Is LUA really popular in China for some reason?
All the LUA Linter extensions for VSCode are chinky.
Which is a bit odd, I only see LUA used for game development normally

>>108101230
i want to kms

>>108101586
you need vitamin D
t. internet doctor

File: ygg-1770657063.webm (2.18 MB, 1920x1200)

2.18 MB WEBM

Might make some zombie characters today

>>108101965
Hopefully there's necromancy so players can engage in two way bullet hells of zombies and projectiles.

here's a fun little "sepples fact" for you: the order that the "initializer list" runs in is undefined. it can go in any order! yeah, that was fun...

>>108102594
didnt they fix that in a later version? c++17 maybe?

>>108102610

-std=c++26

it could be a g++ regression, idk

Do NOT use malloc.

>>108102594
It's defined as indeterminately-sequenced
same thing with function call arguments

>>108102542
Would this be the only real implementation of necromancy in gaming? I'm talking spawning your own monsters from the safety of your home and then sending them out with full pathfinding to stalk your enemies. The realm is persistent and we can disable any LoD which can despawn/adjust tick rates of entities with no nearby players. I want chaos on some server types.

>>108102629
odd, cppreference states its evaluated in order,
https://en.cppreference.com/w/cpp/language/list_initialization.html
> Every initializer clause is sequenced before any initializer clause that follows it in the brace-enclosed initializer list.
> This is in contrast with the arguments of a function call expression, which are unsequenced(until C++17)indeterminately sequenced(since C++17).
are you forwarding the values through a function call first?

>>108102713
so it depends if it's () or {} then

>>108102685
>>108102713
I'm talking about this shit: https://www.cppreference.com/w/cpp/language/initializer_list.html
>The order of member initializers in the list is irrelevant,
so not undefined, just defined according to convoluted rules which I surely will not learn.

>>108102700
That would be broken as hell, but also extremely rewarding. Just letting someone keep spawning those before unleashing is asking for trouble due to balance and performance reasons.
However it can unlock interesting dynamics if players can build, as you could allow players place repelling holy sigils all over the place to counter the periodic zombie inflow long term.
No idea about necromancy examples in mmos, but just luring enemies in those can cause absolute chaos.

>>108102913
It's in order of declaration, should be simple enough. I totally didn't learn that by having my program crash in a cryptic way, haha.

>>108103057
-Wall -Wextra

I know it's a PITA at first, but it'll basically warn you whenever a statement isn't 1:1 with the generated machine code. It's saved me a lot of time, over the years.

>>108100843
bro we don't even have 5 regular posters

>>108101051
Why not a static inline or something? You can use an anonymous namespace (effectively a "static namespace") if the former doesn't fit your conventions

>>108101276
Chinese people love to use good stuff for some reason.

>finally get back to my project after a couple months off to study other things (and play some RPGs from the 90s)
>Undefined reference error
>spend a couple hours trying to figure out what's going on
>turns out I didn't add a file to cmake

God, I'm glad to be back at this.

>>108101586
I tried to work the Zebra 3 beta synth all day. I feel the same. Genuine sense of impending doom.

I gave in and started using AI, I'm sorry

>>108104270
Good morning sir.

>>108104270
good morning sir

>>108104270
good morning sirs

>>108104270
You're a retard if it took you this long.

>>108105353
Good morning sir

>>108105353
I'm using it because I'm too retarded for some of the math

File: lotus.png (21 KB, 1600x1000)

21 KB PNG

does anyone know how those retro racing games rendered terrain and tracks? I suspect precomputed sine tables, some extremely simplified projection method, and lots of blitter trickery, but maybe someone has already deconstructed how they did it exactly.

>>108104270
AI is great. Literally 3x's your dev speed

File: Ygg_1224_Random_Mode7.png (185 KB, 1919x1050)

185 KB PNG

>>108105680
mode7?

>>108105847
I don't think any of the platforms Lotus was released for had anything like that. still pretty cool, I didn't know about mode7 before.

>>108105829
reading, understanding, testing/stepping through AI slop in the debugger and having to edit it takes longer than if I just wrote something myself from my experience


I genuinely don’t get the appeal outside of nocoders who just want to slam the output in and see it work

>>108106076
The less skilled you are at programming the more useful AI is

>>108101230
Do you guys think using SDL and BGFX for engine dev is a good idea?

Like if anything BGFX is more future proof than openGL, and webGPU support in C++ is lacking.

