archived-code-advanced

[iron pagoda]

I discovered it twice. Keep forgetting about it.

[merry notch]

I got it working for characters 😁

[twilit chasm]

Is there a way to see if two agents are on the same navmesh? I wasn't able to find some kind of "navmesh ID" in the docs. Does something like this exist?

[zenith ginkgo]

https://docs.unity3d.com/ScriptReference/AI.NavMesh.html if it exists, it'll be in here somewhere

[shadow seal]

https://docs.unity3d.com/ScriptReference/AI.NavMeshAgent-navMeshOwner.html

I suppose this would be the same for both?

[midnight violet]

If you really only need one particle, you might use a single particlesystem for that or just a fake particle. But you can try to set some statements in the shadergraph too to only effect particles on a specific height or position

[vernal kindle]

mmmm i understand, i'll have to delete the comment because i noticed there's a vfx and particles channel

thanks for the help

[gleaming thistle]

ello, how can I detect via linecast if I'm hitting an object.. if I end up inside of the object? it seems to not count as hitting if the linecast winds up even a single pixel inside of the object.

imagine I cast a linecast downwards, and it detects that there is something there, and moves upwards.
but if it doesn't happen fast enough, then what can happen is it ends up partially inside of the object, and then it cannot detect a hit anymore

[regal lava]

Do you mean like your line that originates is drawn inside of the collider if you get too close such that it wont detect it?

In my game I have two stages for my linecast; one for which I attack when I get close enough to the collider, and another collider for where I deal damage which is a smaller collider in that.

[static patrol]

There's something that I don't quite understand here. I have a cubic Bezier curve system, which generates a procedural road along it.
However, when I rotate the object containing the control points, the Bezier curve does not match the rotation, and makes it impossible to place accurately in the world, where I have to do weird positions and rotations to make it align with other instances of the road.

[gleaming thistle]

@regal lava yes exactly.
so you're suggesting I make another collider essentially? in case it gets too close to the actual block/object?

I manually use position move the object around and was really hoping that I could make it go up ramps smoothly, but if I need a lot of colliders to do that that's chill ig

[regal lava]

Well, it seems you're drawing the linecast too close, so perhaps you can fix it by originating it further back if you continuously need the linecast to touch a collider.

[gleaming thistle]

ohh

so if the origin is further back

then it will still detect the collision

thanks for the advice mate

[regal lava]

aye

[static patrol]

The three scripts that I employ for this
https://hatebin.com/behfvrrzcq
https://hatebin.com/vthapsltgz
https://hatebin.com/bovfddbsfb

[heady bane]

Could I have a coroutine call an action<string> to return the value found in the coroutine?

Im trying to find a generic way of returning a value from a unity web request coroutine that I can then plug into several scripts that need to get information.

[heady bane]

Good idea? bad idea?

[mild laurel]

bad idea.

[somber tendon]

Do you mean something like a callback function? I'd say yes, why not

[heady bane]

yea a callback function

[heady bane]

why is it a bad idea?

[novel plinth]

anybody knows if Unity supports FastSpan already or not?

[regal olive]

sorry for repeating but its so strange for me and dont know if its intended :/

[fresh salmon]

Do not cross-post. This isn't an advanced code question.

[hollow wing]

is it possible to call Instantiate with an empty GameObject?

if I do Instantiate(new GameObject(), ..) i have it twice :)

[humble leaf]

You don't need to instantiate, just create the new GameObject.

[hollow wing]

I think there are cases where not calling instantiate confuses the physics

when you set position afterwards

therefore I consider it good practice to use Instantiate always when physics is involved

[fresh salmon]

You could have a prefab of an empty GameObject you drag into a field. I wish GameObject.CreatePrimitive supported empties, it doesn't

[hollow wing]

I think I'll be using a static new GameObject intead

I wish GameObject.Empty was a thing 😛

or Instantiate without a GameObject

[regal lava]

Couldnt you just make a naked gameobject prefab and use that?

[hollow wing]

I think static is the better approach here since I have one less dependency of my code

[somber swift]

can't you just use GameObject constructor?

[fresh salmon]

Probably, no idea

[somber swift]

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

it should work fine

[hollow wing]

hm I'm not sure maybe though the issues with physics I had before would have gone away if:

• I had the prefab GameObject set inactive
• I enabled the GameObject from Instantiate after setting the position

There was some issue where Physics tests wouldn't detect new GameObjects immediately

[sharp otter]

Hi guys got a problem with my photon 2 this is the code https://paste.myst.rs/xv38y2ko and this is the error i get when ever i try to create a gameCreateRoom failed. Client is on MasterServer (must be Master Server for matchmaking)but not ready for operations (State: PeerCreated). Wait for callback: OnJoinedLobby or OnConnectedToMaster.
UnityEngine.Debug:LogError (object)
Photon.Pun.PhotonNetwork:CreateRoom (string,Photon.Realtime.RoomOptions,Photon.Realtime.TypedLobby,string[]) (at Assets/Photon/PhotonUnityNetworking/Code/PhotonNetwork.cs:1782)
CreateAndJoinRooms:CreateRoom () (at Assets/Scripts/CreateAndJoinRooms.cs:15)
UnityEngine.EventSystems.EventSystem:Update () (at Library/PackageCache/com.unity.ugui@1.0.0/Runtime/EventSystem/EventSystem.cs:385) i realy need help

[midnight violet]

#archived-networking maybe better for this, and please domnt paste code in raw messages... #854851968446365696

[sharp otter]

okay

I fixed but thx

[patent wadi]

hello i have got a question?
can we add a file to an existing executable so as to customize the keys? attention i'm talking about an exe compile no code sources.

[timber flame]

Suppose a player with ads mode ability in shooting games.
I have a simple script AimDownSightAbility. It has a boolean field IsAds (getter,setter) and an event AdsChanged.
In different situations, we want the player to exit ads mode.
For example when reloading, if they are in ads mode, they should exit and then they can reload or when equipping/switching between weapons, picking items, consuming medkit, enter vehicles, player died, etc.
They are different business logic and procedures.
My question: how do you handle it? You will have one script, listen to these events (reload,consuming,etc.) and then set IsAds field of AdsAbility instance or you will write different scripts (listeners) for each?

[formal lichen]

I usually have a singleton eventmanager that listens for events and notifies listeners when one is invoked. This is not much different in amount of code to just having one event that is triggered with references but saves lots of time in that: 1. You don't need a reference to the specific object to invoke the event or listen for it. 2. Removes spaghettification of code and helps when scaling things up.

[timber flame]

So, do you mean something like below?

@formal lichen

public class AdsHandler:MonoBehaviour{
    private void Start(){
        _eventManager.Subscribe(EventType.ReloadStarted,OnReloadStarted);
        _eventManager.Subscribe(EventType.PlayerDead,OnDead);
        _eventManager.Subscribe(EventType.EquipmentChanged,OnEquipmentChanged);
    }
    privae void OnReloadStarted(){
       _adsAbility.SetAds(false);
    }
    //... others
}

It is my first approach. You just suggested global event manager instead of separate events, perfect.
By the way, I do not like globally event manger.

[formal lichen]

yes, like that

[lilac jacinth]

Hi everyone, I have developed a VR keyboard for my university project and I have a problem with my enter key. The player to access the rest of the map has to enter his name which I save to display later in the game. Can anyone help me please?

Here is a video of my problem, I would like my input text to be displayed in the sample text and saved using the enter button.I tried to store it in a player prefs but it didn't work

[formal lichen]

I'm sure there are other ways of doing it too 🙂

[formal lichen]

is that enter key a Button?

[lofty falcon]

Is async void okay for Start() because exceptions will be thrown on the main thread anyways?

[lilac jacinth]

no, it's a cube with input text and i take the content of textmeshpro to display it, like all key

[timber flame]

It hides dependencies.
Also, arguments are different. You have to use something like object and then cast or use a bit different approach like Mediator pattern. This pattern hides dependencies as well

I use zenject, therefore, injecting dependencies is really easy and straightforward

[formal lichen]

You can also add another parameter to your eventmanager that will pass an object to the listener. I've used dynamics in the past but that doesn't work on all platforms

That's kinda similar anyways to just doing object and then cast though like you said

[lilac jacinth]

me or gladiator ?

[undone coral]

you probably wanted to use Input System and QueueInputEvent

to solve this problem

then your VR keyboard could behave like the native keyboard

[undone coral]

no, i think that's too weird

use

void Start() {
  UniTask.Void(async () => { ... });
}

or wrap it to deal with exceptions. there's no documentation for the behavior of async void Start, and besides, Unity's default synchronization context is bad

[undone coral]

what is your objective?

[undone coral]

use local space instead of world space for your bezier curve rendering code

[lilac jacinth]

yeah but can we declare a storable variable in a playerpref by saying (if Key.text =="Enter") you store the result of the text input to the text meshpro?

[undone coral]

this is #archived-code-advanced i'm trying to give you a great answer here

creating a mock keyboard device is just two lines in input system

and queueing a keyboard event is self explanatory

then you can queue a keyboard event with the Enter key

and everything would work, specifically eventsystem-driven stuff like UGUI input fields

[raw path]

Hi there
how do you guys store/manage localization entries in your projects?
I'm using a simple .json file with key/value pairs per language but I find that a bit tedious to edit

[lofty falcon]

Unity's talk on async vs coroutines best practices shows examples of using async void Start() to start other async Tasks. Would you not reccomend this?

[undone coral]

No, use UniTask

Unity’s people don’t make games, but UniTask’s authors do

it works better

[undone coral]

The Localization package

[raw path]

yeah I've stumbled upon it just now myself 😅

[wintry turret]

I am currently following the Unity documentation for Graphics.DrawMeshInstancedIndirect. I would like to rotate each mesh individually in the Z axis. How can I manipulate the unity_ObjectToWorld or unity_WorldToObject to achieve this?

[sly grove]

you can either rotate the object in your compute shader or in the vertex shader

[wintry turret]

I'm working in the shader provided by Unity's example:

void setup()
        {
        #ifdef UNITY_PROCEDURAL_INSTANCING_ENABLED
            float4 data = positionBuffer[unity_InstanceID];
            //handle scaling
            unity_ObjectToWorld._11_21_31_41 = float4(data.w, 0, 0, 0);
            unity_ObjectToWorld._12_22_32_42 = float4(0, data.w, 0, 0);
            unity_ObjectToWorld._13_23_33_43 = float4(0, 0, data.w, 0);
            unity_ObjectToWorld._14_24_34_44 = float4(data.xy, 0, 1);
            //retrieve rotation from buffer
            float rot = radians(data.z);
            float c = cos(rot);
            float s = sin(rot);
            float4x4 rotateZMatrix = float4x4(
                c, -s, 0, 0,
                s, c, 0, 0,
                0, 0, 1, 0,
                0, 0, 0, 1);
            unity_WorldToObject = mul(inv(unity_ObjectToWorld), rotateZMatrix);
        #endif
        }```
No matter how I modify unity_ObjectToWorld, my meshes won't rotate in the Z axis. How should I do it?

