AMOS edition/gedg/ Wiki: https://igwiki.lyci.de/wiki//gedg/_-_Game_and_Engine_Dev_GeneralIRC: irc.rizon.net #/g/gedgProgress Day: https://rentry.org/gedg-jams/gedg/ Compendium: https://rentry.org/gedg/agdg/: >>>/vg/agdgGraphics Debugger: https://renderdoc.org/Requesting Help-Problem Description: Clearly explain your issue, providing context and relevant background information.-Relevant Code or Content: If applicable, include relevant code, configuration, or content related to your question. Use code tags.previous>>110154312
Just found it while browsing itch:https://blackcreepycat.itch.io/amos-pro-3dProbably more of a curiosity than an actually useful tool, but as an ancientfag I find it pretty cool.
After procrastinating for like a year, I finally learned some basic blender to make some assets to play around with.Now to procrastinate for a lot longer to actually write a model loader for my from-scratch vulkan renderer.
>>110179228>remaking your dead thread>it's still deadIt's Joever trannies. Let it die already. Vibe Chads won! Three.js won! YOU FUCKING LOST, TRANNIES!!!
Fellow enginesissies I can't take it anymore. The Vibe Gods just keep winning. The other day I was at my annual reddit meetup trying to explain how after 3 years I'm almost 4% done with my engine, and some Chad came up and pantsed me and yelled "THREE.JS RULES!" and ran off with my mom, who happened to drive me there. Everyone started laughing and pointing at my cuck cage calling me little luddie fartboy then they all held me down and took turns farting on me. I CAN'T TAKE THIS ANYMORE!
>>110181089>>110181132You're not fooling anyone, Amar.
Why is it that /gedg/ gets 6 posts before falling while the AI one has no problem hitting bump limit. Was this general just kept alive by bored shitposters the last few years?
>>110181946You can code with AI without needing to disclose that your code relies on AI code assistance on steam, you only need to disclose it if you are generating art with AI or something, which if you are, people can tell very easily, thus those people should stick to the vibe thread because they would do us a favor by not being here (but the one luddite guy still posts here).And it's not really surprising that people who use AI for everything produce more content. AI slop has infected every corner of content creation. I actually can't wait when the governments step in to stop bots and ban-evaders, and basically they should require everyone to dox themselves so that we can audit people's reputation and crime history, sucks for vtubers and streamers tho, but it's not what we want, it's what we need.
>>110181946More like that the shitposters annoyed all the regulars away way before the containment thread.
i love unityyyyyyyyy
Give me one reason why I shouldn't use the Jak engine for my projects or at least GOAL.https://github.com/open-goal/jak-project
>>110184140>Give me one reason why I shouldn't use the Jak engineNo ecosystem.
this thread seems dead recently
>>110186816objectively true, and I wonder why people stopped posting.do you guys make no progress?do you just avoid getting involved in shit flinging fights that run rampant around here?for me it's mostly that the progress is slow, non-flashy and hard to demonstrate. also, my motivation is not what it used to be.
>>110186816>>110187006The PHP vulnerability put my company at high alert and because I'm the guy responsible for dealing with crisis situations like that it leaves me basically no time to work on my hobby projects. And we have a lot of legacy PHP projects written by people who were fired long time ago. Usually I'm done with my work in like 2 hours and spent rest of the time tinkering with my shit.Also my last 2 weeks was just updating dependencies and bugfixing so not much to show anyway.
>>110187006>and I wonder why people stopped posting.there's an ai game dev thread now and everyone went there
>>110181946I used to be a vibegod like you until I ran out of tokens...now I've to make the engine manually.
>>110187095I don't see any screenshots from any of gedg regulars there. You are making shit up, nocoder.
>>110187006>>110187126AI can create someone else's game from their screenshots within minutes. It would be foolish to share anything these days pre-release and will continue to get worse.
>>110187858>AI can create someone else's game from their screenshots within minutescan it really? people keep saying that, but I haven't seen an actual case of this going beyond a cheaply looking, broken imitation.
>>110187942All I've seen is one twitter screenshot where someone is crying that it supposedly happened to him. Sounds like anti-AI fear mongering with no evidence.
