/v/ - Video Games

>>15623279

Just wait until you refactor it another two times to get something that's actually done properly

>>15623279

I'm thinking about writing up my own. Any suggestions on where to start first? It's my biggest hurdle right now. Feel free to call me faggot.

>>15623300

I've done it enough times to know this won't be the last time and that there is no such thing as a perfect engine, but every time it happens I reach new heights.

>>15623382

Your own engine? There's many ways to go about it, and I can't really give you anything but my personal experience, which is also a very low level "starting with nothing" kind of route and I know that isn't for everyone. Handmade Hero, including the "intro to C" part got me started with making a compiled program, opening a window, and putting stuff into it. However I already had quite a bit experience working on projects with scripting languages and other engines, so I can't tell if it's a good place to start.

Searching for answers online and not killing yourself because all the information and advice is more or less shit and you have to churn 70 different sources of bullshit in your head into a viable comprehension of the topic is an invaluable skill.

>>15624113

Why did you add some kind of "ploc ploc" sound every time your character makes a step?

>>15624547

I just added SFX calls to everything I might want a sound effect on so I know which I need to do, or until I decide that action doesn't need a sound effect.

>>15623440

Looks pretty ugly without the proper shading.

Pretty cool otherwise.

Also, is the chandelier is a mesh rather than a texture?

Hit a little stumbling block on my texture packing thing. Let's say I do a Texture.Load("character.png") and generate a Texture object, which I then pass into my TextureAtlas object.

What should happen next? I have it so it packs it correctly, but then I'm not sure how best to relay the offsets to the texture. Like, should it internally maintain a Dict<Texture,Vector2> for offset coordinates? Should the Add() method return the offset? I can't really expect to keep the original Texture object in memory and use its object reference to find it in the atlas, since I'm basically doubling up on texture information at that point, and also Add() shouldn't return anything

>>15624658

It's a mesh.

>>15625716

How come you guys don't setup a patreon. It's been a good long while and you could have been making spank bank with a small following.

>>15625841

post the original I really want to see the goofy suit again.

>>15627566

Use is_action_just_pressed instead of is_action_pressed

>>15627572

I'll be adding in automatic weapons later on. I want the signals I'm sending from the character controller to work with any weapon so I need is_action_pressed. My solution is to lock the hammer back until the player releases the trigger.

>>15627594

or just make a variable for semi or auto fire, then you can even do a semi-auto switch.

if auto == 1

is_action_pressed

if auto == 0

is_action_just_pressed

>>15627642

Oh I used that nomenclature for a while but it got confusing. Try something like IsUp, IsDown, as well as WasPressed, WasHeld, WasReleased

>>15627669

>>15627642

I figured it out. I added the variable "hammer lock back" that starts out as false, is set to true after the gun fires, is set to false when the trigger is released, and must be false for the "pull trigger" function to run. If the gun is full auto then the hammer isn't locked back.

>>15627868

Try:

[code]

Shader "Custom/CutOutShader" {

Properties {

_Color ("Color", Color) = (1,1,1,1)

_MainTex ("Base (RGB) Trans (A)", 2D) = "white" {}

_Cutoff ("Alpha cutoff", Range(0,1)) = 0.5

}

SubShader {

Tags { "Queue"="AlphaTest" "RenderType"="TransparentCutout" }

LOD 200

CGPROGRAM

#pragma surface surf Standard fullforwardshadows alphatest:_Cutoff addshadow

#pragma target 3.0

sampler2D _MainTex;

fixed4 _Color;

struct Input {

float2 uv_MainTex;

};

void surf (Input IN, inout SurfaceOutputStandard o) {

fixed4 c = tex2D (_MainTex, IN.uv_MainTex) * _Color;

o.Albedo = c.rgb;

o.Alpha = c.a;

}

ENDCG

}

FallBack "Transparent/Cutout/Diffuse"

}

[code]

>>15625716

>the grass

my dick

>>15627190

I feel we don't have enough to show yet, but the patreon exists

https://www.patreon.com/laetus

Oh no why did I go on Twitter

https://twitter.com/RSSLiam/status/1054901753710370816

This'll probably have a shitty theme like always

https://itch.io/jam/game-off-2018

>>15623440

awesome modelling and texturing, as usual.

>>15625716

really looking forward to seeing more of this. I liked that discussion a thread or two ago about tubers and variances in plant reproduction.

>>15629047

>inb4 another xbox dev-kit situation happens

He's already shot himself in the foot for posting about it before he has it.

>>15630972

I will try to keep the plants as realistic as possible, but sometimes it is hard to get all the info i need, im not a farmer and i don't know any to consult.

>>15628087

That worked thanks Anon. All the other forum posts I found had outdated shader code.

>>15628435

>I feel we don't have enough to show yet

>Has 2 demos ready

>makes weekly update posts

Nigger don't do yourself a disservice. Patreon is a good way to advertise your game and also grow a fanbase so when you release your game, you have people who are already looking forward to your game and aren't up shit creek without a paddle.

>>15632762

After this demoday then

>>15624959

Give each texture a name. The file name could work. Map this name to the atlas location. Have users of the texture be expected to know this name and hide the details of how atlasing actually works. So when you do Sprite.setTexture(’image.png’), have the function query Atlas for the name, then Atlas checks if the name is registered (i.e. is found within the name->offset map). If not, look for the file, open it and allocate the image into the atlas. Return the location. Sprite saves the location. When doing Sprite.buildQuad() use the location and dimensions to compute uv coords for the vertices. If your language of choice can’t operate efficiently on strings, map them to integers and use that internally. Do the mapping when defining the objects, I.e. don’t do setTexture(string) every time a sprite is instanced. Do it once when starting the game. Alternatively, build a list of unique sprites at startup and pass one of them to the constructor of a GameObject. So a particular Enemy or Item instance has a reference to the Sprite instance used to draw it and multiple such GameObject may share the same Sprite.

>>15629034

Looks great. Are you going to support custom scripting in levels as well?

>>15634454

>. I don't want the OST to take up three quarters of the game size.

Use midi or tracker format problems weren't.

>>15634454

It's pretty normal for textures and sounds to take up most of the game size, except you're generating textures through shaders and sound through midis. Code/Gameplay is generally the least space waster. Anyways, as long as you don't go fucking ham and end up with dozen fucking gigabytes as "modern" games do you shoudn't worry too much.

>>15634469

If your music is a good fit for a mod-tracker then you could really decrease the file size by using that. But if it's not then Ogg Vorbis is the way to go.

What compression settings are you using?

>>15634454

Sage for double post. I should mention in-case you aren't already aware that ogg isn't a compression format, it's a container. You can have an ogg mp3. Ogg is just a container for Vorbis the vast majority of the time so I assumed that's what you are using.

>>15634526

A few songs I have planned could potentially work in modtracker form but most won't.

Here are a few examples

https://files.catbox.moe/hsuq6v.mp3

https://files.catbox.moe/pfy7h9.mp3

>>15634533

I write them in FL Studio and adjust volume and such in Audacity. I output them in FL Studio at 192kbps, then adjust the volume, set the project rate to 48000 Hz and render them at quality level 5. Anything less than that (primarily 44100 Hz) fucks up the audio quality. No clue what it comes out to exactly after all that but it sounds good.

>>15634613

Very comfy music.

If you only have a few that work for modtracker than I would say keep them all as ogg Vorbis for consistency.

You probably saw the same post online that I did. I was going to suggest 44.1kHz at quality level 6 but it sounds like you already tried that.

I'm not sure, but it sounds like you're exporting from FL Studio as lossy. You should be exporting a lossless wav file from FL Studio if you're just going to run it through Audacity anyway. The file should only go from lossless to lossy during the final step.