SDL has support for web builds, since godot uses it.

>>108106076
i've found it to be helpful for tracking down stupid little bugs hiding around as well as for quick reference lookup or just an extra "pair of eyes" to review my shit before i commit
asking it to write features or even design them is a fool's errand, you will get 20% functionality with 80% tech debt neatly packaged up in a professional-sounding box of deceit

>>108106236
OpenGL is probably the most future-proof thing you can use at this point

>>108106255
Really? Wasn't vulkan supposed to eat OpenGL's lunch eventually?

I thought BGFX would get rid of the headache of writing graphics code for each platform?

>>108106276
Futureproof doesn't mean "here's what we think everyone will use in the future" futureproof means "here's what everyone already uses and is already the most compatible and companies have the biggest interest in keeping it working"
That's OpenGL (and DirectX)

>>108106288
Well BGFX is open source, mojang backs it as well as many smaller studios who don't want to reprogram graphics code.

File: ss_b39b5e2c683f3692c486d3(...).jpg (363 KB, 1920x1080)

363 KB JPG

OP Pic reminds me of populous

>>108106303
I don't know if it's any good I'm just saying OpenGL is as future proof as graphics APIs get
I wouldn't take Mojang as a credible source though

File: electrostatic singularity.webm (3.68 MB, 1280x720)

3.68 MB WEBM

Of fucking course simple additions and subtractions from the charge balance would end up in a singularity.

File: electrostatic bullets.webm (3.46 MB, 1280x720)

3.46 MB WEBM

This is way better than I ever imagined the concept.

>>108105680
>I suspect precomputed sine tables, some extremely simplified projection method, and lots of blitter trickery, but maybe someone has already deconstructed how they did it exactly.
That's pretty much it. Cutting the road into horizontal slices and moving the memory locations around to precomputed locations.

>>108106076
You can literally just get it to fix it for you with simple feedback. I only debug stuff if it's really stuck and needs pointers, but even then tracking down the bugs is faster with AI too.
>t. 10 years of professional software dev experience in web, android, graphics (openGL), and windows client apps who's barely written any code in the past 3 months

File: lmao.png (458 KB, 861x714)

458 KB PNG

>>108106703
>t. 10 years of professional software dev experience in web, android, graphics (openGL), and windows client apps who's barely written any code in the past 3 months

>>108106703
>10 years experience and using AI to program
what a wretched existence

It's still programming. I just code in plain English inside markup files.

>nodevs pretending the compilers haven't been using machine learning to fix their shitty code for the last 30 years

>>108106784
are you insane?

>>108106786
Do you not know how compilers work?

>>108106796
I do know how compilers work, that's why I'm calling you insane

>>108106762
I was doing it longer than most of my peers. Most of them just became tech leads or went into management and stopped actually programming a long time ago. Even then most of my AI-assisted work is far closer to the code than sitting in stakeholder meetings all day and rubberstamping PRs.

>>108106786
Modern compilers compute hundreds of different code paths based on your source and pick the best one, often times completely rewriting poorly implemented loops or condition statements.

>>108106805
Machine learning is old, the only new thing is that it produces text now. You guys are dumb. If you know how to develop software in a team environment, you don't need to write code anymore. It's like having a list of work, and having your coworkers do all the shitty parts.

File: Sandy Petersen 🪔 (@Sandyo(...).png (197 KB, 600x735)

197 KB PNG

If you ever feel bad about your spaghetti code, this is how Age of Empires 2 was programmed.

>>108106805
>Modern compilers compute hundreds of different code paths based on your source and pick the best one
Not really, doing that would be slow
>often times completely rewriting poorly implemented loops or condition statements.
Yes but that's got nothing to do with ML

>literally no news on jai in weeks
It's so over.

>>108106814
>Not really, doing that would be slow
Yes really, that is the whole basis for the LLVM project having an intermediate non-human read source code representation in its compilation chain.

I remember my first AI boom.

>>108106824
No it's not

>>108106828
Okay, it just converts your C++ into an entirely different representational language for absolutely no reason before compiling. There are no optimization steps in the compiler where it figures out the best code path, often times rewriting branch statements and conditions.

>>108106838
You said it compiles hundreds of different code paths and picks the best one
It doesn't do that, because it would be really slow
There are many stages and optimizations and strategies for picking optimizations, "try a whole bunch and pick the best" isn't usually one of them

CPUs themselves have largely been designed wtih AI tools. Do you really think someone is placing 50 Trillion transistors on silicon by hand?