[cedar ledge]

are you supposed to be using data.w everywhere?

[wintry turret]

data.w is the size that I pass in

data.z is the rotation angle in degrees

[cedar ledge]

what does it do, if not rotate on z axis?

[wintry turret]

It scales fine, but does not rotate the meshes

[sly grove]

multiply it by a rotation matrix

[cedar ledge]

well yeah

[wintry turret]

I did

[cedar ledge]

that's what unity_WorldToObject = mul(inv(unity_ObjectToWorld), rotateZMatrix); is

[sly grove]

oh sorry

missed that somehow

[cedar ledge]

does positionBuffer contain rotation information?

[wintry turret]

yes. positionBuffer is a float4 containing (xPos, yPos, zRot, size)

[cedar ledge]

have you confirmed that setup() is being called

[wintry turret]

Yes.

[cedar ledge]

so changing the xyz of data does nothing at all?

forgive me but can you explain the purpose of data.xy in unity_ObjectToWorld._14_24_34_44 = float4(data.xy, 0, 1);

[wintry turret]

That updates the position of each mesh

[cedar ledge]

i feel like that might have something to do with it

do x and y rotations work?

[wintry turret]

No rotations work. The lighting on objects changes, however.

So maybe the normals are being rotated without the mesh rotating?

Just a guess that's probably wrong

[cedar ledge]

show me your vertex shader

i was going to suggest what happens if you change that line to unity_ObjectToWorld._14_24_34_44 = float4(data.xyz, 1); but lets put that on hold for now

i think that ^ will add z translation not rotation

[wintry turret]

The render call looks like this: Graphics.DrawMeshInstancedIndirect(instanceMesh, subMeshIndex, instanceMaterial, new Bounds(Vector3.zero, new Vector3(100.0f, 100.0f, 100.0f)), argsBuffer);

[cedar ledge]

oh its a surface shader

[wintry turret]

I'm following the example here: https://docs.unity3d.com/ScriptReference/Graphics.DrawMeshInstancedIndirect.html

[cedar ledge]

did you change float4(data.xyz, 1); to float4(data.xy, 0, 1); yourself?

[wintry turret]

yes, as data.z is not the zpos

Rotating around the X gives this:

Around the Z gives this:

These are cube meshes

[gleaming thistle]

say we have object A here [plr] and object B [can pathfind around walls]
at the start, object B is at the same position as object A, however, it then pathfinds around walls etc.
now, suppose it takes a moment for object B to find the path.
as such, to make movement look smooth, object A is to smoothly move towards object B.

however, object A does not pathfind, so, as a potential solution, I've done something like so;

obj 1 follows 2, 2 follows 3, etc all the way to object B

object A follows obj 1

but despite my best efforts, replicating this actual behaviour seems impossible

something like this happens at every corner

imagining all those cubes as obj 1,2,3, etc

basically it follows the path just fine following object B, but anytime B needs to figure out a new path, there is a slight delay.

and because of this, the player will sometimes reach a point on his way to an end point, and stop briefly.

I want to make the player slower overall and never stop moving until the end

without running through the walls

[wintry turret]

MonoBehavior:

