Відео

Thanks!

Unfortunately that is how oop works. You start seeing the value in OOP when you are already creating a project for 6 months and everything starts breaking. There is nothing wrong with writing if-else statement at the prototype level. But when this one condition grows to 100 or more lines of code 2 things happen: 1) it will become hard to work with (and there is where errors come from) 2) you can be sure that it will grow even more. But I agree - initially its very hard to find the value in SOLID and learning OOP takes very long. Data Oriented programming seems like a much better approach (so unity DOTS for example)

I believe that we use URP in the project. Still this is definitely a shared problem where you fix it by swapping the shader of your material to a compatible one.

Thanks!

Thanks for watching! 🙂

I'm really sorry to hear that! I will need to create a new version of this tutorial that uses the new input and ui toolkit

Since Libre Sprite is a fork of Aseprite it depends on what commit it's file exporter is based off of. Aseprite has meta data inside of the file that keeps track of layers, animation settings, and so forth. As long as Libre Sprite has not edited any of the data that is exported from the animation section it most likely will work unless there is some hidden checking for what program the file came from on Unity's side, but I think it is just checking the file extension type.

Thanks for answering this question 👍

@@SunnyValleyStudio No problem. I should mention I don't know the last time Libre Sprite did a branch merge from Aseprite, so not sure how many commits the main Aseprite repo is ahead of it.

Glad it helped! 🙂 It isn't an easy topic to explain but I did my best to explain it as I understand it.

I think it displays only the 'public' methods. It also will not show methods with an argument that can't be passed through the inspector. It might the case that you are missing UnityEvent<..> (triangular brackets) in your UnityEvent definition? I try to always pass the correct argument there or not pass one at all.

Thanks for watching! 🙂

There are 2 ways to make child object follow some 'Target' object. 1 - you convert the Target object position to child local space using child.transform.InverseTransformPoint(target.transform.position) or 2 - detach the child from parent, make it follow the global position and at the end attach it pack to parent using public void transform.SetParent(Transform parent, bool worldPositionStays); (there 2nd argument is set to true)

@@SunnyValleyStudio thank you very much, the second method was able to solve my problem👍

Hey! It is placed in ObjectPlacer only because I try to separate "code data" from "visual data". Why? Because when you do any kind of save / load functionality you save what is in PlacedData and load it. Next you can call the ObjectPlacer to again place those objects like we would do when placing an object when just playing a game. It is not something that we need to ever save because we can always recreate it from important data - PlacedData. But that is what I like to do. You do what works for you I am using it State pattern. refactoring.guru/design-patterns/state . Its not the same as "Data State". Yeah come to think of it it can be confusing. But I always think of it in terms of State pattern so I didn't think that it might be confused with anything else.

@@SunnyValleyStudio Thanks for your quick response! Makes sense, I'll definitely go over the recourse you linked. I stuck with your architecture and I'm retrieving the data from PlacedData as you suggest here. Cool stuff ✌️🙏

Glad it was helpful! 👍

Thanks for that! 🙂

In the game view you can select camera that you are viewing at the top - so this might be the issue. Or just delete "main camera" or maybe you don't have it in the hierarchy? Definitely something to do with the camera. If you have one in your scene and select it in the hierarchy you should see the preview that shows the same thin as the game view.

Yes 🙂

I would place "interactive" objects like chests or work bench as GameObjects that we spawn. Now do mind that I haven't touched lighting system so I am not sure how light would affect such objects. Still that way we can add any logic we want and we can just spawn and remove them when we travel around the map. If you have some special types of blocks you could add the logic in our "raycast" logic for breaking stuff. Sorry for not covering those things. But the amount of time it took me to create vs the the support I got for it from community makes it look like a waste of time 😅 That is why I didn't continue it.

I'm glad that you find it helpful 👍

I'm using Quest 2

Thanks for watching! 🙂

I can't recall the code but the idea was that if you find a tile with only 1 neighbor it means that it is a dead end. For bigger corridors it would not work so we would have to remove dead ends before we make the corridors larger.

Oh I got it thanks, I think for larger corridors you just need to increase the neighbor limit to 2 for example (if the corridor width is 2) and that would work no?

I'm really glad that those videos are still useful 🙂

Many thanks! 🙂

Thank you! 😊

Interesting. Well maybe I have messed something up in the explanation. I have used this a few times and usually it works out great. Especially when you use something like State pattern to separate states. Maybe its just that despite the delay some other code that triggers attack again runs? Like some "bool" flag check is missing?

if im correct then you can set the delay to any amount and you can still spam it

@@ggnouhiiam The idea is that you connect input with logic that triggers the attack. Next you block the trigger and use a delay to unblock it. The only way to spam it would be if the trigger is not blocked. In that case the delay indeed doesn't matter. That is why I think it might be a simple bug in the code where the bool flag is not set if is not blocking the trigger.

yeah i figured it out when i saw the block attack

Thanks for watching!