>>15623440

Nice room, but the colors are mismatching horribly. Either change the blue part of the room to a lighter color, or remove the brown tints from the room. The room is to busy to notice important things/events. Moving the colors/saturation closer together would also work, but the main problem are the VERY blue walls.

>>15634811

>If you only have a few that work for modtracker than I would say keep them all as ogg Vorbis for consistency.

Yeah, that's what I was planning on. Is ogg Vorbis playback hard to implement in C++?

>Very comfy music.

Thank you.

>I was going to suggest 44.1kHz at quality level 6 but it sounds like you already tried that.

Yeah, most songs sounded okay but the first one I linked ends up garbled at 44.1.

>I'm not sure, but it sounds like you're exporting from FL Studio as lossy. You should be exporting a lossless wav file from FL Studio if you're just going to run it through Audacity anyway. The file should only go from lossless to lossy during the final step.

That's a good point, not sure why I didn't think of it first. I'll try that out, thank you.

>>15629034

This is really cool man! Looking forward to seeing more. Makes me wish I had done something similar for my game.

>>15633641

Okay, so that's more or less what I was thinking.

Is it common for an atlas to remove allocated textures? It's easy enough to flag the square region for reallocating, but having it re-merge or move textures around, kind of a shitty thing to have to implement.

Specifically, I made pic related as a "SpriteSheet" class. This let's me easily pass a number of parameters and get an animated set of sprites from an image, basically chopping shit up easily. If I shuffled the atlas textures around, I might have to eg update all of these values, and that was my biggest concern. Now mind you my game isn't going to be very big or resource intensives, but I don't want to shoot myself in the foot too much later.

I was going to have a manifest sort of file to load all the textures at startup (eg "res/img/spr_player.png" which becomes simply "spr_player" to load it) but some larger games load and unload textures as needed.

>>15634847

>Is ogg Vorbis playback hard to implement in C++?

Use a library. SFML, SDL, Fmod, Wwise. Don't try to do this from scratch.

>>15634886

If you actually need unloading, implement it. If not, don’t. You can add it later if needed, so don’t bother doing it now. If you do it, use the observer pattern to notify Sprite instances when their atlas position changes/gets deallocated. They should be flagged so you avoid drawing them (or more likely, ask to reload if they get used). The draw loop should tally the active sprites and pass that info to the atlas, so it can decide what to unload. But all this is for if and when you implement actual resource management.

Also, my ID has probably changed.

>>15635287

Implementing a resource management system is one of the next things on my todo list, so mostly I'm looking for dos and donts

>>15634454

I used ogg/vorbis and Portaudio library to write my sound and music system. Works really well for sfx and music playback. At some point I added some mutexes and made it fully multithreaded, which I think was my first threaded application. Pretty fun and educational.

>>15635554

It probably just spawns the appropriate objects. Since it doesn't receive player input, but instead follows a script (eg at frame 200 it calls player1.moveRight()). It could also fuck with physics during that time, then set them appropriately when it's time to play.

>>15635554

It's kinda like the Mario 1 demo if I had to guess. They probably had a sequence of actions that Mario and Luigi take and the rest of the enemies just follow their regular AI.

>>15634454

What everyone else said about .ogg but use Opus instead of Vorbis, it's the newer audio codec from the same people that compresses better/sounds better at the same file size. This is not something you would implement from scratch, you'd find an open source library and use that, which should pretty much work like "pass compressed audio in, get uncompressed samples out"

If you really wanted to get autistic about file sizes, you'd do the demoscene thing and write your own software synthesizer as a VST that "records" all the commands and timings you give it in your DAW, then can play them back in the game engine.

>>15635554

>>15635728

>>15635848

After reading 6k lines (thank god it was commented), I figured it out, the absolute madmen actually hardcoded everything, they didn't use ingame physics, instead they faked it with a lot of while and jump statements, the title screen is totally isolated from the core game.

Goto makes cutscenes really easy since you can jump between labels representing steps of the scene while also updating the screen, it's sad that this isn't the case with high level languages, it's the same how palette swapping isn't as easy as it used to be.

I'd like to keep my code clean, I hate the fact that my game loop is bound to have things that will happen only once, the only way left is let a state machine handle cutscenes.

>>15636258

If you're using a language that supports coroutines, you can implement cutscenes in a similar way. Even if it doesn't I would try to implement similar functionality myself because trying to do cutscenes with just states is going to result in a massive amount of different states and conditionals depending on how complex your cutscenes are.

>>15636258

I mean, C++ and C# have goto statements

Any vulkan gurus here?

I'm slowly learning it, and there are all kinds of things in vulkan that make no sense to me.

Like, why would you want a command buffer for every framebuffer in your swapchain? What could you possibly want a different command buffer per framebuffer for?

>>15637990

I barely know shit about Vulkan but

>Like, why would you want a command buffer for every framebuffer in your swapchain? What could you possibly want a different command buffer per framebuffer for?

So you can render different things in parallel. Like in a AAA engine that is very pipelined and doesn't give a shit about input lag, you could be rendering the "previous" frame while at the same time rendering a simplified version of the "current" frame to be used for occlusion culling. It's flexibility that you almost certainly don't need but is there if you do.

>>15637990

The Red Sky anon has to know a lot about it since he's been making his own engine with it.

>>15636350

I've done plenty of state machines and they weren't as bad, given that they stay simple, but I will use something akin to coroutines just because it's easier to write even but they're slower than a tight state machine, but hey, it's a cutscene, I doubt it'll need that much of speed.

>>15636393

Yeah, but it's considered to be bad practice, I'm using Nim which doesn't have goto for that same reason, but you can use long jump, either way, It'll be really bad using it within a loop, I got an Idea tho, I can easily update the game from different places with specialized loops.

I learned a lot from reading professional code, I keep stressing that my methods aren't the right way even if they get the job done, especially since I want the most performant way possible, after all I want to have thousands of objects at once on old hardware, but even they have hacks all over the place and it still works out.

>>15638981

>Yeah, but it's considered to be bad practice

Actually goto has one very good, specific use: Jumping out of a nested loop. This prevents you from having to use sentinel variables and checking them every iteration

>>15640011

>Quaternions came from Hamilton after his really good work had been done; and, though beautifully ingenious, have been an unmixed evil to those who have touched them in any way

- William Thomson, first baron Kelvin, 1892

Take solace in the fact that your hate of Quaternions is part of a long-standing tradition.

>>15640055

Thomson definitely knew his shit. Nothing else gives me that many surprises, not even shader magic, and of all the videogames i want to make the most of course it's always those that are heavily reliant on rotations like games with space fighting and hover vehicles.

>>15640324

Do two actors occupy the same tile at one time, or they just "pass through each other"?

Also does this seem like overkill?

https://hastebin.com/raw/sijuvusezo

>AssetManager class that lets me index any T object by index, name, or path

>Name and path are optional

Eg I can prepare a texture or spritesheet and shove it into this thing, and then I can then search by "texture 0" or "texture player"

>>15642740

rated

>>15642740

Spooked me into a snafu

>>15642740

Would/10

>>15642740

How does it look from above, like during the game?

>>15642716

This scenery looks extremely comfy.

>>15630141

Shouldn't the leaves at the beginning grow up thin and spread out later? The carrot grows like this and it looks much better than simply scaling it on all axes.

>>15642917

Is he okay?

>>15634869

What type of game are you making?

>>15638981

>Nim

My nigga.

>>15642716

Yeah having a palette to select entities from would be nicer, I've put that on my to-dos. I really like dusty kinda feel your use of particles gives.

>>15643437

Truly the Dark Souls of Diablo-likes

>>15643437

Don't take it out. But temporarily disable ragdolls that haven't been touched for a while, until they're touched again, or you might have memory leaks.