computeShader.SetBuffer(kernelIndex, "buffer", buffer);``` Computer Shader:
```int i = buffer[0];``` I have not filled the buffer, so `i` will be the default value. What is the value of `i` here?

[long ivy]

undefined; will vary by platform

[wintry turret]

yikes

[undone coral]

what's your objective? what are you trying to draw?

[undone coral]

you should use vfx graph

[undone coral]

what is the issue you're trying to work around?

i don't see how this is an issue for pathfinding. that kind of world is easy for navmesh, provided you choose the correct agent radius

[gleaming thistle]

I wrote the pathfinding, not using navmesh

I have a pathfinding script.
it generates partial paths
when you click, it generates a path towards that position.

it's grid based

[undone coral]

hmm

[gleaming thistle]

it works fine

[undone coral]

why aren't you using aron granberg a* then?

[gleaming thistle]

because it was too hard for me to wrap my head around lol

look I don't need advice on the pathing itself atm lol

it works fine, it's just when I generate a new path

theres some time in between

and I want to hide that by smoothly moving towards the object pathing

[undone coral]

you can generate partial paths. that's for you to implement

in A*, there's no guarantee that its path x_0 ... x_i will be part of the fully computed path x_0 ... x_destination

[gleaming thistle]

I'm not sure I'm wording myself well here.
it pathfinds. it avoids all the walls and gets to the end position

I'm just trying to make it smoother

[undone coral]

that's just a limitation of what it is. the right approach is to make pathfinding as fast as possible

you're saying there's time between asking for a path and getting the solved path?

[gleaming thistle]

yes, it's very little but it's visually displeasing

[undone coral]

and you'd like to somehow lerp towards an incomplete path?

i'm telling you that it's not possible to make that behave well in the general case with A*

[gleaming thistle]

no, I'm trying to lag behind the completed path a bit, basically.

[undone coral]

well are you computing a partial path at all?

a partial path could mean, "just in the direction of the destination"

think a bit more abstractly, if you're aspiring to not use aron granberg's A* or navmesh

otherwise you've made things harder for yourself, because you're attempting to replace an asset you don't understand

[gleaming thistle]

I'm not sure why you feel the need to just, idk, insult me. I'm asking for help smoothing stuff out, not for the actual pathfinding

I feel like I've been responded to a bit harshly without much reason.

[undone coral]

sorry, i'm getting to the core of the issue

in math terms

what you're trying to achieve.

[gleaming thistle]

ok, I'll try and be exact

[undone coral]

it sounds like you're saying, "i ask for a path. it takes some time to get a full path. i want th eobject to start moving as soon as i ask for a path"

is that correct?

as opposed to, waiting until you get the path

[gleaming thistle]

not exactly

sorry

[undone coral]

you want to smoothly move between points on a grid?

[gleaming thistle]

basically. but I can do that, to an extent already.
the problem is this:
I click somewhere
it generates a partial path towards that somewhere
I begin to follow that path.
at the end of the path, it generates another partial path.
Repeat until at the end point.

which works fine, but in between, when it generates the next partial path, there's a bit of delay.

before that path is generated

so to make the movement look smoother, I want to follow the pathfinding object with another object, a bit behind it

so that in the time it takes to generate the next path, I won't have stopped moving [or at least, not for as long]

[undone coral]

why are you doing it that way

why are there multiple partial paths

[gleaming thistle]

because the grid is around the player. the world is going to be huge

[undone coral]

why doesn't it just compute a path to the end point

how would you know which points are "in between"

you might choose wrong ones

[gleaming thistle]

it does compute a path to the end point. here lemme grab a short video without any of the trying to smooth stuff

[undone coral]

there's no guarantee that

Vector3[] ChoosePointsInbetweenHereAndPath()

and

Vector3[] AStarPathfindingPath

will share any points at all

so why are you generating partial paths

i don't understand what problem that solves

conceptually, you want to do something liek

AStarPathfindingPath()
 // tween between all the grid path
 // points
 .Select(gridPath =>
   DOTween.DOMove(gridPath, duration: 1f))
 // sequence them
 .Aggregate(new Sequence(),
   (seq, tween) => seq.Append(tween))
 // play the sequence
 .Play(); 

i don't think you want to/ need to generate partial paths at all

i don't see what problem it solves

are you saying because the player moves and changes the path?

that's why you are using partial paths?

"partial paths are my approach to solve the problem that the agents may run into the player, which is moving, if the player intersects the computed path"

[gleaming thistle]

the grid is around the player. the player of course aligns gridded.

I can't grid the whole world because it would take up too much resources

or lag

or something

[undone coral]

hmm

[gleaming thistle]

idek the technical terms here, but if I were to generate a grid and check for all the walls, it would either lag or take ages

if I did it for the entire worldsize

[undone coral]

okay well

what if i told you it's impossible to just ad-hoc a solution to that problem

you either use the solution from literature, or you will do something that won't work

do you believe me?

i don't mean to be disappointing

it's just really mathy

i hear what you are getting at

[gleaming thistle]

I'm tempted to believe that.

[undone coral]

i'm not sure that aron granberg A* even deals with it correctly

i can check for you

[gleaming thistle]

but I also believed that I wouldn't be able to make a pathfinder

so I've been hopeful

lol

[undone coral]

yeah it's still a good thing to have done 🙂

[gleaming thistle]

this is what I mean, visually.
in case I still wasn't clear enough.

but I think you got the gist already

[undone coral]

yeah

i get it now

looks good

[gleaming thistle]

it just feels clunky when it stops midway for a split second. even just a tiny bit of smoothing it out somehow would make it look so much better lol

of course, I could quite easily smoothly follow the pathfinder, but then it could clip through corners and stuff

[undone coral]

the best performing graph will be navmesh, and then turning the navmesh path into a grid using bresenham's algorithm

https://gist.github.com/Pyr3z/46884d67641094d6cf353358566db566

this is bresenham's

this turns the floating-point lines from navmesh into grid movement

[gleaming thistle]

||just wishing I didn't make a 600 line script for a click to move game for nothing||

[sand raven]

is there a way to increase the limit of the StackOverflowException? I know my recursion isnt infinite, its just really big

[tropic lake]

https://stackoverflow.com/questions/54584224/is-there-a-way-to-increase-the-stack-size-in-c

first answer

read the warning because if you're running into an overflow you need to reconsider what you're doing

[glacial wedge]

Hi, I want to know if there is a way in C# to use GetComponent<T> where T is customizable from inspector

[brittle pumice]

wondering if someone could help me with a custom package workflow question

[flint sage]

@brittle pumice Ask away

[brittle pumice]

@flint sage so i have a custom package

with samples

the package in on remote git

so now i want to update the example

with a unity asset

my first thought was to clone the package into a new project under assets folder

update with a new unity asset and push

but the new asset is not short path

so it Asset/Samples/SampleA/MyNewAsset.xxx

so what is the workflow for a package change like this?

[brittle pumice]

or am i way off in the whop whops

[austere jewel]

What is the actual complication?
You can pull a package locally, reference the location as a local package (just add the package.json via upm), and then update its contents.
If you want to make a sample in the package, you just follow this https://docs.unity3d.com/Manual/cus-samples.html
Have your stuff under a folder named Samples~, remove the ~ while working and add it back when you commit (or use another method like symlinks)

[brittle pumice]

ok maybe it was that i committed without ~

[mental rapids]

hello !
Does somebody know how to get the parameter of a unity event by script ?

[midnight violet]

You mean https://docs.unity3d.com/ScriptReference/Events.UnityEvent.html ?

Or better this, to give you an example: https://docs.unity3d.com/ScriptReference/Events.UnityEvent_1.html

[mental rapids]

no, i don't want to add event by script
I need to know the type of the unity event parameter, to know the parameter is a bool, and get his value (true or false)

i am comparing two unity event, here, i am just doing that :

but if the parameter of the uniity event is a bool, i'd like to know if its false or true

[midnight violet]

Oh, so you want a generic parameter

[mental rapids]

i want a generic parameter ?
Don't know what you mean by that :/

[midnight violet]

you want to listen to multiple unity events, without knowing, what params they have when get triggered, some have bool, some have int or stirng, right?

[mental rapids]

Nope ahah

i explain what i have now

i'va made a tool, wich will take all the unity event variables in the scene, every where in every script

it will compare all these event to one unity event

[midnight violet]

What does it compare? so you got an array of UnityEvents and you compare it to what object?

[mental rapids]

i compare it to an other unity event

like this :

you see that i try to find all the unity event like this one :

i find two object where there are this event, but i don't compare the parameter, event if the parameter, here the bool, is true or false, it will indicate me that its the same event

[midnight violet]

Oh wow, okay. Not gonna ask why you do that 😉 but you want to do what now?

Like what is your script and where are you struggling

[mental rapids]

so here i have my unity event comparaison method

i check the target, and then the method name
and now i need to check the parameter

so i need to get the parameter of the function

[midnight violet]

https://docs.unity3d.com/ScriptReference/Events.UnityEventBase.GetValidMethodInfo.html ?

Oh no, you need the arguments, right

[mental rapids]

ouuuuh yeah its maybe that

okay, i don't find this functioon on my event ahah

[midnight violet]

What unity version are you on?

[mental rapids]

2020.3.16

[midnight violet]

Hm, should be available

I am on 2021.2.2f1

[mental rapids]

yeah, but here its just the method name

[mental rapids]

ooooh, i think i'm just stupid ahah

i think i can do something with that, thank yoou very much dude

[midnight violet]

You are very welcome 🙂

[mental rapids]

after some test, i can't really get the parameter of the function like
I spend my time on forums and i don't find any solution, if somebody have an other solution, tell me !

https://answers.unity.com/questions/1310341/how-to-get-parameters-of-unityevent-that-i-set-in.html
i've find this, wich work, but i need to knwo what is the type of the parameter, and i will not know it

[small badge]

I wouldn't expect the parameter of the function to be available; it's very much something that user code shouldn't need to worry about. I'm surprised there are APIs to get any information about the persistent subscriptions.
If this is in editor code only, though, a fallback option would be to serialise the event and read the information you want out of the serialised object.

[mental rapids]

i'd like to don't have it in the editor, but i can probably mdodify my code to have it only in editor

but i don't need this at runtime, if it was the "question"

[small badge]

Oh, actually that's the approach the link is saying, too. Well, then that's that; you're not going to get the type of the method out of the event because it doesn't store that information; you'd have to look it up yourself by looking up the method on the target object's type.

(And if there are multiple methods with the same name on the target object I guess you'd have to experiment to see which one UnityEvent calls? As far as I know that's not documented anywhere.)

[mental rapids]

i had an idea, but idk if it can be done
Like as parameter, we can have an int, float, string, cinemachinepath and a lots of other things
Do you think i can get the parameter as string ?
Like imagine the parameter is a GameObject (nammed "myGameObjectAsParameter")
do you think i can do somethink like mySerializeObject.getParameterAsString ?

and don't need to check if its an int, a bool or other ?

[small badge]

No. Even if it allowed you to get the properties as the wrong types, the different argument value properties all have different names.

Depending on what behaviour you actually want you could possibly get and compare all of the serialised argument properties, including the unused ones.

[mental rapids]

i've found that, but cant define what is the "method" variable

https://newbedev.com/how-can-i-get-the-values-of-the-parameters-of-a-calling-method

[violet fjord]

Hi, I'm struggling to understand how to connect to a database and wondering if anyone can help. I have a PHP script to check info for a login against my database but I'm not sure how to run it. I'm following a tutorial but it doesnt match with the database type I'm using

This is the line that fetches the PHP script but I'm not sure if this is the correct way to do it or not

[hollow garden]

me when the WWW www = new WWW

[violet fjord]

I'm just following the tutorial : (

And making changes where required.

[shadow seal]

Nowadays we use UnityWebRequest, it probably tells you WWW is obsolete

[violet fjord]

Yeahh, I saw that but I really am not sure how to go about this other than following the only tutorial I've found that has made some sense.

[sweet briar]

here is something I've used.

IEnumerator GetUsers(string URL)
    {
        using (UnityWebRequest www = UnityWebRequest.Get(URL + "/GetUsers.php"))
        {
            // Request and wait for the desired page.
            yield return www.SendWebRequest();

            string[] pages = URL.Split('/');
            int page = pages.Length - 1;

            switch (www.result)
            {
                case UnityWebRequest.Result.ConnectionError:
                case UnityWebRequest.Result.DataProcessingError:
                    Debug.LogError(pages[page] + ": Error: " + www.error);
                    break;
                case UnityWebRequest.Result.ProtocolError:
                    Debug.LogError(pages[page] + ": HTTP Error: " + www.error);
                    break;
                case UnityWebRequest.Result.Success:
                    Debug.Log(pages[page] + ":\nReceived: " + www.downloadHandler.text);
                    break;
            }
        }

[violet fjord]

maybe this is a dumb question but why does every tutorial or explanation involve a http address. I just want the script to be accessed through the project.

I have no idea what I'm talking about tbh

Did database stuff for my last uni module but crazily enough it taught me literally nothing because it was just copying things off a slide.

[compact ingot]

what kind of database is that?

is it your own? do people with your (game)client have to access it eventually?

[violet fjord]

For a login yeah

using mariaDB

[sweet briar]

In my example it's getting a PHP file, running that PHP file's logic, which picks information from an SQL database and returns it as json, which C# then serializes into a class

[compact ingot]

do you want to connect to the database directly (with all the security problems that has)?

[violet fjord]

I'm not sure I understand what you mean

I think maybe I should give up on this, its way above my level.

Any recommendations for courses or anything

[compact ingot]

usually you would build a server that accesses a database on its local/private network and processes all requests for data and offers responses to the internet

for security reasons people will be very reluctant to recommend making a database accessible from the internet directly (without such an intermediate server)

[sweet briar]

I'm pretty sure all they need is a sandbox database to play around with in their project. Don't really need to worry about where its stored yet

[violet fjord]

I mean, in the context of a simple uni project (all this really is), do you think its necessary to not have it be direct?