>>108106845
>There are many stages and optimizations and strategies for picking optimizations, "try a whole bunch and pick the best" isn't usually one of them
Funny because that's exactly how the branch prediction unit on my CPU works.

shut the fuck up nerds this is /gedg/ post about engine tech not your fag ass AI girlfriend bull shit

>>108106856
yes, branch prediction isnt compiling

>t. is afraid of using AI because of whats on his computer

>>108106236
when you understand the real limitations of opengl lacks in comparison to Vulkan / DX12 (DLSS/FSR/etc, raytracing, faster shader compilation, better mesh shader support, Nsight Aftermath, maybe VR, future AI compression?), BGFX is not necessarily better in terms of missing features.
Since even if BGFX could use Vulkan / DX12, it probably doesn't expose the right api to allow you use features listed above (or you would end up doing a similar amount of work as doing it all from scratch).
So obviouls this makes it clear that vulkan is the best option, but DirectX is a tiny bit better than opengl and vulkan because the DGXI independent flip mode has better desktop fullscreen / overlay support (like pressing the windows key + . to type an emoji IME while in Fullscreen, supporting unicode/emoji's in games is hell, but you can fix this on Nvidia GPU's by going into the nv control panel -> 3d settings -> vulkan / opengl present method (bottom of the list)->dxgi swapchain, and you could also break this by right clicking on the program, compatibility settings, and disabling fullscreen optimizations).
Also SDL's built in renderer uses Vulkan as the default backend on windows (I think it's because of SDL_gpu stability, which supports DX12, and it has it's own shading language it seems).
I use opengl with Angle (GL es to DX 11 / Vulkan / metal) to get a similar result that bgfx gets, but I still have the option to use the older tools that work natively with opengl (such as shader debugging in nvidia Nsight AKA clicking on a pixel and stepping through the shader, Nvidia removed this feature and you require to use a older verison of Nsight, BUT Angle does kind of work with shader debugging from testing with RenderDoc + Vulkan + retainSPIRVDebugInfo OR Angle Debug build, BUT it's still spirv... so not great...).
OpenGL could die very easily, since it seems like only Minecraft and WebGL is keeping it alive. But GL is still getting vendor extensions added to it, like VR.

>>108106995
>OpenGL could die very easily, since it seems like only Minecraft and WebGL is keeping it alive
or the tens of thousands of other applications that use it but hey who cares about those

>>108106493
I like the shell physics

>>108106995
I want to make a 2D engine that supports making Mobile releases and web releases. OpenGL does work, but i prefer to write graphics code base once. I just need the portability. Am just a bit iffy when it comes to OpenGL since it seems nooby.

>>108101586
pls no anon
i have a favorable opinion of u

>>108107158
for simple 2D, if you go with OpenGL you will likely end up with one graphics code base for all platforms. that's what I did in my framework, and I can't name a single platform-specific hack related to OpenGL I had to make. just stick with WebGL (OpenGL ES 3.0), and you should be fine.

just added fullscreen support to my framework. finally some progress.

>>108104270
its good for common problems. But having it try to implement something niche is very difficult. I tried with a anti gravity racer it produced a lot of slop thats not possible to change. Then I changed the approach and wanted claude to explain the physics of the wipeout decomp project that also didn't work out.

>>108104270
YOU FUCKING RAT
YOU PIECE OF FUCKING GARBA-
Sorry, sorry. Love you anon :D

>>108104270
use AI as a tool to help you and make you better, don't vibe code

I will vibe code. I will ask everything to AI.

Because it's too comfy.

>>108108621
good morning sir

made a custom model format based on iqm. writing the exporter in python for blender. i hate python. i can't believe that variables can be declared inside loops and still exist afterwards, or acquire properties that you just give them randomly. it's fucking insane

>>108101965
I'd pwrsonally would like to see some gore in the death animation, your game looks amazing

I have been trying to learn programming for 15 years! Then AI came to help! AI is awesome!

Be quiet now! Or just continue whining about it.

File: fb6e1472177100479f5e24ffd(...).jpg (135 KB, 500x521)

135 KB JPG

>>108101230
Looking to dip my toes in gedg. Dont know any C/C++.
Experienced with Go, Python, Bash, etc..
Will need to learn some C first?

