Poll of the Day > Unity is going to start charging devs per install

(+) Yeah! (+)

what I typically do is use [Export] for any node I need to avoid hard coding

like

[Export]
TextureRect _progress;

though that's only for current (or nested) scene like you said. For nested or imported scenes I usually use that as an instance of the top level script name, and exported node references in that if I need it elsewhere typically, or use PackedScene like that

can also use groups for categorising (and I'd use some constant string for the names)

ah yeah here's how I typically use PackedScene

my own nodes

[ExportGroup("References")]
[Export]
Control _savesContainer;

things not currently in my scene

[ExportGroup("Scenes")]
[Export]
PackedScene _saveDetailsScene;
[Export]
PackedScene _startScene;

then use that like

var saveDetails = _saveDetailsScene.Instantiate<SaveDetails>();
_savesContainer.AddChild(saveDetails);
saveDetails.Init(save);

and then that has its own node references

looks like I don't have GetNode anywhere in my project using this method (though that might change, I don't spend that much time on my project unfortunately -- currently have 31 .cs files and 16 scenes for an idea of my current size, not sure how best to quantify that lol, barely any aesthetics so mostly mechanics)

I use plain Nodes as containers a lot too for grouping something of a particular type so I can go like this

-------
generally have a lot of exports for node references and configs

[ExportGroup("Config")]
[Export]
YamlResource _recipeYaml;
Recipe _recipe;

[ExportGroup("References")]
[Export]
Node _equipmentContainer;
Array<Equipment> _equipment = new Array<Equipment>();

------
then on ready do this

public override void _Ready()
{
CacheComponents();
Setup();
ConnectSignals();
}

void CacheComponents()
{
foreach (var child in _equipmentContainer.GetChildren())
{
_equipment.Add((Equipment)child);
}

_recipe = _recipeYaml.Deserialize<Recipe>();
}

-----

apparently spacing doesn't work well even with the code formatting..

What I did was just create a public field for the parent node to set on its _Ready() function.

So it's kind of like a constructor that way.

That way I can keep it modular. Blazor actually does something very similar, and it works wonderfully. Maybe in Godot we should have [Required] attributes that will throw an error if some field/property is not set before adding it as a child.

Took me about 3 days to come up with that, now I can keep working.

(The UI needed to grab the grid on the map)

https://cdn.discordapp.com/attachments/841109666158215168/1161095690013917264/villages2.gif

Called it, unifree is dead. AI is not a magic bullet that can solve all problems. It's a real tool, but people who know how it works know how to use it. There were a thousand red flags with that one, but I doubt there was any ill intent like some people are suggesting. Just another idea guy with money in a DIY space. (Like the plague of the 10,001 amazing cryptocurrencies)

I found it weird that the community entertained this jank ChatGPT/Python/AI project. The real tools grow quietly on the side without any PR.

Anyway, my game is coming along without any hitches. I've created some base units for my strategy/sandbox/clicker(?) game. A soldier, a wolf, and a villager. Wolves eat your units, soldiers scare off wolves, and villagers need to be protected.
https://media.discordapp.net/attachments/1096779647594811412/1164381647882436689/villages6.gif

Thinking about adding lava, archers, meteors, dragons.

Imagine science was treated like software is, and every study had to involve carbon fiber nanotubes the strongest structure on the planet. What do you mean I don't know what I'm doing? I have money. This is going to turn out great. Simply use AI and ChatGPT to create the nanotubes and solve world hunger. Plant trees to reduce carbon emmisions.

It probably doesn't help that with the walk-back of the fee structure to something infinitely less insane that doesn't get applied retroactively (not that that would ever have actually stood up in court), there's little to no further market for converting existing Unity projects into another engine. A ton of people have still dropped Unity for other engines because they've burned any trust people had in them, but by and large past and current projects are fine to continue existing as they always have and there's no need to cut Unity out of them like a tumour.

adjl posted...

It probably doesn't help that with the walk-back of the fee structure to something infinitely less insane that doesn't get applied retroactively (not that that would ever have actually stood up in court), there's little to no further market for converting existing Unity projects into another engine.

It's just that this was comically poorly done. ChatGPT doesn't even recognize anything that existed after 2021.

There are tools for porting assets, and they work fine for the most part, but porting your code over is going to take about a day or two of manually fixing random errors if your project is massive. (That's not bad at all). There's no AI that can understand the logic and expected output of your code, yet.

So on top of all that it was basically already a solved problem.