[compact ingot]

that depends on the database host

if your school allows public connections, then do it

to connect to a database directly from C# its easiest to use Dapper https://github.com/DapperLib/Dapper

[shadow seal]

I tried to access my own database from uni, boy did they block me

[sly grove]

Universities are like weirdly obsessed with traffic monitoring.
One time while working in a research lab I got called into someone's office because apparently my computer was making suspiciously high numbers of DNS calls.

It turned out this was because I liked to keep google news open and refresh it all the time, which made tons of DNS requests for all the various news organizations.

[sweet briar]

I'd just use my phone's hotspot since it's three times faster and not logged half as much. Take that, Uni.

[violet fjord]

Can't even do my database stuff offsite so I have to be on campus whenever I want to test it.

Because they won't allow us to connect when off uni wifi

[sweet briar]

What kind/amount of data are you trying to load Hevii?

[violet fjord]

literally just login details for a website/game

itd be shared

oh and a highscore from the game

and hardly any data, we're not gonna have to test loads.

[sweet briar]

So your "Scores.json" and correlating C# class shouldn't be too lengthy or require much 'fancy' json parsing

That playlist I shared should have you up and running for your project within the first 3-4 videos. If you still need help feel free to mention me, I'm no expert but I've been toying around with DB handling and loading for a bit and I need a challenge that isn't my own mess for once lol

[violet fjord]

Alright, thanks dude.

Think I'm done for tonight though because its been like 6 hours and its just making me depressed now.

[sweet briar]

Take a break and you'll have some new approaches for tomorrow. Cheers!

[pulsar isle]

I want a 3D Model to appear at a certain location using GPS when the user is near. Like I have the location of the user, the location I want the 3D model to be present, the distance between the two(approximately) at every given time and ar foundation. Any guidance on how to achieve this, will be deeply appreciated.

[midnight violet]

#🤯┃augmented-reality might be the right place to ask

[timber flame]

Why not

   private void Start(){
       Method().Forget(); // or add Exception in Forget param
   }
   async void Method(){
      await ...;
   }
   async UniTask Method(){
      await ...;
   }

[small badge]

Or even (as the UniTask documentation suggests) just async UniTaskVoid Start() { await ...; }.

[drifting galleon]

just do not use unitask

it's not true async

[small badge]

In Unity it is exactly as "true async" as System.Threading.Tasks is; they both run all tasks on the main thread unless the task specifically switches context to another thread.

[wind vine]

   private void Start() {
        GameObject PlayerPlane = GameObject.Find("Plane cessna");
        PlayerControllerX playerScript = PlayerPlane.GetComponent<PlayerControllerX>();       
    }  
    private void Update() {
        if(playerScript != null)
        {
            if(playerScript.playerHealth <= 0)
            {
                gameoverUi.SetActive(true); 
            } 
        }

playerScript does not exist in the current context

line 7

how can i fix it

pls

[small badge]

You should read up about the concept of variable scopes in C#, e.g. https://www.tutorialsteacher.com/articles/variable-scopes-in-csharp (no idea if this tutorial is any good, it was just the first result from google). If you still have questions about how they work after reading up on them I'm sure #💻┃code-beginner will be able to help you out.

[wind vine]

alright will do thanks a lot

[tight junco]

Hey all, I'm using a NavMesh right now to do pathing, but I really don't like how all my units have to accelerate/decelerate. How can I make it so that units start/stop moving instantly?

[wind vine]

so is my code supposed to be in method level scope ?

[tight junco]

Declare your variables out of the scope of the functions, and initialize them inside of Start()

when you type GameObject playerPlane for example, you create a new object that exists only within that scope

in this case, only the scope of Start()

[wind vine]

i see interesting

[tidal skiff]

I'm making a music-based game where creatures play a part of a song (so for example a vocal or a piano) and I am looking for a way to make them play at the right time. I have an idea in my head but I don’t know how to write the code for it. An object called "SongManager" would send a signal to the objects and tell them to play a certain animation at a certain time. I am not really capable of writing complex scripts, could anyone help me with this?

[fresh salmon]

You could have a list of timestamps, where each timestamp denotes when an animation will play. The list elements are sorted. And the script will basically check if the current song time is greater or equals to the first timestamp in the list. If it does, play the animation and remove the first element from the list, repeat.

[tidal skiff]

That sounds logical, how would I do that though? I am not really experienced 😅

[fresh salmon]

Paradoxal isn't it

You just posted in #archived-code-advanced

Anyway, learn about lists, how to access elements, remove elements, sort the elements. And a bit of class creation if you want to store more than just a timestamp in a list item

[tidal skiff]

Okay, thanks for the help!! ^^

[small badge]

Surely you would get better answers in a forum or discord that is in some way related to any of those topics, rather than in a discord for an unrelated framework that almost exclusively uses C#?

[rocky otter]

Oh sorry, wrong server

[undone coral]

this is alright, i just like the UniTask.Void because it's the most clear about what's really going on

[empty vortex]

hey guys, is it possible to access _ScreenSpaceShadowMap in deferred rendering? setting it to a global texture thru a command buffer on a light source works fine for forward rendering but i couldn't get it working or find any resources on doing the same with deferred rendering except recreating shadow mapping manually haha

[undone coral]

hmm, the whole point of deferred rendering is that you can mess with it no?

are you sure that's what it's called inf rame debugger?

[warm spade]

If I make a game full of easter eggs, is it possible to protect my assets (textures, sound files, models) from getting simply pulled by the players?

[sly grove]

not really

Not without dramatically increasing the complexity of your game/code, and even then it's not 100%. You basically can try to obfuscate/obscure it to the best of your ability but it'll never be foolproof

[dense aspen]

@sly grove I've always been curious, could a game be 'streamed' to completely hide its inner workings from the player?

[sly grove]

sure, launch it on Google Stadia only

[dense aspen]

such as with whatever the equivalent of stadia is

I thought it's defunct

[sly grove]

there are other services aren't there?

[dense aspen]

I'm waiting for the day one of those big corporations decide to make a proprietary game streaming service that their games will exclusively be played on

[sly grove]

That would be awful IMO

[regal lava]

too much money

[dense aspen]

no kidding

[regal lava]

if google couldnt get it to work it aint happening

[dense aspen]

at least not yet

[regal lava]

there was also something a decade ago that tried a game streaming service and it died just as quickly

[dense aspen]

Then the technology doesn't appear to be capable enough yet

[sly grove]

anyway you can certainly stream parts of your game. by which I mean have your game download assets only as they are needed, e.g. via Addressables or Asset Bundles

[dense aspen]

One of the better benefits of letting players snoop around is that they can mod it

I'm going completely off topic

I was thinking that there must be a way to put something into your game that is undetectable in the files

[sly grove]

You can obfuscate it

break the file into 3 different binary files and then combine them all in your code into a single file for example

Or encrypt it and hide the encryption key somewhere else in the code

or break the encryption key into multiple pieces and likewise hide them in multiple places

Ultimately clever people will solve all of these "puzzles"

[dense aspen]

at that point

is your easter egg worth locking up inside fort knox

[sly grove]

up to you

prrobably not

[warm spade]

The game is all about easter eggs and secret references to other games and movies, so it's basically the whole point of the game. If people can just pull out a program and extract all the assets, the game is ruined

[sly grove]

If players want to ruin games for themselves they will. If they don't want to, they won't.
For evidence of this: See the entirety of YouTube Lets Play content, the entire existence of GameFAQs/Game Walkthroughs, and the existence of cheat tools.

Why would someone "pull out a program" to find/extract assets from your game when they could just watch a YouTube video or read an article detailing the easter eggs?

[warm spade]

The thing is that the youtube videos are created out of players actually playing the games and finding out stuff. To my knowledge, it is extremely easy to pull out assets out of unity games. I want the players to slowly find out stuff about the game, I want it to be that kind of game that even after 10 years players still find out new stuff. What I don't want is someone just pulling everything in the first day of playing and pretty much ruin the 'quest' for everyone

The game will actually be a normal, horror survival story game. It's just that I want to make it special by including lots and lots of secrets. That's part of my plan on making the game popular. If people are able to just take everything in one touch, it's not worth the work.

[flint sage]

Pretty sure that progress bar runs on a different thread so it's not blocked

[fervent parcel]

nvm solved it, it wasn't processing anything, it was just a huge lag spike from GC

[empty vortex]

i mistyped it in the message here LOL but i could get it just fine in forward rendering, but it doesn't seem to work in deferred and according to a forum user the command buffer method only works in forward with no further explanation
https://forum.unity.com/threads/accessing-screenspace-shadowmap.737840/

[glacial wedge]

I want to encapsulate a class to another class to add some info.
but I wont to change the full code.
How to write my new class to give the direct access of it's property old class?

class A{}
class B{public A data; public string info;}
print(b); //gets the A data
print(b.info); // gets the string info

[flint sage]

Inheritance?

[glacial wedge]

without inheritance?

[stark sail]

Anyone here familiar with UniRX ?

[flint sage]

Well print calls tostring on an object, so you can just override tostring and return a tostring from a

[fresh salmon]

So a sort-of wrapper / proxy?
Here's a versatile class you can use with any type

public abstract class Wrapper<T> {
  public T Data { get; }

  public Wrapper(T data) => Data = data;
}

Inherit at will to add your own properties

[glacial wedge]

not enought, to access it I have to write .Data. my goal is to get a direct access.... i mean... return the property instead of the class as default

[fresh salmon]

You can't

C# language limitations

You can't do b.Whatever and expect it to implicitly access b.A.Whatever

Unless you create the properties yourself of course

Which could considerably fill the class with a lot of code

[small badge]

An implicit conversion operator would let you use b in contexts that expect an A, but it's still going to be a B, not an A.

[small badge]

If you have a question about something then it's generally best to just start by asking the question itself. If someone knows the answer, they'll answer.

[glacial wedge]

ok 😦 thanks anyway

[stark sail]

Well, firstly, a question regarding vs. I have installed a nugat package and the warning message doesn't appear in the error list

[fresh salmon]

Because it's an analyzer warning and not a compiler warning. It's internal to VS and doesn't get reported back to Unity

[stark sail]

I am not trying to see it in unity but find the code line that produced it. Based on that I should be able to find it

[fresh salmon]

Double-click the error in the list, or check the other columns, Source File and Line

[glacial wedge]

You should replace the ObservableWWW with UnityWebRequest instead of trying fix a deprecated library

[stark sail]

I do not think that the problem appears in this function since the message is received normally

Despite the fact that the library is obselete

[undone coral]

the source to URP and HDRP are open. you can look how to access the pass that way

[undone coral]

i wrote the observable web request that is correct
https://gist.github.com/doctorpangloss/b5677d2b925b5f07a8525760af25d818

[radiant wave]

I want to build a statue editor. You can select 3D pieces and attach them to another 3D piece. I'm using a Raycast to check if the player is currently looking at a valid place to position the new object. I can then use the Raycast Hit point as the spawning position for the new element. But this results in the object being halfway inside the other object (since the pivot point is in the objects center), see teh attached image. Since newly placed objects are fully rotate- and scaleable, I can't just move the pivot point. I need to determine the position of the 3D object where it would be just touching the other object. Is there a way for that? Basically "Set this 3D object as close as possible to the point"();

[cursive horizon]

there's https://docs.unity3d.com/ScriptReference/Physics.ComputePenetration.html

but you might be better of manually placing 'snap points' on each of your pieces and having those snap by distance

[radiant wave]

Basically I would use the raycast hit point, then iterate through positions on the axis from there back to the camera and check for those positions whether they overlap or not using ComputePenetration

[cursive horizon]

ummm doesn't sound right? iteration positions? I think you just spawn the thing inside of the other thing, then see how far inside it is (with compoute penetration) and push it back out

[radiant wave]

I was confused by the bool return value. I just saw the out direction, out distance variables!

I thought I needed to iterate different positions until it returns false lol.

[glacial wedge]

is public event Action<T> MyEvent; the same of public Action<T> MyEvent;?

no it isn't. with event I can't assign another Action<T> to MyEvent

[vague granite]

Anyone with Tnet3 Experience here?

[hushed fable]

@shadow urchin Job posts go on the forum.

[fallow lantern]

you can only do += and -= on fields with the event keyword. Think of it as properties for delegates

[mint sleet]

Hello

Is there any way to swap the AWS SAM Template by Jetbrains Rider's AWS SAM Template

they use different cloud formation files: Rider uses 'yaml' and VS uses 'JSON'. and could not get it worked.

So, I am going to create the SAM application with VS first, and open that .sln file by Rider and everything will work smoothly?

and AWS Toolkit for Rider will be able to work with...

Am I asking too much 🙂

[drifting galleon]

list of methods for your delegates as well

[mint sleet]

  public async Task<Restaurant[]> GetRestaurantsAsync()
        {
            ScanResponse result = await _dynamoDb.ScanAsync(new ScanRequest
            {
                TableName = "Restaurants"
            });

            List<Restaurant> restaurants = new List<Restaurant>();
            
            if (result != null && result.Items != null) {
                foreach (Dictionary<string, AttributeValue> item in result.Items) {
                    item.TryGetValue("Id", out var Id);
                    item.TryGetValue("OrgName", out var OrgName);
                    item.TryGetValue("OwnerId", out var OwnerId);
                    item.TryGetValue("RestaurantName", out var RestaurantName);
                    item.TryGetValue("Phone", out var Phone);
                    item.TryGetValue("Email", out var Email);
                    item.TryGetValue("RestaurantSlogan", out var RestaurantSlogan);
                    item.TryGetValue("Address", out var Address);
                    
                    restaurants.Add(new Restaurant
                    {
                        Id = Id?.S,
                        OrgName = OrgName?.S,
                        OwnerId = OwnerId?.S,