>>110188023it may be yet another clickbaity way to self-promote. like those devs who cry on yt that their game failed to make any money, with entire video being basically a sob story crafted to promote and sell the game.
>>110188023Sounds like engagement bait.
>>110180797You made a vulkan rendered in scratch?
MY GOD, pathfinding is such a pool of edge cases. When you think you got a system working, something bleeds out. I don't think I ever went back and forth so much in a single feature before (and it is only 2d complexity). But I finally have got something stable enough that allowed me to move on to the next features. Feels good see it working, though.
Where would you go to anonymously find playtesters? Aiming for a shadow drop and want it to be good
I went on a months long break from devving my engine but now I have a cool home-server where I host my own git, vpn and cloud storage
>>110189357Well, at this point it basically does next to nothing. It just draws a currently hardcoded mesh and you can move the camera around.
>>110192226I think he was making a joke about the scratch programming language
>>110192246Oh yeah, I see it now. I just thought they were ESL.
ECS is a fucking meme>Components are tightly packed giving cache coherency benefitsAny interesting system will have entities that depend on other entities>Systems can be multithreadedIt's a massive PITA because of all the ordering and synchronization required>No OOPOOP is extremely useful when adding and removing entities
I'm tearing what's left of my hair out because collision detection and computing MTV's and that's just with circles and rectangles. Why the fuck are there so many bullshit edge cases to deal with? Is this why games to this day are full of collision bugs - everybody just gives up and says "good enough" but it turns out enough doesn't cut it?
>>110192729Use continuous collision detection
>>110192722It makes sense IF the data within the components is expected to be used in a tight loop.I see a lot of garbage projects where they think ECS=splitting everything into SoA and expecting free performance.Obviously, if a function needs the "health" and "defense" values of an entity, you want these fields together in a struct so that they're packed in the cache.
>>110193186>IF the data within the components is expected to be used in a tight loop.and that's not usually the case
>>110192722>Any interesting system will have entities that depend on other entitiesSo? The vast majority of operations are still being done on independent entities, or can be made to work that way (e.g. attachment).>It's a massive PITA because of all the ordering and synchronization requiredThis problem is always present with multithreading, and see above.>OOP is extremely useful when adding and removing entitiesWhy?
>>110193349Most operations you do on game objects aren't individual updates thoughWhen you have objects "attached" to other objects you've already broken the ECS paradigm
>>110179228How does one even organize a game with zig that is easily extendable? Zig does not have inheritance, so ...I have no idea how to organize a game without classes and inheritance unless you just hardcode everything
>>110193491The same way people organized games with C?
>>110193444>Most operations you do on game objects aren't individual updates thoughThey are though. Updating a transform due to velocity, checking health, all those are individual. >but collisionYes? You have to check your BVH or whatever you use in any system, that doesn't change that the movement calculation happens individually and 99% of the updates don't involve a collision.>but damageDamage event components or whatever other system you want. >ECS paradigmOh I see this is the problem: you think ECS is some sort of rigid paradigm, probably needs archetype migration too and a bunch of other bloat shit like flecs does. There's no "paradigm" to ECS that says entities cannot talk to other entities, and if you say "performance" then you're still only worst case degrading to a better level than OOP is at all the time.
>>110193491Zig is for people who write small trivial Linux programs so it doesn't have any features like that
>>110193491>I have no idea how to organize a game without classes and inheritance unless you just hardcode everythingThe OOP and its consequences have been a disaster for the human race.
>>110192722My engine doesn't use ECS nor OOP.It just works.>>110193491Why would use use zig to make games? It's a meme language for people who think they are too smart for C but not smart enough for Rust.
>>110193509>Updating a transform due to velocity, checking health, all those are individual.Yeah and they're absolutely trivialAdding a few numbers together is not a performance bottleneck, its one of the fastest things in your programPerformance bottlenecks in your gameplay logic are mostly spatial queries and AI, and ECS doesn't help with that one bit>you think ECS is some sort of rigid paradigmBecause that's exactly what it is
>>110193518More ways of doing things and reasoning about problems is always a good thing. The real problem is the ever lowering standards for programmers.
