs.unity3d.com/6000.0/Documentation/ScriptReference/MonoBehaviour.StopCoroutine.html

[swift falcon]

i see ok

[swift falcon]

and if i wanna start couritine i just put c; in a line?

[leaden ice]

I'm not sure what you mean by " i just put c; in a line?"

My example shows how to start the coroutine.

You use StartCoroutine

[heady iris]

if you want to start a coroutine, you call StartCoroutine

nothing else will start a coroutine

Praetor's code stores the resulting Coroutine object in a local variable

It then passes that object to StopCoroutine.

[leaden ice]

Of course to call my example "proper usage" is maybe also silly, beacuse you wouldn't StopCoroutine on the very next line after you start it. You would do it sometime later.

[swift falcon]

if (couritineON)
{
    StopCoroutine("actionAfterDelay");
    couritineON = false;
}

StartCoroutine("actionAfterDelay");
couritineON = true; 

----```

this is what i have

[swift falcon]

this is kind of why i was confused with waht u said

bec mine starts with stop

then start

well this works so i wont change but i still wanna understand for future reasons

[leaden ice]

You would replace this with:

if (couritineON)
{
    StopCoroutine(myCoroutine);
    couritineON = false;
}

myCoroutine = StartCoroutine(actionAfterDelay());
couritineON = true; ``` if you want to switch to the other form.

[swift falcon]

i see!!

ok

[leaden ice]

Where you declare Coroutine myCoroutine; as a member variable outside the function.

[swift falcon]

and then i do mycouritine = startcouritine(actionafterdelay()))

inside the local

[leaden ice]

as in my example

[swift falcon]

tysm much apperciated!!

[thin aurora]

So in short, you get a reference of the Coroutine and you use that reference to stop it

Because StopCoroutine(MyCoroutine()); just calls a new reference, and immediately stops that one

Also StopCoroutine with a string sucks because I don't think there's any way to know which one it stops, in case you run multiple ones

[swift falcon]

yeh that is true

i had 1 couritine

i will use the one u guys taught me for long term reasons obviosuly

tysm

[thin aurora]

Yeah that is the best idea

Even if you never have multiple, there's no reason to use the string variant unless you start Coroutines dynamically

[swift falcon]

also if am not mistaken

string is also less optimized , performance wise

[heady iris]

That's the least of your concerns!

more importantly:

• You can't use the same coroutine method more than one reliably (which one gets stopped?)
• You might type the name wrongly...

[thin aurora]

Fixable by making it a constant

[heady iris]

nameof can help with that, at least

[thin aurora]

But yeah, doesn't really matter anyway

[heady iris]

But it's still going to be worse than just calling the method directly

since Unity will have to use reflection to find the method you want to run

[swift falcon]

😂

[thin aurora]

Strings can be incredibly performant

[wheat spruce]

oh ffs, was typing out a question then suddenly my PC became weirdly unresponsive, before showing a "your pc is low on memory"

[thin aurora]

Constants, Stringbuilder, Char Spans, you can remove all allocation to an extend

[wheat spruce]

seems like I accidentally have a memory leak in the code

[thin aurora]

Plain strings by itself are shitty though

[swift falcon]

😂 lool

[swift falcon]

ik stringbuilder a bit

from java

but not much

i def will eventually learn as i continue coding

[thin aurora]

Yeah so that's one way to build a string without constantly allocating new ones

Because strings are immutable, but do allocate

And Char Spans are nice because they don't allocate at all

However, this isn't a drop in replacement because they are not truly strings

[swift falcon]

subbed u on utube btw

[thin aurora]

It's just a way to mutate strings without again allocating extra strings

[thin aurora]

Which one? 🤔

[wheat spruce]

how strange, im leaving the project running and looking at task manager, no increase

[swift falcon]

roydefined

[thin aurora]

Oh, nice, thanks lol

[wheat spruce]

it definitely showed unity was consuming all memory

[swift falcon]

how old r u?

i am asking cause i am 18 and new to game dev

[thin aurora]

Mommy told me not to share that with strangers

Also probably not a good idea to use this channel/server for these things anyway

[swift falcon]

lmao fair

i was just curious bec u have a good game

[thin aurora]

But by all means feel free to join the Discord in my bio to chat

[swift falcon]

alrd did

[thin aurora]

Thank you very much 😎

[soft escarp]

why isnt this working? it worked fine with raycast but spherecast refuses to work(i have another problem because i used raycasts, this is why im changing to spherecast)
so, this is the code and how its implemented

private Vector3 GroundDir() {
        if (Physics.SphereCast(_gCheckPoint.position, .2f, Vector3.down, out _slopeHit, .3f, allowedLayers)) return Vector3.ProjectOnPlane(wishDir.normalized, _slopeHit.normal).normalized;
        return wishDir;
    }

[heady iris]

Spherecasts ignore colliders that they start inside of

[leaden ice]

Fen is faster than me today!
Yes your options are basically

• Start your spherecast higher up (e.g. inside the the player object, not touching the ground)
• use OverlapSphere instead

[soft escarp]

but overlapsphere has no raycasthit T-T

[leaden ice]

then do the other thing

[soft escarp]

it needs to be touching the ground ._.