                        RestaurantName = RestaurantName?.S,
                        Phone = Phone?.S,
                        Email = Email?.S,
                        RestaurantSlogan = RestaurantSlogan?.S,
                    });
                }
            }
            return restaurants.ToArray();
        }```

Hey people! I want to get the "Address" value from dynamoDB.

Address is a "MAP"

that is why I can not assign it directly to a UnityObject as I have done to "restaurants" object.
Do you have any suggestion? thank you in advanced.

[empty vortex]

i use SRP for now:/ but yeah a migration to hdrp/urp seems perfectly logical, it would've been nice to support everything tho

[undone coral]

you're making stuff really hard for yourself

[mint sleet]

I know! But I had to, I think... Because I have some fields that I should not retrieve at all.

[undone coral]

what is a "MAP"

i feel like you already know the answer to this

you have to copy your blah into your blahblah

is this something you plan to distribute in google play / the app store?

[mint sleet]

I have tried but I got {}

DDBContext.ScanAsync<Restaurants>(null);

is there any way not to retrieve all the fields by using data persistence model in AWS

        public string RestaurantSlogan { get; set; }
        public Address Address { get; set; }
        public List<Catalog> RestaurantCatalogs { get; set; }```

[undone coral]

is this something you plan to distribute in google play / the app store?

or on steam?

or in any way publicly really?

[mint sleet]

for mobile

[undone coral]

so for google play or the apple app store?

[mint sleet]

yep

[undone coral]

you won't be able to access dynamodb this way then

[mint sleet]

I give user policy when they are authenticated

so they are good to go.

[undone coral]

are you using an access key or cognito?

[mint sleet]

cognito

        public async Task<APIGatewayProxyResponse> GetRestaurantsAsync(APIGatewayProxyRequest request, ILambdaContext context)
        {

            context.Logger.LogLine("Getting restaurants");

            var search = await this.DDBContext.ScanAsync.<Restaurants>();
        
            var response = new APIGatewayProxyResponse
            {
                StatusCode = (int)HttpStatusCode.OK,
                Body = JsonConvert.SerializeObject(search),
                Headers = new Dictionary<string, string> { { "Content-Type", "application/json" } }
            };

            return response;
        }```

this is my api I'm working on it.

there is a Restaurant that I will retrieve but not all the fields.

[undone coral]

use https://docs.aws.amazon.com/amazondynamodb/latest/developerguide/DynamoDBContext.ArbitraryDataMapping.html

configure a mapping for the map attribute type

[mint sleet]

@undone coral could you have a look the API above?

thıs ıs the way that I want to use in fact

[undone coral]

it looks fine to me

[mint sleet]

But I do not want to get the whole fields in a Restaurant object

[undone coral]

item.TryGetValue("Address", out var Address); it looks lie you have to just do this a little smarter. are you sure you should be doing TryGetValue like that?

i'm sure it's in there

[undone coral]

i think the attribute must have the value as an IDictionary<string, object> accessible somewhere

if it doesn't try pinging AWS support

[mint sleet]

hmm to make a query, you mean?

[undone coral]

no i'm saying in the returned object

isn't that your objective, to get the restaurant? the data is probably there

you just have to access it correctly

so yes, to make a query

[mint sleet]

restaurant have 10 fields. one of them is a dictionary and have hundrets line. if I retrieve 100 restaurants, it makes thousands of lines of code.

that is why I am going to make another API for retrieving them separately.

[undone coral]

dynamodb supports creating a class

that will be used for mapping

so does JsonFx / newtonsoft JSON

i wouldn't overthink this

[mint sleet]

hmm.

you mean, I will make this operation in server side and get data directly as json

smart

Thanks I'll give it a shot

[undone coral]

you can do that too 🙂

[uneven ivy]

How to host server?

[wooden cedar]

using which framework?

[wintry turret]

I am currently working with a compute shader. I am passing back and forth a buffer, buffer = new ComputeBuffer(data.Length, sizeof(float) * 5);. When data.Length <=21503, everything runs as smooth as butter, 200+ fps. As soon as data.Length reaches 21504, the whole thing chugs. Is there an explanation for this? I've determined that the size of each element in data does not matter. I dispatch the shader with computeShader.Dispatch(kernelIndex, data.Length / 1024, 1, 1); I suspect it has to due with the max thread size being 1024. I've noticed that the data in excess of 20480 is not processed.

[compact ingot]

you have to make tests at which group count your shader runs fastest. In most cases the group count should be a constant value and not depend on data-length. if it depends on length as part of a strategy pattern, the actual dependency function is most likely a step function with clearly defined thresholds.

[undone coral]

it sounds like you've figured out the problem. is this on windows directx 12?

you should be able to have larger buffers

[wintry turret]

No, I still don't have a solution yet. I'll try messing around with the buffer size more as I haven't fully explored it.

I'll check the DirectX version but I just created a project with the newest LTS build

[undone coral]

i think what anikki told you is the most accurate

[wintry turret]

When setting computeShader.Dispatch(kernelIndex, x, 1, 1); in the MonoBehavior and [numthreads(y, 1, 1)] in the compute shader, what relationship should x and y have? Or is it guess and check?

And a whole lot of

[frozen stump]

Hey! Been having trouble trying to get steamworks to work in my unity project. This is the error code I keep getting and unsure what else to do for it

[untold pilot]

@frozen stump seems like you may have that dll in multiple locations.

[uneven ivy]

I have little to no knowledge about hosting a server so I gave no idea what you're talking about

[undone coral]

what are you trying to do?

what are you trying to accelerate with compute shaders?

[undone coral]

for absolute beginners, you will probably want to use firebase. if your goal is to host a unity headless server, don't

[uneven ivy]

What Is a headless server?

[undone coral]

also try changing to directx12

[undone coral]

i think google "Firebase Unity" and check it out

there's also "Playfab unity"

[uneven ivy]

Ok thanks

[undone coral]

which people use for a lot of stuff

[wintry turret]

Boids

[undone coral]

oh yeah

compute shaders are so hard to use

[wintry turret]

Directx12 does nothing for me :/

[small badge]

As far as I know optimal numbers are GPU-specific. I don't think there are any "invalid" numbers, though, so you shouldn't be getting GPU crashes/timeouts unless you're doing way too much in one call, or you're indexing outside the buffer?

[humble loom]

oh hey i didn't know there was a code channel

can someone help me with an interesting problem., already out the gate unity recorder is out of the question because it doesn't work in builds. but what i need is a system where i can run my simulation manually, and record each simulation tick as its own frame. The idea is i want to do this as fast as possible, record each frame, and then set the frame rate to what it would be if it had played normally

if that makes sense

public IEnumerator RecordSimulation()
{
  // I need a recorder like this
  SOME_AWESOME_RECORDER recorder = new Recorder();

  for (int i = 0; i < frameCount; i++)
  {
    simulation.Tick = i;  // Updates screen with new sim state
    recorder.CaptureFrame();  // Adds frame to video
    yield return null;
  }

  recorder.SetPlaybackFramefrate(60);
  recorder.SaveFile();
  yield return null;
}

this is basically what i would want to do

it seems like most plugins are geared towards realtime gameplay recordings though, which doesn't make sense for my application

i'm making a debate timer, so i'd want to record the debate in real time, go back and scrub it to make it tidier, and then save off the whole thing as an overlay to video

[small badge]

So, you just have code that produces simulation frames slowly (i.e. less than 60fps), and you want to turn those images into the frames of a video that you play back at 60fps?

[humble loom]

the opposite. i have a simulation that runs at normal speed, but i want to play it back as fast as the processor will allow, and create a video file that plays it back at normal speed

[small badge]

Hm; I'm not sure you'll save much time running the simulation faster than 60fps, since encoding the video will typically be slower than real time?
A straightforward way of recording at runtime is to save out a PNG every frame and encode them into a video with something like FFMPeg.

[humble loom]

so you're saying encoding a 1 hour video will always take at least 1 hour?

[small badge]

Well, obviously it will depend a lot on the video resolution, the compression and the hardware you're using.

[humble loom]

i don't mind necessarily how long the offline encoding takes. that's whatever it's going to be

i just want it to generate the image sequence as fast as possible

[undone coral]

There’s an asset you can use

It’s just expensive

Just use that

[humble loom]

what is it

[undone coral]

It’s called like av recorder pro or something

[humble loom]

avpro capture

from what i've seen that doesn't support manual frame capture

[undone coral]

Don’t overthink it

Just use the screenshot capture and save to png

[humble loom]

can you even control this from script? the only thing i can see in their documentation is via the GUI

[undone coral]

Yes

Search capture screenshot Unity scripting

[humble loom]

oh just the unity method

i see

[undone coral]

Yea

[small badge]

I haven't tried the AVPro assets yet, though they do have free trials so they might be worth investigating.

[humble loom]

so would avpro just be used to combine the png's as an image sequence at a target framerate?

i don't mind price. if i can get it working with the watermark then i'll buy it, ez

[small badge]

I'd just use ffmpeg for that part; it's command line, fairly efficient and free.

[humble loom]

does it play nice with the unity runtime?

[undone coral]

if this is only going to run on windows, you can copy and paste ffmpeg into your streaming assets folder, then call it from there

you can use it to composite too

[humble loom]

most likely yeah

[small badge]

It's just a separate (windows) application, so you run it with Process.Start. We use it for our runtime video production, we've had no problems.

[humble loom]

oh ok

is there any async callback for the encoding process?

the image sequence solution seems even better tbh because i can "record" in a format that doesn't care about audio

and i could probably even record them async potentially

ehh scratch taht

[small badge]

You can observe the Process object and see when it exits; I don't know if there's a good async wrapper for that, but you can just poll every frame.

Ah, there is: Process.WaitForExitAsync.

[humble loom]

yeah just wrap it in a coroutine and trigger an event when it's done. that's fine if i can poll state

oh ok

[small badge]

(Though that's for real async, not coroutine async.)

[humble loom]

right

could you link the ffmpeg package you use? just so i don't download something silly

i really appreciate the help btw ^_^

[small badge]

(In theory I think it should be possible to save out frames a lot more efficiently than the built-in screenshot mechanism can, but it's a complex process; it's easy to waste time stalled waiting for data to copy off the GPU.)

[humble loom]

as long as it's faster than realtime that's all i care about

i would still be bound to fixedupdate though wouldn't i?

[small badge]

Hm, I'll see if I can find the source. I think we might be using a version bundled with a different asset (VR Panorama), but it should be identical to the standard versions.

[humble loom]

so i'd have to change the framereate of the application itself to actually see any discernable improvement

[small badge]

Not FixedUpdate, but yes, Update. In theory you can produce images faster than that, but at that point you'd be writing your own renderer which isn't tied to update, which seems like overkill.

[humble loom]

ya

well my concern with update is if i update the simulation but the next frame isn't drawn, i'm just taking two screenshots of the initial state

right?

do you generally download the source and add it to your project or literally just bundle the executable inside your project directory?

[small badge]

We just use the executable. I couldn't tell you which build provider we use, unfortunately, but any of the builds linked from http://www.ffmpeg.org/download.html should be fine.

They just include different libraries, I believe. I think everything you'd need for most purposes should be in the essentials?

[humble loom]

okeydokey. seems simple enough. thank god i have git LFS lol

oh you know what i can do

private IEnumerator Foo()
{
  myProcess.Start();
  yield return myProcess.WaitForExit();
  yield return myProcess.ExitCode;
}

would that work? or would the coroutine just hang anyway

[small badge]

That will block the main thread (and thus stall the application) until the process exits.

[humble loom]

yeah i think i should just do something like poll every few seconds

OH

process does have a callback

myProcess.Exited += (sender, e) => {/* do stuff */};

that's handy. so i'll just yield break the coroutine once i fire the process and just have some feedback mechanism in the callback

so where in the project structure do you usually drop the exe? just at the asset folder or what

just in terms of tidyness

[small badge]

If you want to use it in a player build, put it in a StreamingAssets subfolder and it will be copied to the output directory when you build.

[humble loom]

oh gotcha

is there a particular path you have to reference an executable from streaming assets? i've never done that

just

Process.Start(Application.streamingAssetsPath + "/ffmpeg.exe")

[sly grove]

Application.streamingAssetsPath is the path

[humble loom]

danke

[mint sleet]

JsonSerializer.Serialize(restaurants)

    public class Restaurant
    {
       // [DynamoDBHashKey]
        public string Id { get; set; }
        public string OrgName { get; set; }
        public string OwnerId { get; set; }
        public string RestaurantName { get; set; }
        public string Phone { get; set; }
        public string Email { get; set; }
        public string RestaurantSlogan { get; set; }
        public Address Address { get; set; }
    }```

yo!

My json returns with the empty value of Address.

I do not know why? The other attributes get converted to Json but Address.

Do you have any idea?

[gentle topaz]

what's Address?

[mint sleet]

    public class Address
    {
        public string Lon;
        public string Lat;
        public string HumanReadable;
    }```

[gentle topaz]

you have to mark it as serializable

[System.Serializable]
public class Address

[mint sleet]

hmm.

let me give it a try

No it did not work.

[gentle topaz]

you marked Address as serializable? I had accidentally written Restaurant the first time but I corrected it

[mint sleet]

I wrote to them all

    [System.Serializable]
    public class Address
    {
        public string Lon;
        public string Lat;
        public string HumanReadable;
    }```

[gentle topaz]

what if you don't make it a property

instead just put public Address address;

[mint sleet]

oh my

I did not put getter setter

maybe that ıs why

let me try

[gentle topaz]

these seem like pointless properties right now, you aren't actually using the getters or setters for anything special

you might as well just make them normal variables

[mint sleet]

what do you mean by normal variables?

We got it.

[gentle topaz]

I mean get rid of { get; set; } on all of them

you aren't using it for anything

[mint sleet]

but that will be weird but, I put just getter and setter

and it worked.

is not it strange at all?

[gentle topaz]

it is more strange to have { get; set; } that does nothing

this is basically what your code compiles to right now

public string Id
{
  get => _id;
  set => _id = value;
}

private string _id;

for every single property there

which is very pointless since you aren't using the getters or setters to add any logic

[mint sleet]

yes you have a point but I'm curious now why it ran

[drifting galleon]

incorrect. you should never have public fields. only properties. that is c# standard