>>15643437

only if it screams constantly when you are near it

>>15644174

Well this puts my game to shame

>>15644328

something looks wrong with your Quake, try putting this in your console. Should fix it right up "gl_texturemode GL_LINEAR_MIPMAP_LINEAR"

>>15644328

Have you looked at the Widow-Maker from Arcane Dimensions?

>>15642843

The leaf structure of the potato is very different from the carrot, had to be like this to look more like the plant, but it is harder to animate.

how many years of schooling does it take to make a game? I'm 2 years into my game design degree and so far we've only learned javascript

>>15643510

yeah, I just finished implementing that. Ragdols screwed with performance in my gamejam games.

>>15643511

That may be too spooky, but I'll have my bro record a few sounds of suffering for curiosity's sake

>>15643766

that's very clever. I'll keep this in mind for some flavor lore

>>15643766

so spasms? I'm spooking myself just thinking about it. good idea

Your game/editor looks very cute, btw. Did you post it earlier?

>>15644328

how easy is it to post Hexen 2 models to Quake 1?

And a PSA: Occultus dev is still alive and started a kickstarter campaign. Those who played the demo day games likely remember this project.

https://www.kickstarter.com/projects/blackwoodsoftware/occultus

>>15645590

Ran that command, no changes.

>>15645653

Gonna look into it

>>15646981

I imported them into blender, then exported as a quake2 .mdl, the palette will be fucked up since Hexen 2 uses a different one.

>>15647440

>my name is inhuman Pete

>I dance the rumba beat

>>15642487

Muuuuch better.

>>15642716

You're Speebotanon right? WEBM looks great.

>>15627570

And here we have an anon who discovered why there is such thing as disconnector.

>>15647286

My problem is that I want to achieve that sleekness as a target, but I'm also too stubborn to do things in a certain way.

Incidentally, did you write the UI stuff all by yourself?

>>15648841

Legend of Zelda has various games with different fishing minigames. Link's Awakening comes to mind, where you cast the rod to get an initial distance/depth, but try to reel it in to avoid catching certain fish, eventually getting a heart piece

Shmup anon here

With a cold that didn't want to leave (it's leaving now), people at work waking up and asking various things, and adding the polyboard stuff for lasers in the engine, I almost fell into complete despair.

Started with basic, static, on the side rendering example. I thought that a triangle strip would do. Then seeing the result with sharp edge I decided that having the side vertices with alpha at 0 would fix this. Except that there would be interpolation from 0… To 0. So I needed an additional vertex in between, which mean the triangle strip wouldn't be nice to use. So, basic triangle.

Then I need to put that into the engine, but if I separate the whole thing, that would be code duplication for this specific laser object and the need for additional branching and testing.

[despair intensifies]

After a short break, I realized that if I consider the laser as a fast stream of points (which is not wrong) and only change the rendering for lasers from single point to group of points, then internally nothing changes (animation, collision detection), and I only need to manage those specific "projectile" differently when adding new points.

Hopefully I had some padding in my structure and I could hack into it a definition on how many points I need to parse to render the laser.

Now I would need additional shader work to make it lood "lasery enough", but right now polyboard works. I also checked with other shmup (two) and they also had some "wavy" laser when going left/right while shooting, but I can arrange this without too much difficulty if it's bothering me too much later on.

Taking a break is good from time to time.

>>15648856

I'm using SDL2 for input and SDL2_ttf for rendering fonts to textures but beyond that yeah, I'm not using a UI library or anything.

>>15648841

Breath of Fire 3 and 4

Our game will have a system based on this

>>15647286

How are you doing the bevel drawing on the UI, then?

One thing that's frustrated me in both XNA and SFML is that I can't really define a clipping region and only draw in that region. Eg, if I made a UI form and drew text to it, it could bleed off the edge

How did they handle modding using shared libraries (dll, so…)? I'm using a lot of function pointers to do so and I'd like to avoid scripting languages and their headaches, I've read valve's SDK and they don't seem to use function pointers, I'd like some guidance with this.

>>15650003

Oh so it's literally just a premade texture of a specific size?

>>15649743

You can probably achieve what you want with a scissor rect. It's one of the few pieces of GPU state that is cheap to set willy nilly.

>>15650759

What kills me is that both of those don't provide it out-of-the-box, with XNA/FNA completely obscuring it, but SFML requiring you to hit up external dlls to do OpenGL shit. It doesn't even have a depth buffer, either.

But yeah, if I need it, I'll probably end up doing that.

>>15650003

Don't use dynamic linking, especially for games on Linux. It is the entire reason why all games that get released on Linux are Ubuntu/SteamOS. You don't have to worry about distros having the right versions of whatever shit you use if you use static linking. You might not have much of a choice depending on what you're using, though (For instance, SDL hates it so much that they put in backdoors that let people override it: https://hg.libsdl.org/SDL/file/default/docs/README-dynapi.md).

The only downside is that if your game isn't open source, you are going to have to make new builds any time you feel it's necessary to upgrade library versions. Conversely, you won't have to worry about incompatibilities with newer library versions; the game will work the same as it always had.

P.S. pretty much every supposed advantage of dynamic linking is a lie. http://harmful.cat-v.org/software/dynamic-linking/

>>15638981

"If you want to go somewhere, goto is the best way to get there." – Ken Thompson

Gotos are only considered a bad practice because idiots use them in stupid ways. If you refuse to use anything that stupid people misuse, you won't have much to work with.

Don't follow other people's "best practices" blindly. Experiment and if it works for you, do it. Do not fear the rewrite. This is the only true way to learn.

>>15651488

How much of the detail is from the geometry, versus normal map, versus the texture?

>>15651500

That's the diffuse with curvature and AO baked in, so that's just 1 texture. I could make the normal map later. The poly count is 6k for the long one and 4k for the short one.

Well that's neat. My MathHelper class is my first "real" code file to ever reach over 900 lines

https://hastebin.com/siqixihugi.cs

>>15651484

I can create 3d models/textures and I can program. Let me ask some questions.

>What skills do you bring to the table?

>How complete is your design document?

>What engine are you using?

>How much of the game have you worked on/completed already?

>What tools would we be using to collaborate? I would like to know version control and chat software in particular.

Also, you should pitch your game with more information than the name of a series that spans 31 years, 6 console generations, and has gone from 2d, to 3d, to open world 3d, to micro-transaction cash grab. All I know is that it will be stealth related, lean heavily into military aesthetics, and that any engravings I may or may not give the guns will offer no tactical advantage whatsoever.

>>15650003

> really dislike the lack of a package manager on Windows, what a pain in the ass. Even OS X has brew.

Technically it has Chocolatey.

>>15651606

>any engravings I may or may not give the guns will offer no tactical advantage whatsoever

I'd disagree with Big Boss on that. A well-crafted and engraved gun could give you a psychological advantage: that the weapon you're holding is of absolutely superior quality and not just random trash you found, that THIS tool is going to carry you to victory

>>15651629

He never said it doesn't give an advantage, just that it doesn't give a tactical advantage. It's the difference between making the football team feel like they can win, and giving them the play that will actually take them there. Both are important to winning but one is tactical and one isn't. If confidence is a tactical advantage then "tactical" is too broad a term to offer a useful distinction.

Not to mention Big Boss does appreciate a well crafted gun. The context of the scene is important. Even though Ocelot has this supremely well made firearm and enough ego for 20 men, his gun isn't loaded with live ammunition so none of that matters.

>>15650003

>>15650822

Whatever you do, don't fucking static link OpenGL.