[leaden ice]

no idea what you mean by that. It can check the ground, the cast just needs to START outside the ground

and then it can HIT the ground

[soft escarp]

what

[soft escarp]

its ground check but for slopes
and the groundcheck im using is based on triggers

[leaden ice]

you currentl;y have the spherecast on the left

you ened to switch to the one on the right

the circle cannot START inside the ground

it must start above the ground

that's all

do you understand what spherecast is doing?

[soft escarp]

no, this is why i used raycast before the spherecast

[leaden ice]

Imagine a sphere in the scene, and you throw it through the scene in a straight line and see what it hits

that's a spherecast

ist's just like a raycast but instead of a dimensionless point, it's a sphere

the sphere cannot already be touching the object when it starts its motion

it has to start outside the object, and hit it while it moves

Raycast works the same - the start point of the ray must be above the ground to detect the ground

[soft escarp]

so it just works like
you shoot a paintball and then check where it hit?

[leaden ice]

yes

[soft escarp]

it makes much more sense to me now

[wheat spruce]

my tilemap is designed to return the center position of all my tiles. pretty much the same as shown here. The plan is that I will give these points to my marching squares function so it can return the mesh where theres a quad for every tile.
what concerns me with the idea is that every other marching squares algorithm, uses the corner points rather than the centers.
I figure the option would be either
1: stick with my original plan, which might make it hard to try and find resources online to help as nobody does it like this
2: alter the way my tilemap outputs the points so it is more like a typical marching algorithm

so far im not really intending for the mesh that gets produced to resemble what youd typically see in marching cubes like below. for the moment I just want the grid to have basic quads

[warm badger]

how to implement mouse look for touch(android) in new input system that resembles getasix("Mouse x/ Y") from old system?

[robust dome]

catch the touch and put it into the world with ScreenToWorldPoint

[warm badger]

sry, i couldn't get it ...a bit of explanation would be great

[dusky oar]

Hi everyone I have a rotation problem but I can't fix it however the Axes look good if someone could help me

[leaden ice]

#854851968446365696

[soft escarp]

how do I actually use spherecasts?

[hexed pecan]

Did you check the doc?
https://docs.unity3d.com/6000.0/Documentation/ScriptReference/Physics.SphereCast.html

[soft escarp]

yes, i did

[hexed pecan]

Which part is confusing you

[soft escarp]

the part where it doesnt work

[hexed pecan]

Show what you tried

[soft escarp]

this one returns Vector3, i have one other that returns bool but its very similiar

    private Vector3 GroundDir() {
        if (Physics.Raycast(_gCheckPoint.position, Vector3.down, out _slopeHit, .3f)) return Vector3.ProjectOnPlane(wishDir.normalized, _slopeHit.normal).normalized;
        return wishDir;
    }

[hexed pecan]

We were talking about spherecast though

[soft escarp]

yes, the check is a spherecast

[hexed pecan]

I can't help you if you don't show an actual issue

[hexed pecan]

Then show that?

[soft escarp]

crap, wrong version

    private Vector3 GroundDir() {
        if (Physics.SphereCast(transform.position, .4f, Vector3.down, out _slopeHit, .54f)) return Vector3.ProjectOnPlane(wishDir.normalized, _slopeHit.normal).normalized;
        return wishDir;
    }

[hexed pecan]

Looks like you are projecting your move direction on to a slope

[soft escarp]

the spherecast checks for a angled surface and if its on one, it returns a Vector3 with a proper direction and if its not on angled direction it returns wishdir(input direction
and the check was partially inside of the plaer and partially inside of the ground
and the new one is from inside of the player to ground

[soft escarp]

yep

[jaunty sleet]

anybody know what is needed to set the source image of an image component? I have the image I want to set and I have set it's type to sprite but it still won't let me select it

[hexed pecan]

I still don't know what the issue is, but you can use the Physics Debugger to see what your spherecast is doing

[soft escarp]

i have opened it but i have absolutely no idea what it does

[winged tiger]

[ExecuteInEditMode]
public class NetworkObject : MonoBehaviour
{

    private string persistendID;

    
    public string ObjectID
    {
        get { return persistendID; }
    }


    private void Awake()
    {
        if(string.IsNullOrEmpty(persistendID))
        {
            persistendID = System.Guid.NewGuid().ToString();
#if UNITY_EDITOR
                UnityEditor.EditorUtility.SetDirty(this);
#endif
        }
    }
    
}

Hello! Why is the persistendID changing value every time the function get's called?

[maiden fractal]

what function?

[hexed pecan]

It lets you visualize physics queries like ray/spherecasts and more.

[soft escarp]

what is it even meant to do?

[winged tiger]

Awake()

[somber nacelle]

Show what you tried

[still jungle]

you missing [SerializeField]

[hexed pecan]

You aren't going to get far if you don't use the manual/docs/google