enforces encapsulation.

[rugged pollen]

can anyone point me to any good resources on building a moba-style ability system in Unity? preferably using scriptable objects -- my google fu is only turning up simple examples or very complex ones

[gentle topaz]

I supposed if you are making some public library or something.. I just think it's pointless to have the overhead if you aren't going to use it for anything—and especially for a class that presumably just stores already validated data for serialization

but I guess I can't really defend it since I hardly ever use public fields anyway, just don't see the value here

[humble loom]

[JsonProperty] public int Foo {get; set;}

that's how you can serialize a property

I am also in the category of people that the only public exposure you should have are properties and methods. public fields is just bad practice, even if your properties are simple getters and setters

that's also how i like doing public constants fwiw

public string FooFolder => Path.Combine(Application.PersistentDataPath, "Foo");

a lot of people like doing public readonly strings that are all caps but that just looks gross and feels wrong imo

[wispy summit]

Take this as context.

public virtual IEnumerator InitScene()
{
    yield return FindObjectOfType<SceneController>()?.InitialiseScene();
}

Why does this not show syntax errors?
If ? returns null, then that IEnumerator block won't execute.
Isn't that essentially the same as a blank IEnumerator block? But a blank block shows return type mismatch syntax error.
So why does this block look okay?

[compact ingot]

it executes either yield return null or yields whatever is returned by InitialiseScene()

[wispy summit]

thanks. this was really eating away at me

[drifting galleon]

it will never return null btw. it will return Null so ? will not work and it'll still throw NREs

[wispy summit]

I started the code with private SceneController SceneController = null;

[drifting galleon]

FindObjectOfType will only return Null though. not null

[wispy summit]

the scenecontroller shouldn't be 'destroyed' through the lifetime of the 'this' object here

Oh

how do you check for Null. Intellisense isn't showing me any references to Null itself

[warped grove]

anyone know best practice for serializing sprites to json.net? right now its a massive block of sprite properties and im trying to imagine how it would be to edit that later

[drifting galleon]

null is c# built in null. also known as true null. Null is unity's fake null. it get's returned for every object that derives from UnityEngine.Object. that's why you must always check unity objects for null with ==. is, null coalescing, and ? does not work as expected on unity types

[wispy summit]

Yeah I usually keep that in mind. that unity might not return true null. I guess i just didn't grasp that this was a unity function return.

So I kinda went out on a limb and tested the code out on a coalesce anyway

deleted the SC object from the scene

and started playing the scene

there was no NRE

[drifting galleon]

it might still happen. better be aware of it beforehand

[wispy summit]

Normally VS notifies about using null coalesce. But I noticed it hasn't been doing that lately

[drifting galleon]

begone fake null!

hopefully unity will fix that at somepoint

[wispy summit]

fair. I just dislike the fact that i have to declare a new variable to store the return data, so i can process it in the next line again

[drifting galleon]

as long as there is fake null, unfortunately you'll have to

[wispy summit]

yeah

[drifting galleon]

fake null is probably my biggest complaint about unity

[wispy summit]

i've gotta do a '?' search throughout the project later

[drifting galleon]

especially now with nullable reference types

[wispy summit]

maybe they fixed it, is why VS isn't notifying anymore

[drifting galleon]

no. fake null is still there

[wispy summit]

i see

[drifting galleon]

and probably will be as long as we don't have a moving GC

[wispy summit]

moving garbage collector? i have to see how that works

[marsh veldt]

I haven't read the whole chat but quick tip, You can return multiple data types like this

(int id, string name) FunctionName() { return (1,"objectName"); }

[mortal lintel]

idk if anyone does java but can yall help me?

[wispy summit]

dont need to return anything. It's just an IEnumerator block. I just wanted to null coalesce a nested coroutine in a single line...

[drifting galleon]

not a java discord

[mortal lintel]

ah ok

[marsh veldt]

I see now

[humble loom]

hey could someone give me a hand? i'm trying to figure out how this ffmpeg execution is incorrect. running the process via unity

oh wow that's illegible

PS E:\Users\strop\Documents\GitHub\debate-timer\StropheumDebateTimer\Assets\StreamingAssets> E:/Users/strop/Documents/GitHub/debate-timer/StropheumDebateTimer/Assets/StreamingAssets\ffmpeg.exe -r 60 -i Frame_%05d.png C:/Users/strop/AppData/LocalLow/Stropheum/DebateTimer\Recording\test.mp4```

"could find no file with path 'Frame_%05d.png' and index in the range 0-4"

but they're certainly there

oh doi

[small badge]

It's looking for them in the current working directory, which is the StreamingAssets folder.

[humble loom]

i'm not specifying the absolute path

i just needed you to see how much of an idiot i am

OKAY so i got this

so every yield return null will re-execute on the next frame right? so all i have to do now to improve the record time is to scale UP my timescale/framerate whatever, and reduce the ffmpeg output framerate by the same factor

and that should give me the original simulation framerate right?

[wispy summit]

is yield return null; and yield return new WaitForEndOfFrame(); the same thing (in effect)?

[undone coral]

alright dr p's office hours last one before the holidays

[undone coral]

https://answers.unity.com/questions/755196/yield-return-null-vs-yield-return-waitforendoffram.html

[wispy summit]

ok let me rephrase myself. How to let a nested coroutine return and let the parent coroutine continue executing before the frame starts being rendered.

what's the most useful one in this case i mean

[undone coral]

https://answers.unity.com/questions/755196/yield-return-null-vs-yield-return-waitforendoffram.html

same thing

[undone coral]

IEnumerator ParentCoroutine() {
  Debug.Log("1");
  yield return StartCoroutine(ChildCoroutine());
  Debug.Log("3");
}

IEnumerator ChildCoroutine() {
  Debug.Log("2");
}

versus

IEnumerator ParentCoroutine() {
  Debug.Log("1");
  StartCoroutine(ChildCoroutine());
  Debug.Log("2");
}

IEnumerator ChildCoroutine() {
  yield return new WaitForEndOfFrame();
  Debug.Log("3");
}

yield return a coroutine idiosyncratically is also a yield return null

if you do not want that behaviour, i.e., if you want to start a ton of coroutines and not automatically wait a frame for each one, you have to do

[wispy summit]

I mean like synchronously nested coroutines

yield return ChildCoroutine();

[undone coral]

IEnumerator ParentCoroutine() {
  Debug.Log("1");
  var enumerator = ChildCoroutine();
  while (enumerator.MoveNext()) {
    // do what you need to do
  }
  Debug.Log("3");
}

IEnumerator ChildCoroutine() {
  Debug.Log("2");
}

[wispy summit]

so you're manually handling the movenext. i see

[undone coral]

that's if you don't want to skip the frame

[wispy summit]

yeah

[undone coral]

are you saying you want a workaround to https://forum.unity.com/threads/yielding-a-nested-ienumerator-waits-a-frame-even-on-break.772352/

[wispy summit]

so when you're 'ending' an IEnumerator block, it automatically yields and the coroutine manager, by design, waits for the next frame before it continues executing anyway?

[undone coral]

yes you can just read that forum thread

and yes what i wrote is the workaround

[wispy summit]

yeah, sorry I had to confirm if what i understood was correct or not

[undone coral]

you can yield return null

that seems all fine to me

[humble loom]

i think i want to yield return null because i want to record the frame, not the end of the frame

[undone coral]

are you trying to take a screenshot "at the right time" ?

[humble loom]

i'm trying to take a screenshot of the frame asap

[undone coral]

okay, but presumably after all the cameras and UI have rendered?

and not any later?

[humble loom]

yeeeeeeeeees?

[undone coral]

then you want

while (true) {
 yield new WaitForEndOfFrame();
 take screenshot
}

[wispy summit]

yeah i wanted the parent coroutine to continue executing from that point on without skipping a frame. Anyway, thanks. much appreciated. This clears up a 'lot' of questions for me.

[undone coral]

coooolio

[humble loom]

that's what i got now

[undone coral]

start by studying the moba asset in the asset store. all of them are based on the starcraft II / WC3 scenario stuff. unity also borrows a lot from starcraft II's architecture so

[undone coral]

this seems fine to me

what is the issue?

[humble loom]

nothing's the issue, i was just responding to messages that kept coming in lol

what i'm trying to do now is just run the video faster, capture frames faster, and then generate a slower video

[twilit gulch]

Can someone help me wrap my mind around quaternion multiplication

I have read a host of StackOverflow answers on this and also the unity doc.
What I understood was that when multiplying quaternions the frame of reference is decided by the quaternion on the left
and then the quaternion on the right adds the extra spin in that frame of reference.

Here's a code sample with comments :https://paste.ofcode.org/DBsMtvuqduFdq8ZFGpMPgQ
I've commented the part which is the problem. If someone can help explain this, it'll clear up some fundamentals. Thanks!

[austere jewel]

Picture the left hand side being an orientation, and the right hand side being a rotation of that orientation in its local space

[humble loom]

left is context, right is manipulation

you know how with ISROT every matrix transformation is sort of built off of the last one

[twilit gulch]

So when we do
transform.rotation=transform.rotation*Quarantion.euler(0,90,0)

Say at the beginning the orientation i.e transform.rotation is (0,0,40) so then the right hand side will make the orientation change to one which has been rotated from current orientation to one with 90degree spin on y? So basically just rotate the object as it is by 90 degree along global y axis. Right?

If that's true then idk, for some reason instead of that it's doing the 90 along local y axis. Sorry if I misunderstood what you guys were saying.

[austere jewel]

using UnityEngine;

public class Test : MonoBehaviour
{
    [SerializeField, Range(0, 1)]
    private float progress;
    private Quaternion rot;

    void Start() => rot = transform.localRotation;

    void Update() => transform.localRotation = rot * Quaternion.Slerp(Quaternion.identity, Quaternion.Euler(0, 90, 0), progress);
}```

[twilit gulch]

Huh, it's sort of what I did. The only difference is that the object was nested in another object. Would that change anything?

Such that the parent object has the 40 degree spin in z but the nested object in question had 0 local rotation.

[grave spruce]

So, i am trying to make my game save the variable called "firstCheckPoint" which is a bool to know if the player marked the first checkpoint (in my game the player can mark the checkpoint whereever they wants) but since the variable changes at the same time that the "SavePlayer()" function is called, it doesn't save, how could i make it save?

        {
            lastCheckPoint = transform.position;
            SavePlayer();
            firstCheckPoint = true;
            if (!GameObject.Find("Bandeira Checkpoint(Clone)"))
            {
                Instantiate(bandeira, lastCheckPoint, Quaternion.Euler(0, 0, 0));
            }
            checkingPointEnded = false;
        }```

[humble loom]

so i'm curious why this is happening. when i was manually iterating through the number of frames in the file and capturing each frame, the video was captured beyond when the video ended (and was faster)
i decided to modify the code to just start capturing frames and stop when the callback for video end is hit and it stops at the right frame it just still plays too fast

https://pastebin.com/NmVBDSbu curius what yall wizards think

[gentle topaz]

shouldn't you be capturing frames at the target framerate instead of whatever framerate you happen to be getting in unity at the moment?

just seems a bit weird to have way more frames saved than your target framerate supports, but if the encoder handles that all for you then I suppose it doesn't matter

[twilit gulch]

nvm, it works, I was probably messing up something earlier. Thanks a lot!

@austere jewel

[humble loom]

If I was just trying to record a video that would be ideal. But I'm trying to produce a 60fps video regardless of the framers I simulate

The whole idea of what I'm doing is to generate an hour long video in less than an hour

[humble loom]

so for right now i'm trying to just record the video at normal speed, and output the frames at the same framerate, and then be able to move those values around arbitrarily

something i didn't think about is i could just count the number of frames i've encoded, divide that into the length of the video i recorded and that should give me the exact framerate that should match. idk why it wouldn't be 60 tho

[wispy summit]

since you're running a simulation only to record it, I guess it would be optimal to force the game to run at that frame-rate. or a multiple of that frame rate based on the time-scaling

unless ofcourse, you could ditch timescaling and vsync altogether and provide a custom time.delta based on the framerate you want and just spontaneously render frames till your video is complete

but idk how that'd sync with animator controller though

[quaint iris]

I hope this channel is okay for this question:
I'm trying to upload my project to an empty gh repo but I can't get rid of these git lfs errors... Can I get any help

[thin mesa]

Try to push them with 'git lfs push --all'

[quaint iris]

that... worked?
I swear I tried that command already

Maybe it's something that I did previously that allowed git lfs push --all to work

I tried some stack overflow answers previously that seemed to do something because the push logs changed

Thank you lol

[thin mesa]

lmao i just quoted what was shown in the image

[quaint iris]

I'm impressed that my project is still intact after typing in git commands from stack overflow I only had a vague understanding of

[wispy maple]

When defining and instantiating a list with only one element like this:

List<BossPurpleBugState> bossStatesThatCanRandomAttack = new() { BossPurpleBugState.Moving };

VS does not complain but unity cannot compile due to this error:
Assets\Scripts\Enemies\Bosses\BossPurpleBug.cs(13,70): error CS8124: Tuple must contain at least two elements.

Wth is this crap. I really cannot instantiate a list with just one element and manipulate it later? This strikes me as very weird and i cant find much online its flooded with the same error for other cases

[drifting galleon]

you're apparently not using the latest unity version. your unity version does not yet support new() you have to write new List()

[wispy maple]

Crap yeah. using 2020 but VS 2022 and it "corrected" my instantiation and i did not pay attention to that. Thanks man, it was so weird I couldn't believe that i could not instantiate with 1 element 😄

[vernal kiln]

Hey! I'm trying to figure out whether a certain point on a terrain is a floor or a steep cliff. I have a randomized v3 with X and Z and call terrain.GetSampleHeight to get a Y value on the terrain. Could anyone help point me in the right direction? Do I cast a ray?

oh there's https://docs.unity3d.com/ScriptReference/TerrainData.GetSteepness.html

[regal olive]

Can someone help me with the new mesh api? The documentation is so sparse it's hard to find anything
So I create a mesh using

            mesh.SetVertexBufferParams(vertices.Length, new VertexAttributeDescriptor(VertexAttribute.Position, VertexAttributeFormat.Float32));
            mesh.SetVertexBufferData(vertices, 0, 0, vertices.Length);
       
            mesh.SetIndexBufferParams(triangles.Length, IndexFormat.UInt32);
            mesh.SetIndexBufferData(triangles, 0, 0, triangles.Length);

            mesh.SetSubMesh(0, new SubMeshDescriptor(0, triangles.Length));

But then I need to modify mesh vertices. Modifying the NativeArray directly does nothing, and calling SetVertexBufferData again results in

What's the proper way to modify vertices using the new api?

[drifting galleon]

pass in a new nativearray with the modified vertices