>>15651767

I don't think it's even actually possible. OpenGL does some crazy stuff to get the system libraries, which is why it's such a pain to init on windows without something like glew, iirc

>>15646981

Occultus dev if you're reading this

Get the fucking lead out if you care about your kikestarter and start contacting media outlets and niche audiences and communities(as well as some larger ones if you must)

Barely anyone fucking knows about this faggot

>>15651851

Obviously referring to the unpozzed "alternative" ones when talking about the media outlets

>>15651785

>OpenGL does some crazy stuff to get the system libraries, which is why it's such a pain to init on windows without something like glew, iirc

Not really.

All you do for windows is link against opengl32 (which is a windows system library), create an OpenGL context using wglCreateContext and then get the functions you need by using wglGetProcAddress. (which takes a function name and returns a function pointer)

GLEW saves you the busywork of defining every function, and also queries some stuff in advance (extensions available, device capabilities, etc.) so that you can query GLEW instead of pausing the entire OpenGL rendering pipeline like a fucking idiot.

>>15650003

>>15650822

The reason why you use dynamic linking is so that the libraries can be modified separately from your game. Consider it to be similar to replacing a script file in your project; it wont require a recompilation, which is useful in some cases.

A good example of this is how some people actually replace OpenAL DLL's with different ones to enable HRTF.

The reason why SDL wants to be dynamically linked really badly is because SDL's goal is to be compatible. Being dynamically linked allows SDL to be updated and modified separately from the application; which is incredibly useful when the software in question isn't open source.

>>15647286

Why are you moving from LuaJIT to Nim?

>>15652169

So how do you draw the shit without stretching the texture? Having to divide up and draw each element made from 9 little pieces can't be the best way eitger

>>15652668

I've grown to dislike how mushy and mutable everything in Lua is and have been wanting to move to a statically typed, compiled language for a while. I've tried a few different langues like C, C++ and even Rust (which I really didn't like) then stumbled on Nim and have been enjoying it the most so far.

>>15653276

That's the reason for defining the nine-patch regions, so you only stretch parts of it (edges and fill) when drawing to the render texture. After that, once a UI element at a specific size is in the hash-table (from "button_normal100x20" as an example) that texture pointer is used to avoid the need to redraw it piece by piece again. Nine-patches are a pretty common way of doing resizable texture based UIs.

>>15652163

Well the issue on windows is the opengl it actually ships with is old as shit

>>15653466

>Nine-patches are a pretty common way of doing resizable texture based UIs.

I can usually figure out stuff on my own by looking at how others do it. I didn't know that was the common approach, so that's good to know

>>15653739

What the fuck are you on about?

Do you actually do ANY research in this shit claims you make?

https://www.khronos.org/opengl/wiki/FAQ#How_Does_It_Work_On_Windows.3F

>When you run your program, opengl32.dll gets loaded and it checks in the Windows registry if there is a true GL driver. If there is, it will load it. For example, ATI's GL driver name starts with atioglxx.dll and NVIDIA's GL driver is nvoglv32.dll. The actual names can change from release versions.]

>>15650759

scissor and viewport setup will do it, yep.

glEnable( GL_SCISSOR_TEST )

then glViewport() and glScissor() with identical parameters to set the draw area.

draw stuff

then restore the old viewport and disable the scissor test.

>>15650775

someone may have already mentioned it, but you could render to a texture target and then copy that texture to the display buffer. Does XNA allow rendering to some other target than the back display buffer?

>>15654034

I'm with SFML now, but yeah you can in both cases.

Though, would you create a texture for each UI form? Or just have on that's the same size as the back buffer?

>>15653820

I already tried all of that, tried fixedupdate and update, tried every possible way of continous, continous dynamic, interpolation, extrapolationn and whatnot, tried rigidbodies, tried charactercontrollers, tried triggers. I went through every single way Unity can do collisions. All is fine at lower speeds but i guess Unity collisions simply have a hard time dealing with speeds of over 1500mph.

Spherecasts seem to work fine in my case though so luckily i've solved my problem.

>>15654049

>>15654034

Nevermind, I have no idea what I'm doing

The SFML version I'm using is the C# wrapper, and I can see examples of them using C++ code to call GL functions, but I don't know where it is even located, things just seem to work magically. I have like 14 .dlls already just to make SFML work

>>15654121

It should still work fine; that's one such problem that continuous dynamic was devised to solve.

I actually just tested it, and it works as intended

I tried 100k m/h, 500k m/h, 1mil m/h and it still works great.

If you're colliding with a kinematic rb, it needs continuous enabled too, btw.

>>15654596

You're going full ECS?

>>15654502

You're right and it turned out to be a really stupid thing. Found the reason in the documentation.

>MovePosition

>…therefore calling this method repeatedly without waiting for the next physics update will result in the last call being used.

I used MovePosition at two different times within the same update. I now do all the calculations with custom vectors first and at the very end i call MovePosition once, seems to work fine so far. Thanks.

>>15655117

Nah. So the whole thing is that I need to update actors in order from bottom right to top left. Both actor and tiles are structs for performance reasons, the downside being that individual operations eg MoveLeft() ,being a little more cumbersome.

The big issue I have is defining object behavior and I absolutely do not want to use inheritable classes for this. So rather I took my GameState which is a class strictly for modeling thr playback fof a level/map and broke it into different modiles. Thus I have an API of sorts, I pass in a GameState to the behavior.tick method or whatever, and then it has access to each modules functions in a vacuum, eg Status.AddTime or Tile.Exists

I don't want to define behavior in code but I don't want to define it in raw text either

>>15655199

I'm glad you got it working/figured it out, and happy to assist

>>15654190

>polishing these to not look robotic is hard

If you find any tricks lmk

>>15656249

General graphics tip for unity (since it looks like you haven't done this):

Go to edit, project settings, player and set the color space from gamma to linear

You might just love the result

There are times for gamma, sure, but realism likes linear

>>15656357

What kind of game are you making?

>>15656320

Yeah I'm not satisfied with the placement yet. I still need to think of a better map layout since I need to implement visual checkpoints so large chunks of the map can be occluded for performance, this is a VR game. I'll follow your advice. Was gonna make it a tad taller. Not life size taller though.

>>15656342

Just read up on that after reading your post. Interesting, thanks.

>>15656373

A first person shooter

>>15656373

a pvp game where you get a band of retards together, occupy an empty building, build it into a fuckmassive base that produces money, guns and gear for you then you go shit on other peoples bases from a great height for even more money + unlocks

basically urban rust without the busywork of whacking trees and rocks with an axe for hours on end

>>15656390

Ah, it's an An-Cap simulator. Sounds interesting. Good luck getting something of that scale to work if you're a one man team.

>>15656422

yeah, games like ~20% finished and I've been grinding on it for about a year now

multiplayer and all the basics are functional and stable but the code is pretty hideous to look at and will need a cleanup before (if) i bring anybody else into the fold

>>15656573

>2nd .webm

>All those triangles

Is my whole 3D knowledge a lie?

>>15656605

that guy only works with triangles, none of this is sculpted. Impressive though I'm not sure it's the right way to approach it

>>15656617

Well my boner thinks it's the right way.

Assuming I don't want to use any graphical 2d/3d objects, and want to make an entirely text based game, would it be easy to just do it from scratch in Python?

>>15656911

If you don't use anything except text for your game, you can do it easily in just about any language including Python.

>>15643511

>>15643437

Make it moan, or spazzy type moan or sceam, should fit like someone knocked down and killed, or make it where you can hear the blood pooling out of its body or something. Or just make it moan like a hentai girl.

>>15654190

I hope she can abduct the farmer and rape.

>>15656357

What kind of year is this? Storm or sanitation sewers? I am pretty positive that storm sewers are the only ones that get big. Also not sure how old that sewer is supposed to be but sewers dont make curves. The pipes dont curve let alone, if they do its just short turns, nothing sharp like a 90 degree turn, or even a 45. I can ask my professor if you want more information.