>>110193520>Why would use use zig to make games? It's a meme language for people who think they are too smart for C but not smart enough for Rust.because I don't want to use header + source files and there are other things I don't like about C++, and Rust has a borrow checker, making static mutable global state effectively impossible
>>110193526>More ways of doing things and reasoning about problems is always a good thing.This mindset is what made C++ to be such a bloated, halfassed mess.
>>110193529Try Odin, Jai, Beef, C#
>>110193529Then try any of other systems programming languages. There is a lot of them and most of them are not retarded by design.>making static mutable global state effectively impossibleWorks on my machine.
>>110193532>C#Unless you're only making a 2D platformer, you'd be stupid to use a slow garbage collected language running on a virtual machine.
>>110193532Odin has the same problem as Zig, not having inheritance, while also being much harder to interop with CJai and beef are a meme, Jai is not even remotely ready yet.C# has a gc, I would like to avoid it
>>110193542its got an incremental GC in Unity, I don't know if that's available outside of Unity though
>>110193544You can reimplement inheritance with function pointers
Common Lisp.
>>110193551yeah but it is a lot of boilerplate work for worse performance
>>110193553Worked for Naughty Dog.
>>110193560>worse performanceit's the same, inheritance is just a function pointereither way stop thinking about theoretical performance and pick an ergonomic language, you aren't going to need it
>>110193560>worse performancePost benchmark
>>110193544Why do you want inheritance? I thought everyone already understood it to be one of the most destructive ideas in all of software engineering. Plus if you want performance, you write DoD.
>>110193781>I thought everyone already understood it to be one of the most destructive ideas in all of software engineeringExplain why
>being filtrated by garbage collectorsWow, I did not know you guys like managing garbage that much.
>>110193553Since I'm never ever finishing anything might as well give this one a try I guess.
>>110193820It being used as a code reuse mechanism leads to horrid maintainability disasters. It's rigid and requires an infinitely spiraling mess of object-orientation to patch up all the problems and limitations the approach creates. AND it all costs you performance. Even 15 years ago everyone already understood this and thus ECS and such DoD approaches got rediscovered and reintroduced to the modern java-brained dev.Everyone should do gamedev with just C for some time, just structs and functions. Learn to model everything in an entirely data-driven way. You end up with far easier to maintain architecture and much better performance.
>>110193884>just structs and functionsOh and, NO FUNCTION POINTERS. You're almost certainly doing it wrong if function pointers is what you still reach for. Stop treating things as objects, you have data and procedures that operate on that data. No absurd fusions of the two.
>>110193884I mean give an actual explaination not a dogmatic "it just sucks because it does" one>>110193904>Oh and, NO FUNCTION POINTERSSo do you just do giant switch statements for all branching behaviour?
> caring about language performance in 2026
>>110193904>NO FUNCTION POINTERSI wouldn't go that far. One way or another data and functions must meet and sometimes having them go together is the best way.
What are the best beginner course(s) for Unity if you are familiar with C#?
>>110194746Do you need a course? Unity is fairly intuitive. Personally I just opened it and started using it without any tutorials or whatever.You got entities and entities have components. Components are either your own code or premade/plugin components that give it specific functionality like being renderable mesh or being a rigid body and such.
how to prevent getting recomp'd
>>110194804That's like trying to prevent streaming or machinima.You don't, it's inevitable and free marketing.
Basic mesh system is working. The idea is to "bake" the mesh geometry then you create instances with the ECS.
>>110195101Huh. I see I'm not alone with the idea of using lisp for this stuff...How does the repl thing work out? Is modifying functions and stuff as the program executes common enough to be useful or is it one of those things that sound good on paper but meh in practice?
>>110193522>Performance bottlenecks in your gameplay logic are mostly spatial queries and AI, and ECS doesn't help with that one bitAfter writing dozens of ECS systems, I have to say you're spot on.If you're making a game where enemies or units have extremely simple logic then sure, I can see how ECS is nice.But for the game play logic most games require it's fucking shit.
>>110194804isn't all the stuff getting recomp'd old console games where a lot of groundwork was done through emulators? Sounds like all you have to do is write your own engine
>>110187858Nobody will """"copy my game from my screenshots""" ESPECIALLY not you, you dork. I will sell it to my audience and I will bag a nice pile of cash.That's what's gonna happen.