File: font_loading.jpg (45 KB, 1010x532)

45 KB JPG

not too much to show but I got fonts loading. Using stb_truetype instead of sdl_ttf for a number of reasons. Look forward to having some actual text rendering to show.

>>108108704
try odin, it's very similar to go and is really good. I write C code everyday but for you I would highly recommend odin at least to start with. C++ is fucking disgusting but has all of the libraries and examples so it also makes sense as much as I hate to admit it.

I wish I knew someone who was doing game engines or overall was game deving in 2001.

I’m so curious if I could have made some simple 2D or even 3D game back then because my first encounter with deving was in 2005 and I didn’t even touched the code, I just did maps and models for quake2

File: driver.png (71 KB, 512x512)

71 KB PNG

Which engine is good for making a 2D driving game?
I want it to be minimal effort. This is a spite game to express my hatred.
Hopefully I can use very few objects by using a percentage chance of CPU controlled cars having one of several attributes when they are spawned or in proximity to other objects. Overall each vehicle can probably be the same object. No need for objects to persist when off screen.
I was going to give godot a chance. I have some experience with bevy, but I just don't think it is right for this.
Running Ubuntu and I want this to be a desktop game.
>>108101965
Looks sick. Which engine?

>>108109160
>This is a spite game to express my hatred
For what?

>>108109255
Other drivers.

>>108109160
Just use godot. Old versions of Game Maker had like a vertical arcade style car game bundled with the program.

Should I make a Falloutlike (but make it more like a blend of NV's faction/narrative gameplay and 4's looter-shooter gameplay)?

How do you guys store your sprite data?
I've been thinking of having a csv file or something, just so I don't have to keep hardcoding the data.

>>108109704
Directory tree structure

>>108109704
I have two options: texture atlas PNG file + JSON, and I also can read Pyxel Edit documents

>>108109160
I'm enginedev

>>108109704
A dictionary keyed to the sprite (png) that loads to an atlas

>>108109810
Of course!

>>108109851
>>108109869
Yeah, that makes sense. I've decided to do some macro trickery to avoid having to declare things in two separate files. Hooray, C. I might change things in the future if the need arises, or if compile times get unruly.

>>108109704
>I've been thinking of having a csv file or something, just so I don't have to keep hardcoding the data.
How often do you actually change that data? Do people without access to the source code need to change stuff in that data? There's a big meme about putting shit into csv, spreadsheets, or whatever other "non code" format when in 95% of instances it's just another layer of bullshit to develop and think about.

Do you really gain anything by writing the same lines into a csv compared to a container of your choice in code?

>>108110061
Yeah, so I can directly define my sprites while working on a texture, not having to recompile things when I change proportions, etc.
It's not that I don't have reasons; I'm just more focused on getting my engine (and game) running right now than on making assets.
Shit, I could ship a small-ish game without all that functionality, but I don't have any ideas for a small game that I'd want to do.

File: VgH57Dv.jpg (203 KB, 2048x1924)

203 KB JPG

Post music you're listening to rn
I'm listening to these playlists (semi-irnoically):
https://open.spotify.com/playlist/73qRbp9X5LI2n9M0TWqdy8
https://open.spotify.com/playlist/0zfgWd1IwBQnb0t5Yh4Fl4

File: pepe-sing-peepo-sing.gif (31 KB, 478x640)

31 KB GIF

>>108110382
LIKE A FELLA ONCE SAID
AIN'T THAT A KICK IN THE HEAD

>>108110382
>spotify
lmao

File: 0335558080c6d042.png (444 KB, 1282x749)

444 KB PNG

Tried turning on texture filtering for the colormap and normal map and it... doesn't look bad?

Anyone else doing
World > Region > Area > Tile

>>108105680
It's called pseudo 3D
http://www.extentofthejam.com/pseudo/
https://www.lexaloffle.com/bbs/?pid=69343

>>108106311
I can hear that pic. Also Populous: The Beginning was my first thought too.

A lot of games use Noesis/Ultralight to use html/xaml for their UI. But I’m pretty sure AI is good enough now you can just vibe code such systems so you don’t have to pay for them. I’m going to try it myself. Wish me luck fellas.

>>108110991
exactly what I've been looking for, thanks anon

File: ygg-spritestack-013025-vp9.webm (2.54 MB, 1600x900)

2.54 MB WEBM

>>108111251
Not the specific technique asked, but another pseudo 3d technique (sprite stacking)

>>108111381
>pixelated sprites
>"I-It's art. Okay!"