>>15656573

I hope her pucci is as flexible.

>>15625716

>lawn mowing simulator

the homeowner in me can't fucking wait

the person that loves video games in me hates me

Anyone else recently have a "why am i so retarded" moment?

I spent an hour on a makefile before i realized i could just put shell commands in it to accomplish what i wanted (moving files after compiling). After that I was done in 5 minutes.

Has anybody here tried making a game in WPF?

>>15654606

I suggest just adding a warning to the documentation that memory management must be manual, and patch it later if it becomes a problem. Most game mods are models and skins anyway, so you should focus on streamlining that instead IMHO.

>>15647873

I am, thank you.

>>15657485

Upload the game here. Just change the extension to .pdf and hope people will reupload it to other places.

>>15657485

If you use a different pseudonym for every single thing you release, how is it gonna be tracked to you?

Itch.io or even Steam might be your best bet if you want to sell it, otherwise you can use Gitgud.io and have public releases for it.

Must you release something 100% anonymous? If so, you'd need to upload it to somewhere that can't trace back to you and share the link in anonymous communities, but it's probably better to ask yourself first just how much anonymity do you really require. You might be putting too much effort for the same results.

>>15657485

You could hand it to someone who could upload it for you, though that'd probably get real complicated down the line.

>>15657494

>>15657576

Yeah, pseudonyms is probably the right option. Truly anonymous vidyadev is probably impossible without dumping your game on some torrent tracker, uploading it to an imageboard if it's small enough, or setting up an IPFS site and hoping people seed it.

>>15657769

>impossible

wrong word

>>15657485

I did some research on it since I wanted to make a "controversial" game without linking it to me, long story short, if you want to sell it, you're out of luck, itch.io and steam will happily give your real name to the press, and you should expect the same from every platform where you need to provide your personal data.

Releasing it for free is easier, you can just dump the files and let other anons spread it, lolisim has a repo that isn't maintained by the dev, and the dev only share it on anonymous platforms.

>>15658353

My motives are less about controversy and more that ideally I'd want people to go into my games with fresh expectations. Sure, someone might think "hey, I bet this is made by the guy who did X" and that's fine, but I'd like to pleasantly surprise people and attaching the same identity to everything could weaken that.

why haven't you made a hero shooter yet, anon? do you hate money?

>>15657485

>make new pseudonym

>get new VPN and/or tor

>only talk about your game under that pseudonym and connection

>upload it to a DDL site and post links on a free wordpress site or something when it's fucking ready

>>15658453

Get with the times grandpa. Hero shooters are dead, battle royale is the meme of the fortnight

>>15651851

I'm in here now and then, and yeah you're absolutely right. I'm just getting shit on right now with unexpected issues in other projects that have more immediate(as in days away) deadlines and I've gotta take care of them first. I've got some plans that I'll be acting on early this weekend though if things keep going apace.

>>15659632

>leaked_runescape_4_footage.avi

Took some advice from anons in the last thread and think I've finalized the skill list for my thief character:

1.Poison Strike: Inflict low Sword damage on an enemy. May inflict Poison.

2.Pocket Sand: May inflict Blind on an enemy.

3.Swift Strike: Inflict medium Sword damage on an enemy and raise own Speed until end of next turn.

4.Blade Flash: Inflict high Sword damage on an enemy based on own Speed.

5.Stormblade: Inflict medium Sword/Air damage on one enemy with further low Sword/Air damage on all enemies. May copy negative status from target to other enemies.

6.Steal Armor/Weapon/Accessory/Item/Money

7.In/Decrosa: Enhance/Reduce one ally/enemy's Speed for 4 turns

Alin/Aldecrosa: Enhance/Reduce party/all enemies' Speed for 4 turns

8.Amethyst Shield/Break: Enhance/Reduce one ally/enemy's Air endurance for 4 turns

9.Purple Edge: Inflict weak Sword/Air damage on 1 enemy

Gale Edge: Inflict Sword/Air damage on 1 enemy

Kamikaze Edge: Inflict heavy Sword/Air damage on 1 enemy

10.Gust/Algust: Weak Air magic, targets 1/all

Gale/Algale: Air magic, targets 1/all

Hurricane: Heavy Air magic, targets all enemies

11.Wind Boost: Slightly raise the power of Air magic. [passive]

Gale Boost: Raise the power of Air magic. [passive]

Tornado Boost: Greatly raise the power of Air magic. [passive]

12.Windfilter: Enhance own Air endurance [passive]

13.Locktouch: Treasure chests can be opened without keys [passive]

14.Rush Stance: Enhance own Speed and Critical rate for 4 turns. [trigger]

15.Sabotage: Inflict Poison and Agility Chain on all enemies [trigger]

16.Smoke Bomb: Reduce own Target Rate for 4 turns [trigger]

Trigger skills are similar to limit breaks in that they have to charge up before you can use them.

>>15656573

oh yeah, can u drop a source?

Would like to see their other work and techniques

>>15660543

Gondola Explorer is a game I want to play. The statue of Gondola.

>>15660275

>>15656617

>>15656605

Th "proper" way of making a realistic model is to first make one with all quads, sculpt it to be realistic, then recreate it entirely with a mixture of of quads and triangles trying to overlay the new model as much as possible on the old one while keeping in mind deformations and the like, and finally use the sculpted mesh to bake normal maps and shit.

Of course since modern graphics cards can shit out gorillions of triangles anyway, a lot of companies cut corners and just use sculpted meshes directly, while throwing more quads against it to get the right deformations.

If that guy somehow made those meshes manually without sculpting I salute him, but I highly doubt it.

>>15660543

Nice to see you're still around.

For my final I have to make a clone of a simple game in windows forms with C++.

>>15658453

>6 people make a nearly AAA-tier looking arena shooter in 2 years

Tell me again why AAA companies need many hundreds of people to make games and the games are still shit. You could only maybe justify it at the scale of games like GTA.

Those guys in the video would have just needed a few more people dedicated to designing/build maps, and someone to create character/weapon mechanics specifically.

>>15661021

Because they mostly hire worthless shits;

diversity hires are a hell of a drain.

>>15661021

Cause its just 1 map with 2(?) characters.

>>15661615

How many people do you think it takes to create a map or a character in 2 years if they're specifically hired to do that job every day? 6 people and 2 years includes all the engine work and splitting time across multiple fields for the artists.

I'm not saying you can make AAA games with 6 people, I'm saying there's no reason for it to require 600.

>>15660543

Simply ebin

>>15656605

>>15656617

Arma 3 requires you to triangulate your meshes.

>>15661021

>>15661719

More people speed the process up, which is why they frequently update content like Ubisoft's R6S. However, you're right about it being a drain.

>>15663166

Not sure raycasting is the way to go about your antifa flag simulation. :^)

Is there any reason why 3d animation looks less smooth for mobile games? It looks pretty choppy in most games that I've played in. What's preventing it from having IK?

>>15663166

See >>15663395. You want to do raymixing instead.

>>15661936

thanks

unsurprisingly they're a jap.

to be expected with that level of quality tbh

>>15662824

All mainstream engines I have knowledge of convert meshes to tris upon import, or as you mentioned require a triangulated model.

Afaik this is due to shaders and triangle rasterization.

So in the end it's "all triangles anyways", but obviously as mentioned and as you've read there's reasons to use quads over tris and vice versa when modeling.

>>15663400

Probably because it's the "low-bar" for game dev.

Shit-tier devs make shit-tier animations; who would'a thunk it

Choppy animation usually means low-fps animations and/or lack of interpolation.

It could also be hardware and/or software related.