[regal olive]

So allocate a new array, then pass it in?

Any way to modify it without extra allocations?

[drifting galleon]

no. that's not what it's for

[regal olive]

Is doing what I described above possible at all in Unity?

[drifting galleon]

modifying vertices? the default api should work

[regal olive]

No, I mean modifying vertex buffer without allocating anything new. I pass a NativeArray to a mesh using SetBuffer, but I don't dispose of the array. Since I change the array, but none of the changes are reflected on the mesh, does this mean the mesh just makes an internal copy of array under the hood?

I couldn't find a lot of info about it so correct me if I'm wrong

[drifting galleon]

there's not much info about that. but most definitely

[regal olive]

Then there is no reason for it to break when passing the modified array... Experimental features never disappoint lol

[drifting galleon]

you're probably unknowingly changing some other things. e.g. maybe you're not recalculating normals

[regal olive]

Apparently, .SetUVs causes it

There's probably an advanced alternative I'm supposed to use

Oh, yes, it's in the vertex descriptor

Well case closed

@drifting galleon thanks for guiding me in the right direction

[drifting galleon]

np

[mossy vine]

Does somebody know how I can fix this

    void Crouch()
    {
        previous_y = cc.transform.position.y - cc.height / 2 - cc.skinWidth;
        //previous_y = 0f;
        if (Input.GetKeyDown(KeyCode.C))
        {
            if (isCrouching == false)
            {
                isCrouching = true;
                target_height = 1.2f;
            }
            else
            {
                isCrouching = false;
                target_height = 2f;
            }
        }
        cc.height = Mathf.MoveTowards(cc.height, target_height, 5f * Time.deltaTime);

        head.transform.position = Vector3.MoveTowards(head.transform.position, new Vector3(head.transform.position.x, cc.transform.position.y + target_height / 2 - 0.1f, head.transform.position.z), 5f * Time.deltaTime);

        cc.transform.position = Vector3.MoveTowards(cc.transform.position, new Vector3(cc.transform.position.x, previous_y + target_height / 2 + cc.skinWidth, cc.transform.position.z), 5f * Time.deltaTime);
    }

[sly grove]

What's not working about it?

[tribal helm]

The character Dodges just fine, nothing to fix.

[mint sleet]

Hey people!

I got a restaurant class and want to map the 'dynamodb' database accordingly.

I am not sure which one should be sort key which one is partition

    [DynamoDBTable("Restaurants")]
    public class Restaurants
    {
        [DynamoDBHashKey]
        public string OwnerId { get; set; }
        [DynamoDBRangeKey]
        public string Id { get; set; }
        public string OrgName { get; set; }
        public string RestaurantName { get; set; }
        public DateTime CreatedTimestamp { get; set; }
        public string Phone { get; set; }
        public string Email { get; set; }
        public string RestaurantSlogan { get; set; }
        public Address Address { get; set; }
       // public List<Catalog> RestaurantCatalogs { get; set; }
    }```

My customers should have more than one restaurants.

So when I make a query with ownerID, I want to sort all the restaurants belonged to that owner.

[sly grove]

OwnerId -> Id

[mint sleet]

what does it mean?

[sly grove]

It's analagous to the Artist -> Song Title key in the Music example https://docs.aws.amazon.com/amazondynamodb/latest/developerguide/HowItWorks.CoreComponents.html#HowItWorks.CoreComponents.TablesItemsAttributes

[sly grove]

https://docs.aws.amazon.com/amazondynamodb/latest/developerguide/HowItWorks.CoreComponents.html#HowItWorks.CoreComponents.PrimaryKey

[mint sleet]

thanks! I'll have a look now.

[sly grove]

By having OwnerId as the "partition key" and Id as the "sort key" this allows you to do things like get a list of all restaraurants for a particular owner

[mint sleet]

oh

I understood now.

[scarlet copper]

Hello guys, we use multi-scenes with own physic. We have one scene which contains stuffs like UI. This scene uses default physic settings (so normal). We use there Physics2D.CircleCast to make click and hit on other player. Since we are using multi scene and player is in another scene it is not working anymore. How can I make it possible? To make that possible for all scenes.

[scarlet copper]

Ok Found out scene.GetPhysicsScene2D() has all casts like Physics2D.

Works with it

[regal olive]

if you pre-cache something it reserves the right amount of memory to use it at Runtime right?

So if you DONT pre-cache, and just declare everything as it's required, isnt that causing more computation time for your CPU to reserve the memory needed and then apply the result? instead of just applying the result to the pre-cache...

[compact ingot]

generally, in games, the idea is to load as much as you can into memory and cleverly replace things you haven't used recently with stuff that is likely to be requested soon.

thats what unity does when it loads a scene.

[regal olive]

So basically, better to pre-cache

[compact ingot]

i'd say its required

[regal olive]

nice, just wanted to clarify im doing things in the most efficient way, I dont use var or declare floats or anything inside Update() for that reason.

everythings pre-cached at global level

[compact ingot]

that may not be all as effective as you may think

the key is that you not allocate new things on the heap in update unless you have to

but turning all variables into members is not the right way to do it, that might actually be detrimental as it increases the footprint of all objects of the type

[regal olive]

Depends on what it is, for things like Positions and Rotations i generally use an IF so the CPU only updates it if it changes from the pre-cached value:

if (transform.position != playerPosition)
{
playerPosition = transform.position;
}

That way it only updates when actually required

[compact ingot]

unity already does that check internally

[regal olive]

not sure how thats possible, your either setting playerPosition = transform.position or you arent

dont see where the check would come into play under-the-hood

[compact ingot]

oh, sorry, i read that the other way round

[regal olive]

oh lol

[compact ingot]

still its not neccessary

assignment is the cheapest operation there is

an "if" is already 2 instructions

[regal olive]

surely that depends on what your assigning, a simple IF check is 0 or 1, assigning a Vector3 is 3 different operations instead of the IFs 2

[compact ingot]

maybe, it is however moot to debate this without actually measuring the performance of the relevant code

[regal olive]

well im not the greatest at measuring this, hence why I asked here.

[compact ingot]

chances are that something else in your project the actual bottleneck

i'd say it is wasted effort

if you want to optimize at that level, you have to measure and you have to architect your systems differently to actually benefit from pipeline and cache preloading

you should, essentially, do DOTS

[regal olive]

it's not a current issue I have for performance, Im asking the question so I use the most performant method by default, clearly one must be more performant than the other, even if that benefit is so small people consider it "micro-optimization"

which I always found hilarious, if you have 2 choices, why you wouldnt choose to do the more performant method by default is a mystery to me, the whole "dont worry about it until it's a problem" mantra doesnt wash with me, thats just asking for problems down the line.

[compact ingot]

the most performant method will be also one that is so much work and hassle that your project will never finish, that is why everyone uses different levels of optimization in their projects... the art is in knowing how much optimization each component benefits from

if you are thinking about the right choice of datastructures or not using GetComponent() in Update... those are simple things you can optimize and do right all the time

but real performance code is an order of maginitude more work than 'normal code'

performance code can also be very hard to understand and change

[regal olive]

If what you said is true, nobody would ever do things like pre-caching WaitForSeconds, despite it taking a grand total of 5 seconds to do. It's just lazyness IMO.

"Do at the right time" is "dont worry about it until it's a problem". Like I said, that doesnt wash with me. IMO you do it properly or be prepared to refactor potentially hundreds of scripts down the line, and thats just stupid.

[compact ingot]

optimization is not "don't do stupid things"

and i have never suggested anything of what you implying here

you should make sure that you understand the potential benefits of your optimization... checking if a value has changed before assigning it to a variable has practically no benefit when compared to the benefit of caching a WaitForSeconds

and caching WaitForSeconds has practically no benefit when used in a async routine that runs once every 2 minutes

[regal olive]

DOTS is a completely different data structure which has nothing to do with my original question.

"real performance code" covers a wide-variety of situations, my question was way more specific.

if we stick to Monobehavior, I have read several articles that say assigning the position and rotation of say, a camera constantly at Runtime is a poor thing to do and you should check if it NEEDS to be changed before doing so.

there are also several articles that say pre-caching WaitForSeconds is way more performant than simply declaring it inside the coroutine.

and it doesnt matter if it runs every 5 hours, pre-caching it costs less than declaring it as-needed, regardless of how often your doing it.

[regal olive]

it is posible to modify any shader to add a slider to increase the amount of specular highlights?

[fresh salmon]

Please don't cross-post.

[regal olive]

does it hurt you?

[gleaming thistle]

#854851968446365696

[regal olive]

you know there are orcas and penguin diying

[gleaming thistle]

@regal olive there's rules to the server. it's really not that hard.

[fresh salmon]

Ah, a troll

[gleaming thistle]

don't be a troll.

[regal olive]

not a troll

im just real

[fresh salmon]

It's been some time

[regal olive]

does it really affect you

[gleaming thistle]

you'll get 0 help acting like this. @regal olive

it's the rules of the server. actually read them.

[regal olive]

if i posted in 2 channels?

yea but be real

it hurt?

[gleaming thistle]

man you're not gonna get any help by being an asshole.

[regal olive]

who is the asshole

here

[gleaming thistle]

you.

[regal olive]

?

for?

[gleaming thistle]

there is rules to the server.

you were informed of them.

you responded by being snarky.

[regal olive]

i asked if it hurts

if i paste in 2 channels

where is the assholness?

[fresh salmon]

It doesn't "hurt", it just shows that you're impatient

[gleaming thistle]

it's because if you get an answer in one then people in the other will still try to respond to you.

you derail other people for no reason and waste a bunch of peoples time potentially.

[regal olive]

i can see who is wasting time here

[gleaming thistle]

post in the correct channel for what you need and don't cross post.

[regal olive]

yea trolls

go away

[gleaming thistle]

Informing you of the rules is not trolling.

[fresh salmon]

Enough of this

[gleaming thistle]

you're flooding the chat and derailing it off topic. just drop it already. [hellokitty is]

I'm gonna go back to my project. gl

[fresh salmon]

<@&502884371011731486> Trolling, "aggressive-ish" answer to "don't cross-post"

Origin <#archived-code-advanced message>

[austere jewel]

@regal olive just follow the rules, and don't complain when you're called out for not following them. It's not difficult

[regal olive]

yea vertx i understand you

[austere jewel]

Seeing as you've spammed your question across the server it's extremely obnoxious

[regal olive]

i want to apologies to you vertx

woudl you accept my apologies?

i'm really sorry

i won't happen again

[humble loom]

i have a code design question i'm curious about. to minimize scene dependencies I've been thinking about refactoring the way i structure things. so if i have two monobehaviours that are essentially pretty tightly coupled. one will only ever exist when another does in the scene, what's the best way to handle that? should i just make the dependent script a class that i store as a member of the script and serialize it? or do i just make it so B requires A component and leave it at that? sort of enforce that they can only exist on the same object and not on their own

[sly grove]

Use the [RequireComponent] attribute, and get the reference with GetComponent in Awake()

[humble loom]

i figured that'd be the best practice

i suppose anything that doesn't require unity lifecycle stuff shouldn't be a monobehaviour though

[compact ingot]

there is no universal best way, just one that works best under the requirements you have.

[humble loom]

https://pastebin.com/Ezqj4a7L so like this. the only thing it's really using from unity is the inspector binding

so maybe it should just be an internal member

[sly grove]

And coroutines

[humble loom]

coroutines don't have to be invoked from monobehaviours though no?

oh shoot no they do

[sly grove]

they do

[compact ingot]

the coroutine runner i.e. StartCoroutine has to be launched from monobehaviour

you can however iterate them manually

[sly grove]

The lifecycle of a coroutine is tied to a MonoBehaviour

They can be iterated manually of course, but then you're rewriting the coroutine "engine" yourself basically

[compact ingot]

a coroutine is a C# thing, an automatically advancing coroutine is a Unity thing... those two get easily conflated

[humble loom]

yaya

[sly grove]

Coroutines are a Unity thing. Iterator methods are a C# thing.

Unity exploits iterator methods to create Coroutines.

[humble loom]

i believe unity's coroutine controller is basically just a list of coroutines that get invoked once per update right

[sly grove]

there's a few entry points for coroutines in the game loop.

There's a FixedUpdate one, an End Of Frame one

maybe a couple more

You can see all the places where "yield <something>" are mentioned in this diagram: https://docs.unity3d.com/Manual/ExecutionOrder.html

those are places where a coroutine might continue execution from depending on how it was previously yielded.

[plush hare]

does c# have a command that will give me the number of leading zero bits in an integer

[empty vortex]

Is there a way to get the dominant light direction and intensity from a SphericalHarmonicsL2 probe?
I'm using this to get the probe:
https://docs.unity3d.com/ScriptReference/LightProbes.GetInterpolatedProbe.html
but i don't quite understand spherical harmonics yet to translate the answers i found in GLSL to C# (i want to query over just 1 given point)

[sly grove]

Do log base 2 of your number, Floor() it, then do sizeof(int) - theResult

e.g. 15 gives you:
logbase2(15) = 3.9
floor(3.9) = 3
32 - 3 = 29

might be off by 1, not sure

gotta account for the sign bit too

so maybe 31 - 3

maybe it's Ceil() not Floor()

[plush hare]

Someone told me every processor has an instruction that does it in one operation

[humble loom]

so here's an interesting thing. for anyone that was familiar with my issue yesterday. i modified the code to basically take one snapshot every coroutine iteration until the video completes, and then when its' done, just take the # of frames / length to get the framerate. it looks like even though my video's framerate is 60, my games framerate is 60, i'm only getting 21.4 screenshots per second somehow

this makes me sad. i feel like capture screenshot hangs so if you call it too fast it gets ignored

[sly grove]

There is one but I don't know if C# exposes anything that directly uses that instruction

.NET's IL language might not have an instruction for it either

[compact ingot]

https://docs.microsoft.com/en-us/dotnet/api/system.numerics.bitoperations.leadingzerocount

not available in unity (yet)

[sly grove]

ah nice

[compact ingot]

there is even one using intrinsics https://docs.microsoft.com/en-us/dotnet/api/system.runtime.intrinsics.x86.lzcnt

c# getting really serious with IA level stuff there

[sand raven]

so, i have two sets of code:

                possibleMineLocations.Remove(vect);
                mineLocations.Add(vect);```
and
```vect = possibleMineLocations[UnityEngine.Random.Range(0, possibleMineLocations.Count)];
                possibleMineLocations.Remove(vect);
                mineLocations.Add(vect);```