Sorry to hear that! The problem with 2D sprites are drawn on Quads and use transparency to show only part of the image that has colors. We usually use have to stretch the pixels in 4 or 8 directions to create an outline. Something like here ua-cam.com/video/MqpyXhBIRSw/v-deo.htmlsi=c5eTPqeQPzDhs417

@@SunnyValleyStudio Thank you :)

Thanks for watching! 🙂

Thanks for letting me know! I am planning an updated version that uses Jobs system + Burst or maybe even ECS. Might not be a step by step but an overview of a premade, downloadable project this time though 🙂

If you use the default Rule tile you can define there the rules of what tile to put in on the edges where the rule tile connects with some other tile. This is problematic in case where we have an outer "water" layer to which sand needs to connect and "inner" grass layer that sand again needs to connect to. In my setup I have created a "custom" rule tile (you can copy the rules from your default one to the new one using copy / paste). The point of the new Rule tile is to assume that "Grass: knows how to connect to sand. If it is so "Sand" should treat "Grass" tiles as "Sand". So in my custom Rule Tile you give the "Sand" a "Grass" rule tile - preventing the "Sand" tile for creating the 'water edge' where it meets the sand. I know that it sounds pretty tricky when you would have 3 different tiles meeting up but that is when you can create a more advanced custom Rule tile or do as I did with water - create a "FixTilemap" and manually fix the places where the RuleTile has placed an incorrect tile. I hope it helps!

@@SunnyValleyStudio thank you for the reply, I'm currently working on it, and your system is helping me a lot, another question would be if tilemaps is optimal for handling a very large world, pretty much like terraria, would that be viable somehow? or i am better off just instatiating sprites in a grid?

You can make the detector on the AI Agent detect other agents as Obstacles (I think we have separated the detector and have it some "LayerMask" - so you would just add "Enemy" layer to it). That being said it can be tricky and might require adding some "weights" to the different layers so that you don't end up with Enemies that stands idly because they can't squeeze between obstacle and another Enemy.

It might be that the tile placement system doesn't returns a TileBase so "null" tile is painted (but the tile position was calculated correctly). In a worse case it might be a code issue where one of the methods has a bug and skips some positions creating gaps. I think I link to a github repo where you can download a unitypackage per each vid if you want to duplicate the project / create new one and just import my package to compare the setup 🙂

@@SunnyValleyStudio thanks I will try some stuff out when i get home

Because we are using Kinematic RB for the item if I am not mistaken. If you make the item to have a Dynamic RB the movement will be jagged but any collision between it and wall will be detected. It is the same problem as you can see in VR games. There is no one simple solution for it. The easiest one is using a dynamic Rigidbody to prevent collision. The most "precise" one is using kinematic RB and coding your own collision reaction - so for example if player tries to push object too far into the wall the Player drops the object (the object stays on the correct side of the wall). Again there are many others so you would have to do a bit of testing / research to find out what works best for you.

@@SunnyValleyStudio Thanks for the answer. Yea i would have hoped that this mechanic was the same as we see in example Portal. Not having any success implementing is so far sadly.

I appreciate that!

You're welcome! 🙂

It can be animation related. I recall that we had set the Animator on the weapon ? If so if the animation doesn't loop and stays in the "swing" animation (I don't remember how we have called it exactly) the animator directs the position of the weapon - overriding your input. So my best guess is that the animation never transitions back to "idle" state in the weapon animator / player animator. I hope it helps!

Thanks for watching! 🙂

Must be an issue with the scale. When we import the models we can specify what scale they are in docs.unity3d.com/Manual/class-FBXImporter.html If you don't do it correctly you can see that the prefab / model in the hierarchy when placed has a parent object with scale (1,1,1) but the child model object has some weird values like (100,100,100). I hope this helps!

@@SunnyValleyStudio It works now thanks for the help!

Glad to hear it! 🙂

I have used cc0 sounds from freesound.org/ + Audacity to extract part that works for me

I have run it in WebGL and the generation took 2-3 sec. It is by no means optimized. I am planning a paid (5-10 $) version that uses Jobs + burst since those can work on mobile (but not in webgl). The layered approach works well for testing but many steps could be combined into a single for loop. Again I would prefer to use jobs system but in PCG usually optimization comes after you know that the generation works. Another thing that I don't plan to tackle is map streaming. If you want to have multiple islands you would be wise to separate the map into chunks and only render the visible part. Unless you are cratering an infinite word you can generate all the data at once and just integrate map streaming into the rendering process (which is separate anyhow). Just don't forget about the collider generation step since those 2 overlap. Sorry for making those tightly coupled!

Glad it was helpful! 👍

You can use DOT product using the up direction and the direction from player to mouse position. Just flip the X (I think) axis of rotation (transform component) of the character gameobjects.

it's too complex for me, I try to implement that in my game but I can't change what I want. I want to use Tilmap instead of Tilebase for the ocean because I've got a material attached but I've got a lot of errors, so I will do it my-self. It's just for my project, otherwise Thank you for sharing,

Sorry about that. Yes Procedural generation can be a bit daunting. You could set the ocean tile to "None" in the renderer so that it does not render the ocean / shallow sea tiles (at least I think this would work in your case)

I think that you could use docs.unity3d.com/ScriptReference/ExecuteInEditMode.html on the Update() of the MapGenerator.cs (or MapGeneration.cs) and add there code like "if R key is pressed call GenerateMap() method". But a simple approach would be to just run the game and sizable / enable steps and regenerate the map at runtime. I will try to make it so that steps can be "disabled" from the inspector rather than that you need to click each separate checkbox for each step.

Thanks! Making it work took about a week (about 20-30 h). Coding this is pretty simple. Tweaking the values and adding sprites to handle edge RuleTile cases took the longest 🙂 For example I still have to add a step to remove island of size 1x1 that sometimes appear after using circular mask in the base step. Finding and fixing those kind of things takes the longest.