>What's preventing it from having IK?

hardware/game design constraints, laziness

>>15663507

>Anyone have experience coding totally text based games?

yes, for a uni project in C++, it was fun for an exercise but it isn't really on the scale of a real game. it is good practice for a beginner tho.

also python is good for learning the basics, but I wouldn't recommend it for a game.

>What should my next step be?

collections, different types of loops

after that move onto a C-language (C#, C++, C if you want2 learn low-lvl).

>>15663507

>Anyone have experience coding totally text based games

Yeah.

Look into natural language processing. What you want is a dependency parser. I wouldn't recommend it for a beginner even if you're using a library. Make clones of atari games like breakout, pong, lunar lander, or asteroids.

Hi boys,

I started pluggin away at OpenGl, been followin the tut here: https://learnopengl.com/

Is this a good tut to get started with / are there any better ones?

My ultimate goal here is to recreate og doom.

>>15664630

Unless you're a super genius and/or I'm a super retard, you'll need multiple tutorials. There's no one place stop that gives you proper and comprehensive enough of an understanding of it to be able to make things on your own.

Here's 2 more

http://ogldev.atspace.co.uk/

https://open.gl/

>>15664630

I've used open.gl, but learnopen.gl looks to be a more detailed version of it now. (Especially it going in-depth into a VAO is very useful, since VAO's are very vague/weird objects.) It's good enough for getting started and getting triangles to draw.

If your goal is to make something widely compatible, then be sure to read documentation into what each object actually encompasses.

Implementation differs wildly between Linux, Windows, AMD, NVidia and Intel. Following the spec is the only way you're guaranteed to get (mostly) the same results, so understand what is and isn't allowed is important.

This is especially the case with Linux, as Linux implementations of OpenGL tend to be very strict compared to Windows.

An example of this, which tutorials often get wrong and glance over, is the fact that EBO's can only be bound in a VAO.

Every object in OpenGL can be created and bound to a global context, but EBO's are an exception to this rule. Because it is required to bind an EBO to fill it with data, this means you'll end op creating and binding a VAO and destroying it right after just to accomplish this.

Windows implementations don't follow this spec and instead often have a global hidden VAO that is bound when no VAO is bound, allowing you to bind an EBO anyway. But Linux implementations are significantly more strict. The ogldev.atspace.co.uk tutorials that >>15665912 mentions don't actually do this correctly, and I've had to modify them their code in order to get them to actually work on my Linux system.

Another thing to read into is how shaders actually work. Shader code cannot branch, nor can they dynamically decide which memory objects to access. Yes, you technically can write code, and yes, the glsl compiler of your driver will just accept it and run it, but you'll end up with visual glitches in some cases. Read into dynamically uniform expressions.

Using shaders incorrectly causes graphical artifacts based on a whole slew of factors, which vary widely on different brands of hardware too.

>>15666148

>I've used open.gl, but learnopen.gl

fugg I meant learnopengl.com

>>15634389

Looks really nice

>>15668784

nice

>>15668816

>>15668666

you too, my dudes. Nature is the perfect inspiration

Kinda program would you suggest for developing an FPS game ??

>>15669268

Unity, to make pong first, then arkanoid, then tetris, then asteroids, then you can go make your FPS game

I actually wish there was a "easy to make 3D game" to add to the list, but there is no such thing as "easy 3D game"

If you are making this question, you are just about to get started making game, I'd bet

If that's the case, I suggest you start with easy stuff

Like in the martial arts animes, you work up your basics to mastery before learning the ultimate technique

>>15669295

>Amateur Gamedev General

Can I use unity for 3D games ?

>>15669325

Someone I know at church suggested Unity (should I buy premade models ?)

>>15669327

No, don't buy anything, there's enough free resources online.

>>15669032

Now I want a titty tree.

>>15669327

Never buy models

>>15669333

Can suggest a site for free resources ??

>>15669339

If you go for Unity as the engine

https://assetstore.unity.com/search/?k=rating%3A5+released%3A31+price%3A0&order_by=relevance&q=rating%3A5&q=released%3A31&q=price%3A0&rows=42

>>15669316

And many regrets were made

>>15669345

is Unreal better than Unity ??

>>15669339

https://free3d.com/

https://www.cgtrader.com/free-3d-models

Just look at it in google, there's a shitload of them.

>>15669355

only if you are a experienced team of professionals

>>15669345

I've more tutorials for it on Unreal

>>15669355

Depends, the most important factor is knowing how to use the engine, both can make great games, you need to know what you are doing

>>15669316

I love the schizophrenic feel of you guys posting under one ID

>>15669325

>Can I use unity for 3D games ?

It's better at 3D than 2D.

>>15669327

>should I buy premade models

It's in the spirit of agdg to make your own stuff, otherwise what >>15669333 said

>>15669339

I'd recommend starting by looking at the wiki

http://8agdg.wikidot.com/graphics

>>15669365

Maybe we are one guy with two computers :^)

>>15669325

>>15669327

discontinue your cancer meme spacing

>Can I use unity for 3D games ?

yeah, unity is called "Unity3D" for a reason.

>should I buy premade models

First, you should make an effort to keep your art style consistent.

There's plenty of free resources online; look on the agdg wiki under graphics.

Make your own models, or get an artist after you've prototyped a game; which may attract one.

>>15669032

best post is anime tittie tree post

>>15669355

They're both great engines.

I prefer Unity because C#, better docs, more consistent updates/features, and things like consistent GDC and unite events + community outreach; also I've used the engine for many years so its grown on me (so I'm heavily biased).

>>15669295

>>15669316

For context this ID is a pair of dev brothers posting from the same IP; it's not a schizo

>>15669372

Unity is also not owned by fucking Tencent, but that's more a ideological problem than Engine quality

>>15669295

>pong

>tetris

>arkanoid

>starting on easy modo

Where's the fun in that. It's better to aim for the stars and start with a dream game.

>>15622921

How well does C# work in godot? Is it on par with GDScript yet?

>>15669807

There's no documentation and from what I hear there are things you can't do that you can in GDScript.

>>15669268

>>15669325

>>15669327

>>15669339

>>15669355

Why are you writing so many posts in a row? You seem like an ESL 12 year old due to your retarded questions and lack of English skills.

Does mesh deformation and animation happen on the CPU or GPU?

>>15670373

I think you're trying to remember delegates/events

>>15670662

It depends on the engine, and their implementation.

Animation usually occurs on the CPU, but for skinning (transforming actual mesh); this can occur on the CPU or GPU.

Unity can do either for skinning (tho, afaik it's not included OOTB, you need to grab the plugin from unity's github).

>>15670415

It's a wayward redditor

>>15670862

>Correction

GPU skinning for unity isn't included OOTB afaik.

CPU skinning is the default.

>>15670373

>>15670862

Those are events. Here's a functional code example I just made

https://hastebin.com/fuqokufano.cs

And this is a bit of a read, but it's the best article on events I've seen

https://www.codeproject.com/Articles/20550/C-Event-Implementation-Fundamentals-Best-Practices

>the gdnative tutorial is garbage

>the error messages are even worse

>>15672180

I don't want anything to do with anything that Gitcuck is involved in.

>dedicate a WHOLE MONTH of your year to this

No.

>>15672192

But anon, Github is promoting FOSS projects and what's the fucking point of that if they're owned by Microshit now :^)

>>15672180

>dedicating a month to a github jam

>when you could be delivering something for demo day instead

Working on my input recorder next. The way I have it set up is that my GameState object increments a counter every tick, which is a ushort, so it repeats after 65536 ticks. This means that instead of tracking an int (4 bytes) per change of input, I only have to track a ushort (2 bytes).

The downside is that a replay recording will break if you stand perfectly still for 18 minutes without pushing any buttons, but the timer should have run out by then, so I think it's fine.