The top block of case uses a hashset for possibleMineLocations, and the bottom set uses a list. After doing some testing, with possibleMineLocations being 250,000 entries long, and iterating the code 45,000 times, the top block of code takes about 40 seconds to finish, while the bottom set takes only 15. My question is why does the hashset block take so much longer? is it the getting the random element of the remove? If its the remove, i dont get why, since list.remove is O(n) and hashsets.remove are O(1)

[sly grove]

HashSets are not ordered.

In order to do Enumerable.elementAt on a hashset

you have to iterate through n elements

where n is the number you pass in

not only that but the element you get is basically a random one, since again, the enumeration order is not guaranteed as HashSet is not ordered

Whereas if you use a list you can get any element in O(1) time

[sand raven]

i see

[sly grove]

not only that but lists are actually ordered, so ytou can guarantee that list[n] is the nth element that was inserted in the list

or the element at index n

HashSets don't have such a concept

[sand raven]

so then I could get a random element from the hash set without having to enumerate through the entire list, it WOULD be more efficient since hashsets have a faster .remove right?

[sly grove]

well

I said "random" but what I really meant was "undefined"

[sand raven]

i mean, does it matter if its undefined since im removing it right afterwords?

[sly grove]

not really, but it's not going to exhibit "randomness" in the way picking a random index would

what you'll get is dependend on the implementation of the HashSet

It depends what you're trying to do wherther it's acceptable or not

[sand raven]

im making minesweeper, i start with making a square list/hashset of vector2's, which for a 500x500 board takes about .2 seconds. I then iterate through the blocks of code shown above x amount of times (x being how many mines are being added to the board, in the testing i did for the 500x500 board, this number is 45,000)

            {
                for (int y = 0; y < height; y++)
                {
                    possibleMineLocations.Add(new Vector2(x, y));
                }
            }```

[cursive horizon]

i'm confused as to why you need to iterate at all...can you skip the first step entirely, and then pick some random coordinates to place mines at?

[sand raven]

well i dont want two mines to be able to spawn on the same tile

[cursive horizon]

use a hashset and a while loop? 😛
alternately, there are clever algorithms that solve this

[sand raven]

what do you mean hashset and a while loop? and what algorithms?

[sly grove]

well this is essentially what enumerable.elementAt does

[cursive horizon]

something like this:

int mineCount = 10;
Vector2Int gridSize = new Vector2int (500, 500);
Hashset<Vector2Int> mineLocations;

while (mineLocations.Count < mineCount) {
    mineLocations.Add(new Vector2(Random.Range(1, gridSize.x), Random.Range(1, gridSize.y)));
}