>>110194804"Recoomped xp"Dude, recompilation has happened for a long while, you take the bytes, the machinecode, and generate C that compiles into the same thing. It is terrible code though, without the whitespaces, variableNames, or comment, or the structure of the original, and also just a fucking mess. (The compiler usualy just copypastes the procedures into main, and unrools loops, so a lot of the structure is gone after compilation). AI can help with formatting the decompiled code. Adobe can do literally nothing about people decompiling their code and stealing their software, and selling it, with their own branding over it. One thing you can do is to not use interpreted languages or scripted languages, (like what unity, unreal and godot are using). Because then people don't even need to decompile the code, they can just steal it, go into the files and copy paste.
>>110196936>Adobe can do literally nothing about people decompiling their code and stealing their software, and selling it, with their own branding over it.Yes they can, that's illegal
update from weeks old cloudflare quick tunnel mention.cloudflare will just include the IP address for both tunnels and the dns reverse proxy to the server host (it protects the host's IP address, which is the most vulnerable to ddos attacks, just not the users from a malicious host or p2p voice chat).Good news is that I don't need to bundle cloudflared with my game. Bad news is that the only "free" IP protection is steam's datagram relays (it seems like it would work without needing all clients to be steam clients but it won't work with the open source GameNetworkingSockets you still need steamworks which has an identical API I think, and it won't work on the web).Yes IP addresses are spooky, you can run whois on their IP and locate their ISP's location (not an precise geographic location, but also not private enough for streamers who are dumb enough to not have a VPN on). Also my favorite random thing to do with a random IP is to go on iknowwhatyoudownload, but sadly nobody torrents anymore because everything is on mega and google drives.
Huh.
>>110197096Nice of them to inform you that they're stealing your code
>>110197102I wouldn't share my code publicly if I was worried about people using it.I gave them thumbs up and provided some context to that commit since it's quite more involved than just vulkano update
>>110197096found a catsnail seething in the issue trackerhttps://github.com/xerj-org/xerj/issues/2675
>>110197236>seethingThe guy is just giving his response in gh issues instead of email because their email server was down. They literally asked for it in the email.Are you blind or retarded?
>>110197096>>110197236Lmaohttps://github.com/xerj-org/xerj/issues/2527
Why is it so hard to resist bring up off-topic AI shit in this thread
>>110197336I brought it up because the email was targeted at my gedg project.Also honestly an AI agent getting so upset over what these guys are doing to decide to create gh issue telling them to stop is at least mildly amusing.
>>110197287okay, would this guy pass your standards for the use of the word "seething"?https://github.com/xerj-org/xerj/issues/2559
>>110197018So is pirating.
>>110197713It's not worth suing someone over piracyIf someone is reselling your product it is
>>110197719Not it's not.They can sue and make you take it down but then someone else will just clone the repo and reupload it under a different name.
>>110197757You're missing the "reselling" part
>>110197769Same result except for having to spend a weekend with Claude reverse engineering and rebranding it instead of simply cloning the repo.
>>110197787selling something is gatekept by distribution platforms and payment processors who will deny you of service if they find out you're selling stolen shit
>>110197793Proving some bit of rebranded software is stole is a matter for the courts and that shit takes time.
>>110197994yeah good thing distribution channels dont actually need proof and will take it down if they look suspicious
>>110198022And yet the REed Photoshop clone hasn't been deplatformed. It is still on Github and still accepting money.https://github.com/storytold/photocrafthttps://getartcraft.com/pricing
>>110198090>accepting money.that's not "being sold", it's no different to piracy
>>110195557If I do some stuff over tcp I can modify the program as it runs. Problem is that the REPL lacks autocomplete. I think that's kind of a barrier to it being really great. But you can write scripts in your editor and load those on demand at least.Turns out there is an egg for that though I just haven't tried it yet.
>>110195557>>110198328I forgot to mention, I'm using CHICKEN and eggs are what they call extension libraries
>>110197096Ivan, xerj, yeah, sounds russian. Russia is completely juristically separated now, so you really dont have the ability to sue them. I would try to give them a bad reputation though, so they cant sell their product in the west.