>>15672284

>Theme is Hybrid

>Submission is "hybridized" to be released for Demo Day AND the Jam

>>15671488

Is there a specific reason why this code doesn't work in Unity?

It's just telling me

>The type or namespace name EventHandler<> could not be found

>>15672500

Apparently you have to define them as System.EventHandler and System.EventArgs fuck if I know why

>>15670862

>>15670872

Wew, if I understand it correctly, that means applying a transform with weight to every single vertex in the model? that would saturate the CPU pretty fast

>>15672500

Add "using System;" to your script. There's a lot of shit that needs you to add their namespace to the top if you want to use it like Serializer or Unity.UI, etc.

Also, Anon… It's the first result if you google for it: https://stackoverflow.com/questions/18094063/the-type-or-namespace-name-eventhandler-could-not-be-found

Always google your errors, someone already had them and asked for help before you did.

>>15672500

You can just type "using System;" on top to avoid having to type System. before that. Otherwise you have to type it because you're using that Namespace.

>>15672508

>that would saturate the CPU pretty fast

It's multithreaded in unity, and no not really; as accessing an array of vertices in linear order to transform them works in a DoD manner.

I.e. harnessing how the CPU works, little to no cache misses, and the logic is very simple; so it's quite fast in actuality.

The issue comes about when there's extremely high vert counts with quite a few models on screen at a time.

However, that's what LoDs mitigates, and if you have spare GPU budget with those high poly models then you can always offload it to the GPU (which has many more cores, and this job is perfectly suited to, as there's no branching and fits in the 16kb cache size of most GPUs quite well).

>>15672500

Search engines are your new best friends m8

>Try out Choosenim on Windows because hadn't tried it yet and sounds convenient

>Compile errors about sizes, even after double checking my Nim and installed GCC are both 64-bit

>Turns out Choosenim 'doesn't support 64-bit yet' and installs its own 32-bit GCC without warning

>Get SDL2 and MinGW set up after that

>Last thing I need is SDL2_gfx but the fuck doesn't offer binaries

>Install and set up CodeBlocks to import that VS solution file

>Compile the shared .a library version but it complains when trying to compile DLLs

>Need to download the Windows SDK for the .lib files it needs

>Still complains after that, even when pointing it directly to the folder I can see the .libs in

>Hear it might be an issue with importing to CodeBlocks so install VS and all the bullshit that needs

>Have no luck getting that to compile DLLs for SDL_gfx either so say fuck it, what am I actually using it for anyway

>Was only using it for FPSManager which is just a convenience shit around SDL_Delay so remove that and do my own

>Now game crashes on launch, turns out the placeholder array I had thrown in for history states was bigger than the default stack size on Windows, which was a good thing to learn about

>Game launches from Git Bash fine but when I launch it from Explorer I get an error about not being able to find the entry point for libpng

>Turns out the Git folder has a mingw64/bin too that the Bash shell executable uses and the zlib that libpng depends on in that is newer than the one in my C:\mingw64\bin

>Finally up and running

Well that was a learning experience. Also there seems to be an odd discrepancy between Windows and the other platforms on memory usage. On OS X and Linux I see ~70MB but on Windows I see <20MB.

AHAHAHAH

I feel like a fucking genious.

So here I am, thinking about Serialization to save my game and the state of every object in it.

It works fine for the world and the terrain since they are known classes.

It would even work fine for Furniture or Items since they inherit from a base class too. Except these have their own specific scripts with their own variables that I want to save as well. (And a few components, but that's not important for now).

For example, I have Oven.cs that inherits from Furniture.cs and has a bool saying if it's On or Off. How do I save this?

Well it's simple enough: Make XmlOvenData that inherits from XmlFurnitureData and put it in a List<XmlFurnitureData> on the XmlTileData to be saved!

Well, fuck, it doesn't work. When serializing the tile data, it iterates over that List and either uses the base type XmlFurnitureData or complains that there's something in it that's not XmlFurnitureData and it wasn't expecting that. How do we solve that?

Simple!

List<System.Type> extraTypes = new List<System.Type>();

extraTypes.Add(typeof(XmlFurnitureData));
extraTypes.AddRange(System.AppDomain.CurrentDomain.GetAllDerivedTypes(typeof(XmlFurnitureData)));

XmlSerializer serializer = new XmlSerializer(typeof(XmlAreaData),extraTypes.ToArray());

This lets the Serializer recognize extra type of data, and on top of that, any time I make a new Xml(somethiing)Data for a new Furniture type, it's automatically added to this list and recognized, using that type instead, enabling me to never look at this part of the code again.

This means that, as long as I set up said Xml(somethiing)Data, that piece of Furniture will be properly serialized regardless of anything else and all I gotta do is make sure it's constructor records whatever specific variables there are for that object, even including other components it might have!

And I can freely re-use this for Items and Mobs so every inventory can save any item in whatever state it currently is, while every Mob can be composed by whatever number of limbs (separate objects) I want and even be missing them as well.

I have no idea what the industry standard for saving\loading data is, but the fact I managed to cobble this up together on my own makes me feel fucking great. I bloody love writing code and solving problems.

>>15672572

>Choosenim

Yeah that's a known issue, the devs want to make it 64bit only and people who want 32bit will use the installer, I used the installer because it's less prone to errors.

>SDL2_gfx

I know that feeling, that library is pure trash, currently it's just a wrapper for openGL, took me days to get it to compile only to know that even if you force it to use 1.1 (to support old machines like mine) it doesn't work, it still use recent version texture which made it a total waste of time for me, its code isn't special anyway, you're better off with RayLib or SFML, although I had the same problem with RayLib because the available binaries are set for recent versions.

>default stack size on Windows

How much space did you allocate? I've read that you can readjust it somehow.

>On OS X and Linux I see ~70MB but on Windows I see <20MB

That's odd, if anything, I'm having a problem with win7 not wanting to reclaim memory after freeing it, that may be something worth asking the devs for it, Nim has a discord/irc/gitter with a bridge connecting them, join whatever platform you're comfortable with if you didn't do it already.

>>15672690

Smart approach, and nicely done.

I'll definitely keep this in mind when i do more in depth serialization

>>15672706

>How much space did you allocate? I've read that you can readjust it somehow.

I read that you can increase the stack size but it's not a big deal. It was really just me being retarded allocating around 2MB on the stack needlessly. The old Lua version of my editor was very wasteful with memory and kept the entire level state as separate versions in the history, this time I'll just track delta changes and apply/unapply then when stepping through. I wonder if the memory difference is just in how Windows reports memory in use. There also seems to be a difference between what Task Manager states (~14MB) and Visual Studio debugger (~20MB). I'll probably look into it more under more controlled situation later on.

All this messing around makes me wonder if I shouldn't just go full enginedev mode when I move my main project over in the future and say fuck SDL completely.

>>15661021

As the team grows so does issues with management, leadership, and talent acquisition.

If you don’t have developers with a clear vision and corporate savvy your big game will be mediocre and designed by committee.

Without proper management a large team of highly specialized people will waste a lot of time, you’ll never know when you’ve hit milestones, and miscommunication will kill your game.

Finally You need good people and your competing with unicorns(and decacorns as of late) for the best programmers in a niche job market. So you buy an engine and grab some lower tier devs(by no means bad, but not geniuses) to make your game on top of. Artist and fx guys are easier, but still very competitive.

Doing a creative endeavor as multi-faceted as game development at scale seems like a recipe for disaster that may not even make it out the door, let alone be economically successful, so just copy what’s been proven to work and try to mitigate all the risk.

I never buy into public facing developers for the reasons listed above. PR glamour probably conceals the real people who stick their hands in this and get things done.

Sage because we should be working on games here.

>>15672828

SDL is already enginedeving anon, it's just more simple for 2D

>>15672920

SFML is even simpler. It even has Simple in the name!

>>15674977

>dat trigger discipline

>>15675164

Nice logo.

>>15675801

Thanks.

So is it a copyright infringement if i use the system from a pen and paper game for a video game? Thinking of using cyberpunk2020 btw.

>>15676684

Is the company still around?

Do they have something proprietary and copyrighted that no other pen and paper RPG did?

If so, potentially so trudge somewhat carefully. Though I think the transition from P&P to vidya is significant enough and will create enough changes to put you in the clear, but be careful with any terminology invented specifically for that system.

>>15676710

Yeah thats what I figured. Although is it possible to copyright the math behind a system? I mean, isn't a pen and paper system copyrighted as a pen and paper system and nothing else?

>>15660543

A walking simulator i'd actually play

>>15676744

Not necessarily, Fallout was originally going to use the GURPS system but had to be reworked into SPECIAL because the licensor of GURPS didn't like the violence in the game (since normally it'd be used for text-based games):

http://archive.is/7IK4S

You can probably take general ideas from Cyberpunk/Interlock System as long as you don't make a carbon copy.

>>15676684

As long as it can be qualified as "derived works" you're protected.

Don't copy it word for word, nor should you use trademarked terms; just inject your own creativity into it with your own twists and you're good.

>>15676744

>copyright math/formulas

This is a grey territory, but basically the novel expression of ideas are copyrighted intrinsically by the creator (intellectual property protection); though it's the expression of the idea that's important (results, what it creates, the expression of the idea).

So, this is why only the specific implementations of algorithms developed by the pioneer to solve specific problems can be patented/copyrighted (must be an "original work").

Basically, the algorithm/math/formula itself cannot be copyrighted, but the expression of that work can be copyrighted (so, the application of the algorithm to solve particular problems); in addition to being patentable if they want additional protections (must of been an "original" expression).

>>15676822

Can I look up copyrights? When I go to the website it just kind of shows me the company and the name and whatnot. Where can I get the exact details?

>>15676946

I'm not entirely sure about finding more information, but I think the information provided is sufficient here:

LINK

https://cocatalog.loc.gov/cgi-bin/Pwebrecon.cgi?Search_Arg=cyberpunk+2020&Search_Code=TALL&PID=_2kR57hrjcR3FUVBnO32xAct6Za&SEQ=20181103001214&CNT=25&HIST=1

It's just "text" which is covered by the copyright, and the expression of that text is the RPG systems, lore, narrative, and so on that creates cyberpunk2020.

If you take some of these, inject your creativity, create your own story/lore/etc (so it's original); then you're fine, as long as your work's expression isn't "cyberunk2020".

I.e. if it's substantially different to that original work in its expression, even if some parts of it are derived from it in an original way; you're protected.

To clarify, it cannot be an unoriginal "derivative" work (I worded this poorly in my previous post).

So, you cannot create a carbon copy of their systems/etc while renaming things then just name it cyberpunk2022.

>>15676946

Also, think of it like rougelikes, soulslikes, and so on; as long as it's not a carbon copy it's fine.

Just be creative and make the expression of the work original.

>>15676946

Most games are iterations of previous work, if not clones. Just don't copy exact texts or artwork or trademarks.

>>15677053

this pretty much, it's not worth overthinking tbh

Shmup anon here.

After the laser thing episode, finally got to make the background part. At least for the first stage (Ocean)

With some digging, there was a link showing some example with displacement mapping but with Maya, and from their UI perspective (click here, add this with this setting). Nevertheless I gave a try with using displacement mapping… It looked pretty shitty.

After some more digging, there was a paper talking more in details about this kind of rendering with displacement, and I realized why I was retarded: I had to add some delta in sampling the texture and THEN have those derivate value to calculate the new normal.

I hope I can at least make a small something for demo day.

has anyone implemented A* Pathfinding in godot?

>>15677472

There's a class for it: https://docs.godotengine.org/en/3.0/classes/class_astar.html?highlight=astar

You'll also want to look at the Navigation/Navigation2D nodes.

>>15677472

Also, Godot 3.2 is targeted for Recast/Detour integration, which should aid in the creation and navigation of 3D navigation meshes. The latest build might have semi-working versions available.

We want the Amulets & Armor audience

>>15678443

Huh, I had no idea the autosage limit was increased

>>15678477

Dunno what your specs/software are but it really shouldn't take that long.

>>15678443

>>15678457

Nevermind, I think I've got a more solid grasp on events.

Still though, is there a better way than manually having each of my class instances subscribe to each other's events?

Pizzatower is looking good. I heard rumors some of his stuff got here. Is It true?

>>15678786

this is some good shit

>>15678786

That looks cool but the music ruins it

>music for dark atmosphere with a red sky

>I know, how bout we put some Hawaiian guitar in the background

>>15678914

am I being baited? If not, don't worry. This isn't the music the game is gonna use.

>>15678786

Have you considered not having really short fog and giving the terrain some elevation difference? Or is this just a test from your top-down project?

>>15679083

>test from your top-down project

you got it. The scene can't really have elevation with the iso view I tried, so this is the way of making it look acceptable. FPS mode definitely works better in the dungeon.

>>15679067

nah, it will be on-disk DLC, but before that a Kickstarter stretch goal

>>15678786

Music makes me think of a shitty porno. Looks bretty gud though.

>>15678786

If you are going for first person you need to tone down the fog, for some areas it would be ok, but not the entire time. Also the music doesn't match the setting very well, it would probably work better on a cyberpunk game.

>>15679589

Anime characters don't really have makeup as makeup is only required for 3DPG.

Makeup in anime is utilized to exaggerate a wacky character or some feature of their personality.

Blushies (how cheeks naturally get that rose tint, but anime style so a mesh circle+ texture), deep eye circles, different lip colors, maybe eye liner (darker eye outlines), eyebrow style, hairstyle, and these little things are enough to communicate a character's personality in most cases.

Is there a way I can import 3d skeletons into Godot? My .dae files from blender aren't importing at all, usually giving a blank file.

>>15679589

To be fair, it's very easy to get banned from /a/.

>>15679682

Sometimes it is, sometimes it isn't. At least the moderation and userbase aren't nearly as cancerous as /a/'s until /a/ gets salty and starts crossposting.

>>15679698

Don't use Blender's default collada exporter, it's garbage. Install https://github.com/godotengine/collada-exporter or a glTF 2 exporter like https://github.com/KhronosGroup/glTF-Blender-Exporter

>>15679720

This type of fog is very cool, i was talking about the red fog restricting view range.

>>15679724

I know they are grammar nazis there, but it was an embarrassing post

>>15679731

Gonna lurk there a bit then later

>>15679682

Makes sense, and yeah it fits w/anime style's conventions.

The way I'd recreate that character in a 3d model is way different than your current style though.

I'd do the dark outlines around the eye makeup stuff with a thin mesh w/a black texture, and have mesh eyelashes so it occludes the eye makeup from different angles; while having higher priority in shaders so it's not occluded by hair (w/some alpha blending or something)

Then, the makeup eye liner would be a small mesh matching the profile of the makeup in that pic, which could be enabled/disabled, and is shrinkwrapped to the face mesh; then you can just texture/paint that separately.

The specular blushie highlights would be seperate meshes with a specular material (maybe billboard not sure I'd have2 test it).

Then the blush itself would just be a quick texture edit, but would be JUST that, as the base color of the character would be done via doing a base color in the shader settings (so texture color gets added, so blush would be subtle like in the pic).

So, less a single texture, and more separate mesh pieces with a reliance on shaders for certain effects in that image.